Bulletin of the Geological Society of America Vol. 67

Home , Lake Chippewa, Valparaiso Moraine

BULLETIN OF THE GEOLOGICAL SOCIETY OF AMERICA VOL. 67. 271-288. 6 FIGS.. 1 PL. MARCH 1966

LATE WISCONSIN CHRONOLOGY OF THE LAKE MICHIGAN BASIN CORRELATED WITH POLLEN STUDIES

BY JAMES H. ZUMBERGE AND JOHN E. POTZER*

ABSTRACT Geological interpretation and radio carbon dating of an excellent exposure in a wave- cut cliff on the shore of Lake Michigan near South Haven, Michigan, permits the es- tablishment of the following absolute chronology of events for the Lake Michigan Basin: Two Creeks-Bowmanville low-water stage, 11,000 years ago; end of Lake Algon- quin, 8000 years ago; Lake Chippewa low-water stage, 5000 years ago; and beginning of the Nipissing Great Lakes, a little less than 4000 years ago. Pollen profiles from the South Haven section and from a bog near Hartford, Michigan, on the Valparaiso ground moraine reveal the following forest succession in southwestern Michigan from Two Creeks time to the present: spruce-fir to Jack pine to white-red pine to oak to oak-hemlock-broad-leaved forest to oak-pine. Radiocarbon dating indicates that the close of the spruce-fir period was 8000 years ago; the Jack pine period, 7000 years ago; the white-red pine period, 6000 years ago; the oak-pine period, 5000 years ago; and the oak-hemlock-broad-leaved forest (Xerothermic), 4000 to 3500 years ago. The difference in time between Lake Algonquin and the Nipissing Great Lakes was 4000 years. The Xerothermic is correlated with the Nipissing Great Lakes at 4000-3500 years ago.

CONTENTS TEXT Page p Summary of pollen studies at South Haven . , .. 2% and Hartford bog 286 IntroductioT n 272 f e cited 287 Acknowledgments 272 Re erenc S Resume" of the Late Wisconsin history of the ILLUSTRATIONS Lake Michigan basin 272 Figure Page Previous work 272 L Index map 274 Glacial Lake Chicago 272 2 Wave-cut cliff exposure along Lake Michigan Lake Algonquin. 273 shore near South Haven, Michigan 275 at r 01 111 sta es J; ? . Ate "! ? . g 273 3- Chart showing geologic section and radio- Nipissmg Great Lakes 273 carbon dates from the South Haven Algoma Great Lakes 273 s;te 276 Low-water stages in the Lake Michigan 4. Correlation "chart'of'geologic'events'with basin.. c,'-;,'^ T;;-1;.-"" %* forest succession 277 Geologic section near South Haven, Michigan .274 s Pollen diagram of the South Haven section.. 278 Description 274 diagram of the Hartford bog 282 Geologic interpretation 277 6 Pol]en Summary of geologic events as indicated by Plate Facing page the South Haven section 279 j. Views of the South Haven section 274 Radiocarbon chronology 280 Previous work 280 TABLES Discussion of South Haven radiocarbon Table Page dates 280 l. Comparison of Antevs, Flint, and Zumberge- Pollen studies of the South Haven and Hartford Potzger chronologies 283 sites 281 2. Pollens of shrubs, trees, and herbaceous Introduction 281 plants, and spores of ferns counted at Methods of collecting and preparing samples South Haven 284 for pollen analysis 281 3. Pollens of shrubs, trees, and herbaceous Pollen profiles 283 plants, and spores of ferns, Hartford Discussion 283 bog 285 —— 4. Climatic change inferred from forest history * Deceased, September 18, 1955 in southwestern Michigan 286 271

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 272 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

INTRODUCTION Harbor, Michigan, deserves credit for the initial discovery of the South Haven exposures. For more than 50 years geologists have been Finally, the authors wish to express their studying the late Pleistocene history of the sincere appreciation to the following who pro- Lake Michigan basin. From this mass of infor- vided the C14 dates appearing in this paper: mation a detailed sequence has been deciphered, Richard Crane, University of Michigan Phoe- ranging from the earliest impounding of glacial nix Project Radiocarbon Laboratory; W. F. water (Lake Chicago) to modern Lake Michi- Libby, University of Chicago Radiocarbon gan. Fluctuations of the ice front, shifts in Laboratory; Hans Suess and Meyer Rubin, U. outlets, and differential uplift of the land have S. Geological Survey Radiocarbon Laboratory, contributed to the changes in elevation of the and Wallace Broecker and J. Laurence Kulp, water plane during the final wastage and dis- Columbia University Lamont Geological Ob- appearance of the last Wisconsin ice sheet. servatory. One of the more striking later contributions to this history has been discovery of extreme low-water stages. Prior to the recognition of RESUME or THE LATE WISCONSIN HISTORY OF these stages in lake history, geologists assumed THE LAKE MICHIGAN BASIN that modern Lake Michigan attained its present level through a series of consecutive steplike Previous Work lowerings (owing to various causes). The framework of the present concept of late This paper presents further evidence in sup- Wisconsin history of the Lake Michigan basin port of these low-water stages of the Lake had its inception with the report of T. C. Michigan basin and integrates the generally Chamberlin in 1877. Since then, additional accepted chronology into an absolute time scale findings by other workers have added much to based on radiocarbon age determinations. A the sequence of events. (See Taylor, 1894; further aim is to correlate climatic changes as Leverett, 1899; Alden, 1902; 1906; 1918; J. W. indicated by forest succession in southwestern Goldthwait, 1906; 1907; 1908; Leverett and Michigan during late Wisconsin time. Pollen Taylor, 1915; Wright, 1918.) More recent con- studies by J. E. Potzger have been employed tributions by Stanley (1938), Bretz (1951), and as the basis for this correlation. J. H. Zum- Hough (1952; 1953; 1955) are especially impor- berge made the geological investigations and tant in that they supply evidence bearing on interpretations. low-water stages below the present level of Lake Michigan. ACKNOWLEDGMENTS

During the field investigations, geologists, Glacial Lake Chicago palynologists, and ecologists visited the exposures near South Haven. Among these were J The proglacial lake that formed at the south- Harlen Bretz, the late Leland Horberg, Bjorn ern margin of the retreating Lake Michigan lobe Andersen, Pierre Dansereau, and Stanley A. was named Glacial Lake Chicago by Leverett Cain. Their discussions at the site and exami- (1897, p. 65). Three stages of this lake are re- nation of results is greatly appreciated. J Har- corded by shore features above the modern len Bretz and the late Leland Horberg read the lake: the Glenwood stage, 60 feet above the manuscript and gave the writers valuable present lake; the Calumet, 40 feet above criticism. modern lake level; and the Toleston at 20 feet Appreciation is also expressed to Charles above the modern water plane. The reasons for Shinn and William Grable of Larchmoor for the changes in level were discussed by Bretz their co-operation in providing access to the site (1951) and need not be considered here. itself, and to the Michigan Congregational The term "Calumet stage" as used in this Christian Conference for hospitality extended paper is the same as Bretz's "second attained and for permission to work on the Pilgrim Calumet level" (Bretz, 1951, p. 427) and Calu- Haven property. Mr. Harvey Franz of Benton met II as used by Hough (1953, p. 99).

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 RESUME OF LATE WISCONSIN HISTORY OF LAKE MICHIGAN BASIN 273

Lake Algonquin Nipissing Great Lakes Lake Algonquin was the first body of water The Nipissing Great Lakes occupied the area to occupy the entire Lake Michigan basin. The now covered by Lakes Superior, Michigan, and position of the ice border at the time of the in- Huron. Most of the earlier students of Lake ception of the Algonquin stage is not known, Nipissing considered its level to have been about but Hough (1953, p. 101) believes that it was 15 feet above modern Lake Michigan (Alden, somewhere in the Northern Peninsula of Michi- 1918, p. 337; Leverett and Taylor, 1915, p. 449; gan and that much of the eastern end of Lake Tague, 1946, p. 17) or about 595 feet A.T. in Superior was ice-blocked at that time. This the southern Lake Michigan basin, but Hough would not permit Lake Algonquin to communi- (1953, p. 89-90) concluded that the Nipissing cate with any part of the Lake Superior basin and Algonquin water planes were both at 605 as supposed by Leverett (1929, p. 65) and feet A.T. before upwarping. This difference in Sharp (1953, p. 120). In either interpretation interpretation is not critical to this study al- Lake Algonquin ("highest Algonquin" of though the writers prefer Hough's reconstruc- Stanley, 1936, p. 1954) used the St. Clair- tion of the Algonquin-Nipissing history. Detroit River and the Des Plaines River outlets. The reconciliation of these two ideas is not Algoma Great Lakes critical to this study. Lake Algonquin in the Lake Michigan basin was at about the same Younger beaches 10-15 feet above the pres- elevation as Lake Toleston (605 feet A.T.). The ent level of Lakes Michigan and Huron were exact relationship of the Toleston stage of Lake ascribed to the Algoma stage of lake history by Chicago to Lake Algonquin is not completely Leverett and Taylor (1915, p. 449). Tague resolved, although Hough (1953), p. 37) con- (1946, p. 18) followed Leverett and Taylor's sidered that the Toleston stage ended when the interpretation, but Hough's interpretation Valders ice retreated sufficiently far north to (1953, p. 92) required all beaches referred to the open the Traverse Bay lowland or the Strait of Nipissing Great Lakes by his predecessors to be of Algoma age. Mackinac. Apparently no appreciable change in level occurred when this event took place Low-Water Stages in the Lake Michigan Basin Later "Algonquin" Stages Into the general scheme of events just out- lined, two low-water phases have been injected. As shown by Stanley in his work in the At least one lake stage below the Nipissing Georgian Bay area (1936; 1937), four prominent level was recognized by Alden (1918, p. 332), beaches lie below the highest or main Algonquin recorded in buried peat deposits near Sheboy- beach. Stanley demonstrated that these beaches gan, Wisconsin. He was not certain, however, are essentially parallel to the Algonquin beach when this low-water phase occurred with re- and were produced by successively lower eleva- spect to the other known stages. Baker (1910, tions of the water surface as lower eastward p. 715) also noted a buried peat deposit under outlets were uncovered by ice recession. Chap- Calumet beach gravels near Bowmanville, man (1953) disagrees with this interpretation, north of Chicago, which he considered evidence but, since the present study is concerned pri- for a pre-Calumet low-water stage. However, Leverett (1915, p. 356) did not consider the marily with the Lake Michigan basin, the Bowmanville exposure conclusive in this re- writers can add no new data in support of spect. either interpretation. For the sake of consist- Bretz (1951, p. 410), after Baker's interpre- ency, the four "lower Algonquin" stages (Wye- tation, correlated the Bowmanville lower-water bridge, Penetang, Cedar Point, and Payette phase with the Two Creeks interstadial. The in order of decreasing age) are used as originally lowest water level of the Bowmanville stage is defined by Stanley and followed by Hough unknown, but it must have been lower than the (1953, p. 68-9). present level of Lake Michigan, and Hough

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 274 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

(1953, p. 34) has suggested that it was at least (Hough, 1952). It fits the "standard" chronol- 100 feet below the present lake level. The Bow- ogy between the Algonquin and Nipissing manville-Two Creeks correlation is now a gen- stages and would occur in time after the Payette erally accepted part of the "standard" chronol- stage of Stanley (1936, p. 1948).

FIGURE 1.—INDEX MAP

ogy of events and occurs in time between the The major events of the history of the Lake Glenwood and Calumet stages of Lake Chicago Michigan basin are thus well established, al- (Bretz, 1951, p. 427). though several details remain to be added. Stanley (1936; 1937; 1938) postulated a low- Some of these details are added here. water stage in the Lake Huron and Lake Michigan basin between Algonquin and GEOLOGIC SECTION NEAR SOUTH HAVEN, Nipissing time with a connecting channel MICHIGAN through the strait of Mackinac. Hough's studies of bottom sediments verified an ex- Description treme low-water phase in the Lake Michigan basin (350 feet below present lake level or 230 Between May 1952 and June 1954 the writers feet A.T.) which he named Lake Chippewa made several joint visits to an excellent section

PLATE 1.—VIEWS OF THE SOUTH HAVEN SECTION FIGURE 1.—BURIED SAND DTJNE ON LACUSTRINE SANDS Looking south; see Figure 2 D for location; Melhorn (1954, Ph.D. thesis, Univ. Michigan) FIGURE 2.—BURIED SOIL PROFILE ON LACUSTRINE SANDS WITH AN ORGANIC LAYER HIGHER IN THE SECTION Looking east; see Figure 2 C for location FIGURE 3.—CLOSE VIEW or BURIED SOIL PROFILE Same location as Figure 2 FIGURE 4.—BURIED PEAT LAYER Looking east; see Figure 2 B for location

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 BIT.,: GEOL SOC AM, VOL. 67 ZUMBERGE AND POT7OER PL

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 GEOLOGIC SECTION NEAR SOUTH HAVEN, MICHIGAN 275

near South Haven, Michigan, exposed as a wave-cut cliff along the east shore of Lake Michigan (Fig. 1). The locality lies in SWJ£ NEJ4 sec. 28, T. 1 S., R. 17 W., Van Buren County, Michigan (South Haven Quadrangle, U. S. Geological Survey). Figure 2 is a diagram- matic cross section showing the major features of the exposure. The vertical scale is exag- gerated to emphasize the stratigraphic relationships. The basal formation consists of a laminated, compact, blue-gray silt. The laminae are con- torted and deformed. The base of the silt layer is not exposed anywhere in the area and it presumably lies below the present lake level. It probably rests on till. The top of the silt is ir- regularly exposed at or near the lake level of May 1952 (approximately 582 feet A.T.). The contact with the overlying deposits is uncon- formable and is marked in many places by an accumulation of organic material, which ranges in thickness from a few inches (Fig. 2, B) to over a foot (Fig. 2, A). Wood fragments were X removed at B (Fig. 2) at the top of the silt layer. o § The South Haven pollen profile (Fig. 5) begins H in this organic layer (Fig. 3) at the 107-inch level. t-J Unconformably on top of the silt is a unit of g q bedded littoral sand with fine lenses and stringers of gravel. Medium to coarse sand and fine to medium gravel is characteristic of the formation. At B (Fig. 2) the sands contain thin lenses and streaks of organic matter. These streaks were sampled for pollen and constitute the part of the South Haven pollen profile between the O H 30-inch and 107-inch levels. The thickness

U ranges from 8 to 30 ieet in half a mile of length. At D (Fig. 2) a sand dune lies above this formation. The contact between the dune and the underlying bedded sands is shown in Figure 1 of Plate 1. The top of the lacustrine sands is marked by a buried soil zone (Fig. 2). It can be traced con- tinuously from the mouth of the creek near the north end of the section for about 3000 feet south along the exposure. It truncates the bed- ding of the sands and rises up over the buried !• ' I sand dune near the south end of the section 11 (Fig. 2; PI. 1, fig. 1). This buried soil profile is developed in a high-quartz parent material and has a well-developed B horizon (PI. 1, figs. 2, 3). No trace of calcium carbonate could be

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 276 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

OF SOUTH HAVEN SITE I See fig Z tor location]

SURFACE 25-|

DUNE SAND • - CARBON 14 SAMPLE LOCATQN 24 J M - UN VERSITY OF MICHIGAN LAB C - UN VERSITY OF CHICAGO LAB

L - i-A UONT LABORATORY >Vv . RADIOCARBON DATES 9- ^^vVV\V\ i FROM SOUTH HAVEN SITE

UN1VOFM1CHLAB UNIV OF CHIC LAB LAMONT LAB

4000 * 160 8- \ M-291 4000 * 3'JG C- 149 ;4000 * 350} 4816 * 290 T £ l°H M-290 0000 * 400 6232 * 310 > 7- C-846 < Si 85 * 400 6650 ' 350 5 L-214 20 (5090 * 3001 ',6440 * 230} 5i3O*l50 z

M-288 8350 * 500 30 BSs^l! C-846 7500 ' 500

M1CHIG A (7925 * 400) 6744 ' 530 3=5=!^^:=

. -— —^~=^ 4O LAK E

50 ABOV E FEE T

_ _ - -

1 - 90

LEVEL OF L MICHIGAN I MAY, (952, 582' AT j 100 UD3EOL SURVEY LAB

M 268A H^^^fl|^^^iM-1t ~ lO? W-167 :i,rGO ' 600 ,0,860 ' 350 r!^^^^^^^^^^^^.BLUE"- SILT ~

FIGURE 3.—CHART SHOWING GEOLOGIC SECTION AND RADIOCARBON DATES FROM THE SOUTH HAVEN SITE Figures in brackets are averages where more than one run was made of the same sample.

detected in the sand beneath the B horizon for The peat layer constitutes the upper 30 inches a depth of more than 10 feet, and it is question- of the South Haven pollen profile (Fig. 5). able whether these sands were ever calcareous. The soil zone and peat layer are covered by At B (Fig. 2), the buried soil grades laterally dune sand containing several discontinuous into a 30-inch layer of peat containing lenses of bands of organic material (Fig. 2; PI. 1, fig. 2). iron-stained sand, and wood fragments (PL 1, These bands are similar in gross appearance to fig. 4). Pollen samples were taken from each inch the buried soil profile, but differ in that they are of peat thickness, and larger samples were se- not continuous for a distance of more than 400 lected for radiocarbon determinations (Fig. 3). or 500 feet.

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 GEOLOGIC SECTION NEAR SOUTH HAVEN, MICHIGAN 277

The stratigraphic relationships of the South of the Lake Border moraine sometime during Haven section provide the basis for the inter- Glenwood time. Another possibility is that they pretation of geological events of late Wisconsin represent periglacial involutions associated with time. the later exposure of the silts to subaerial con-

YEARS BEFORE LAKE STAGE IN THE PRESENT THE LAKE MICHIGAN WISCONSIN CHRONOLOGY (based on C-l4doles) BASIN (elev.above seo level)

Lake Michigan (580)

2,500 Algomo (595)

3,500 Nipissing (605) XEROTHERMIC

4,000 N (Depth fromtap v of pollen profile 5,000 Chippewa (230) Opening of North Payette and Bay Outlet 6,000 related stages of "Lower Algonquin' lakes of Stanley. (Depth from lop of pollen profile ICE RETREAT 6,000 in inches) ^~» 8,500 Algonquin (605) Toleston (605) Calumet 1620) VALOERS MAXIMUM

11,000 Bowmonville (below 580) TWO CREEK INTERSTAOIAL

Glenwood (640) CARY 13,000 FIGURE 4.—CORRELATION CHART or GEOLOGIC EVENTS WITH FOREST SUCCESSION

Geologic Interpretation ditions during the Bowmanville low-water stage. Of the two possible explanations, the A cursory examination of Figure 2 reveals former seems more plausible. two major breaks or discontinuities in the sedi- The Glenwood silts emerged when the water mentation record. These are represented by the fell from the Glenwood stage to the Bowman- organic layer on top of the basal silt and the ville low-water stage, thus ushering in Two buried soil higher in the section. Both are typi- Creeks time. The unconformity marking the cal unconformities and record emergence from upper limit of the Glenwood lacustrine silts is beneath the water body in the Lake Michigan correlated with the Two Creeks interstadial, basin at some time in the past. Both stages re- and the peat and wood associated with the un- quired a water plane lower than the present conformity must be equivalent in age to the level, and two such periods of low water have peat at Two Creeks, Wisconsin (Libby, 1952, already been established in the Lake Michigan p. 88). A radiocarbon date of the wood col- area. lected at B (Fig. 2), resting on the silts, was If during the Glenwood stage of Glacial Lake reckoned at 11,200 ± 600 years by the Univer- Chicago lacustrine silts were deposited in this sity of Michigan Laboratory and 10,860 ± 350 area, the basal blue-gray silt must be of Glen- years by the U. S. Geological Survey Labora- wood age. The present elevation of the top of these silts is about 580 feet A.T., and the level tory (Fig. 3). This further corroborates the of the Glenwood water plane was 640 feet A.T.; interpretation that the lowermost organic zone so these silts accumulated in water 60-70 feet is a correlative of the Bowmanville low-water deep. phase of Two Creeks time. The contortions of the laminae are difficult The rise of the waters from the Bowmanville to explain. Overriding by glacier ice is a possi- low-water stage to the Calumet stage at 620 feet bility if, as Hough (1953, p. 96) suggested, the A.T. submerged the organic material and silts Lake Michigan lobe advanced to the position beneath shallow waters of Lake Chicago. Depo-

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 278 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN SOUTH HAVEN PEAT EXPOSURE

WueWooD

» "I|- I •IM-|-|l>"lil|||||

FIGURE 5.—POLLEN DIAGRAM OF THE SOUTH HAVEN SECTION sition of littoral sands and a few lenses of fine thin lenses or stringers of organic matter. These gravel was initiated by this rise to the Calumet organic bands were sampled for pollen grains stage. At B (Fig. 2), about 6 feet of lacustrine and represent the 106-inch to 31-inch section sand accumulated above the basal silt. How- of the South Haven pollen profile (Fig. 5). ever, interbedded with the sand are several Deposition of the sand layers alternating

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 GEOLOGIC SECTION NEAR SOUTH HAVEN, MICHIGAN 279

with thin organic layers requires an environ- the time involved in the drop from Algonquin to ment of minimum wave agitation; in this case Chippewa, through the "lower Algonquin" a protected bay along the shore is most plausi- levels, and the rise from Chippewa to Nipissing ble. A spit or barrier beach must have existed levels. lakeward from B (Fig. 2) in order that the sand Uplift of the North Bay outlet (Hough, 1953, and organic material could accumulate. The p. 86) re-established the southern outlets and sand may have been inwash from waves break- caused the waters to rise from the Chippewa ing over the barrier, or it could have been level (230 feet A.T.) to the Nipissing stage (605 blown into the protected waters by wind. At any feet A.T.). The strand line was thus returned rate, the bedded sands with thin organic layers to the South Haven area, and eolian activity at B (Fig. 2) represent continuous deposition was renewed. Sand dunes began covering the with no significant breaks until the beginning of soil and peat of the South Haven section before peat accumulation. The uniformity of the the Nipissing level was fully attained, and con- forest history as shown by the pollen profile tinued rise of the water eventually drowned the (Fig. 5) for this part of the section also favors pre-Nipissing—post-Chippewa dunes. Shore such an interpretation. erosion since Nipissing time has destroyed all If the interpretation is correct, this part of evidence of the Nipissing beach at the South the section includes all of Calumet, Toleston, Haven site, but presumably it existed west of and Algonquin time. Because much of the shore the present shore line. area has been eroded by post-Nipissing wave Modern eolian activity in the area, restricted action, and because most of the lower part of to the local formation of blowouts and the the South Haven section is badly slumped, it is shifting of sand over dunes previously stabilized impossible to reconstruct the exact shore con- by vegetation, accounts for the discontinuous ditions at this locality during the three stages organic bands in the dunes above the buried of lake history mentioned. The Calumet shore soil. Lutz (1941) used a similar reasoning in lies east of the present shore in the South Haven studies of fine-textured material in sand dunes area, but probably both the Toleston and Al- of Cape Cod and Glen Haven, Michigan. The gonquin shores once lay west of the present local encroachment of wind-blown sand onto shore. The buried sand dune at D in Figure 2 forested parts of near-by dunes is a common required a source of sand west of the present phenomenon today along the entire eastern shore line. shore of Lake Michigan. The waters of Lake Algonquin began to re- cede when eastern outlets were opened by fur- Summary of Geologic Events as Indicated by the ther recession. Hough (1952; 1953, p. 71) pre- South Haven Section sented clearly defined evidence from bottom cores in Lake Michigan for an extremely low- The South Haven section indicates the follow- water stage in the Lake Michigan basin, 350 ing events: feet below modern lake level (230 feet A.T.). (1) Withdrawal of the Lake Michigan lobe Hough named it Lake Chippewa, and his recon- from the Valparaiso moraine and formation of struction (1955, p. 966) shows its eastern shore basal silts in Lake Chicago during Glenwood line about 20-25 miles west of South Haven. time Westward recession of the strand line from (2) Possible overriding of the Glenwood silts the South Haven area at the end of Algonquin by readvance of the ice time eventually stabilized dunes such as the one (3) An extensive retreat of the Lake Michi- perched at D (Fig. 2). This stabilization, owing gan lobe, resulting in Bowmanville low-water to removal of the source of sand, is indicated by stage during Two Creeks time. Represented at the soil profile and the associated peat now South Haven by the lower unconformity locally exposed in the South Haven section. Radiocar- containing wood and peat. Radiocarbon dates bon dates from the top and bottom of the peat of 11,200 ± 600 years (M-288A) and 10,860 ± at B (Fig. 2) indicate a duration of about 4000 350 years (W-167) verify this correlation. years for the peat bog (Fig. 3). This 4000 years (4) Rise to Calumet level owing to glacial includes all of Lake Chippewa time plus most of readvance; Valders stage

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 280 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

(5) Continuous sand deposition behind bar- mum. The C14 date of sample M-288a (Fig. 3) rier beach at the South Haven site during requires a date of slightly less than 11,000 years Calumet, Toleston, and Algonquin time; some ago for the Valders maximum, whereas Antevs dune activity (1954, p. 520) considers the Valders maximum (6) Ice retreat, exposing eastern outlets and to have been reached about 19,000 years B.P. development of the "lower Algonquin" levels. Another difference is that Antevs places the Opening of North Bay outlet culminated in the Xerothermic period (his Altithermal), which is drop to Lake Chippewa. Soil and peat formed the age of maximum warmth since retreat of at the South Haven site. Radiocarbon dates at the ice, at 7000-4500 years ago (Antevs, 1952, top and bottom of the peat bog indicate a p. 204), whereas the writers conclude from the 4000-year time span between beginning of drop South Haven radiocarbon dates correlated with to Chippewa level and return to the Nipissing pollen studies that the Xerothermic was not level. Eolian activity was resumed during post- reached until sometime after 4000 years ago Chippewa-pre-Nipissing interval. The differ- and is, in fact, correlative in time with the ence in age between Lake Algonquin and Lake Nipissing Great Lakes. Finally, for the date of Nipissing is thus fixed at 4000 years. the opening of the North Bay outlet (Mattawa (7) Post-Nipissing time has been one con- Valley) Antevs (1954, p. 520) has proposed tinuous lowering of the lake, but shore erosion about 13,700 years B.P., but the writers place during that time has destroyed all evidence of this event at about 5000-6000 years ago. post-Nipissing levels. Dunes are now locally The writers' chronology is in disagreement shifting owing to blowouts. with Flint (1953, PL 3) in the age of the Xero- thermic. Flint considers the Xerothermic (his RADIOCARBON CHRONOLOGY Thermal Maximum) to have reached a maximum about 5000 years ago. Under his interpre- Previous Work tation, the Xerothermic antedates the Nipissing Great Lakes and had actually ended prior to A recent attempt to establish an absolute the low-water stage in the Lake Michigan chronology of late-Wisconsin history in North basin (Lake Chippewa) between Algonquin and America (Flint, 1953) was based on radiocarbon Nipissing time (Flint, 1953, p. 911). According dates from the University of Chicago laboratory to the writers' interpretation the Xerothermic (Libby, 1952). Strong disagreement with this is a correlative of the Nipissing Great Lakes and basic outline skeleton of dates has been ex- therefore postdates Lake Chippewa. Table 1 is pressed by Antevs (1953; 1954; 1955) who a comparison of some of the Antevs and Flint builds a geochronology of the "Deglacial" and dates with the ones proposed by the writers. "Neothermal" ages based essentially on varve analyses correlated with recessional moraines. The chief reason for Antevs' disbelief in radio- Discussion of South Haven Radiocarbon Dates carbon dates stems from his insistence that an ice cap disappears chiefly by retreat of the Without the stratigraphic information neces- margin. Apparently, he does not consider wide- sary for reconstructing geologic history, C14 spread and almost complete stagnation and determinations are obviously meaningless. catastrophic disintegration in situ a valid pos- Furthermore, geologist must not depend upon sibility. For those who accept the radiocarbon C14 dates that are obviously anomalous. dates as essentially correct, the concept of the The writers present in Figure 3 a comparison rate of disappearance of the final shrunken ice of several radiocarbon dates determined by sheet must be modified to include the possibility three different laboratories. These are among of rapid disintegration in place. the first dates published from the University of The chronology shown in Figure 4 differs Michigan laboratory and represent the first from those proposed by Antevs (1953; 1954) instance in which a cross check on dates for such and Flint (1953, PI. 3). The major difference a large series of samples has been made. None between the Antevs chronology and the one of the three laboratories had knowledge of the proposed here is the date of the Valders maxi- dates determined by the others until after all

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 RADIOCARBON CHRONOLOGY 281

dates were completed. Each laboratory used a late Pleistocene history of the Lake Michigan different method of age determination; Michi- basin. In addition, because of the excellent ex- gan and Lament used carbon dioxide gas, posure, several samples of peat could be easily Chicago used carbon black, and the Washington and accurately obtained for radiocarbon laboratory used acetylene. analyses. The agreement in order of magnitude of the As a check on the South Haven profile, which various dates as well as the consistency of re- was truncated both at the top and bottom, a sults in terms of the geologic section is striking. deep bog was bored 15 miles inland from the In each case, a single sample was divided into South Haven site near Hartford, Michigan equal portions and sent to the various labora- (about 660 feet A.T.), well above levels of the tories. In the case of sample M-290 (C-848, water bodies formerly occupying the Lake L-214), a piece of buried log was split into Michigan basin. several samples so that the radiocarbon dates shown in Figure 3 are all based on the same Methods of Collecting and Preparing Samples piece of wood. The writers had hoped for a three- for Pollen Analysis way check on all samples, but, because the U. S. Geological Survey laboratory was not in opera- A yard stick was placed vertically against tion during the early part of this study, and the exposed peat; a pocket knife cut a clean because all the laboratories were extremely surface and took samples at 1-inch intervals busy with a tremendous back log of samples, for the total depth of 30 inches. The samples only a two-way check for each sample was pos- were placed in waxed paper accurately marked sible. for level and later transferred to small vials and The writers believe that the objective ap- securely stoppered. Samples were taken of or- proach to the radiocarbon dates as presented in ganic streaks contained in sand beneath the this paper provides a sound basis for the peat for 75 inches, giving a total pollen profile evaluation of the process in general. Most of 106 inches (Fig. 5). A sample was also taken geologists are not qualified to judge the methods of the buried soil associated with the peat. and theories involved in the radiocarbon-dating A deep bog on the farm of Mr. Clair Lamb, procedure, but as geologists we can attest to here called the Hartford bog, which probably the reasonableness of the dates insofar as they represents the filling of a kettle lake in the Val- are in accordance with sound stratigraphic paraiso ground morainic upland southeast of interpretation. The South Haven dates seem to South Haven (Fig. 1), was sampled in April fulfill this requirement. 1954. The boring was made with a Hiller borer to a depth of 32 feet. Samples were taken at POLLEN STUDIES or THE SOUTH HAVEN AND 6-inch intervals and placed in stoppered glass HARTFORD SITES vials. The bottom 8 feet of the boring was Introduction through rock flour. Further penetration of the rock flour was discontinued because of the Pollen analysis has established itself as a danger of twisting off the sampler. valuable tool for the study of forest history Because the peat from the South Haven and with inferred climatic changes during late Hartford sites separated readily, the Geisler Pleistocene time. Most pollen data are collected (1935) alcohol method was used to prepare the from borings in peat bogs where there is little pollen slides. In a few instances the separation or no stratigraphic information that can be the was in hot alcohol. Gentian violet stained the basis for geologic interpretation. It was thus a pollen grains which stood out prominently rare opportunity to find a buried layer of peat from the particles of other organic substances, (Fig. 2 B) exposed in cross section on a vertical and thus could be easily and accurately counted. cut, permitting easy and accurate sampling at Except for a few foot levels of the Hartford bog inch intervals. The particular value of the South and some of the organic streaks in the sand Haven pollen profile, however, is the fact that underlying the South Haven peat where counts the South Haven peat could be related to the of 100 pollen grains were made, at least 150

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 282 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

HARTFORD BOG

bbbbbbbbiflbbbbbo'bbbbw'bbbbbbbbbbbbbbu'.Ti

FIGURE 6.—POLLEN DIAGRAM or Tint HARTFORD Boo

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 POLLEN STUDIES OF THE SOUTH HAVEN AND HARTFORD SITES 283

pollens of trees were tabulated at each level. relative of the 21-foot to 10-foot levels of the For the South Haven peat, a count of 200 was Hartford bog. made at almost all the inch levels. The two Radiocarbon dates of the South Haven peat pollen profiles are shown in Figures 5 and 6. (Fig. 3) permit the following correlation of Pollen representation of shrubs, herbaceous forest change with time: close of the spruce- plants, and ferns at the two stations is shown fir period (beginning of peat formation at in Tables 2 and 3. TABLE 1.—COMPARISON or THE ANTEVS, FLINT,. AND ZtTMBERGE-PoTZGER CHRONOLOGIES Pollen Profiles All dates are in years before the present. The lower 42 inches of sand of the South Antevs Flint Zumberge- Haven profile (Fig. 5) was deposited during a Event (1953; 1954) (1953) Potzger decided spruce-fir period (up to 94 per cent). Between the lower 42 inches and the base of the Xerothermic 4500-7000 5000 3500^000 peat (the 68-inch to 30-inch level) spruce-fir Two Creeks 19,000 11,000 11,000 shows a steady decline, with a corresponding increase in jack pine. South Haven), about 8000 years ago; jack pine The 30 inches of peat represent 4000 years period, about 7000 years ago; white-red pine 14 according to C dates (Fig. 3) and record at period, 6000 years ago; oak-pine period, 50001 least five or six major forest changes. Succes- years ago; oak-hemlock, broad-leaved forest, sions are from spruce-fir to jack pine to white- 4000 years ago. red pine to oak to oak-hemlock-broad-leaved By correlating the truncated South Haven forest. Spruce and fir almost disappeared during profile with the Hartford profile, it follows that the pine period. No evidence of pronounced the Xerothermic (climatic optimum) reached tundra is indicated in either profile. During the its maximum development after the cessation of final stages of the growth of the South Haven peat deposition at the South Haven site, 4000 peat, paludification expanded the bog area into years ago. This is indicated by the 9- to 2-foot an extensive wet meadow that can be corre- levels of the Hartford profile. Since the end of lated with the buried soil for a considerable the Xerothermic, an increase in pine at the distance along the exposure (Fig. 2). The pollen Hartford site (0- to 2-foot levels) suggests records of the soil horizon closely resemble climatic deterioration. those of the 2- to 3-inch levels of the South Haven profile. Discussion Shrubs of greatest significance in the South Haven profile are alder (Alnus) and hazelnut The physical features of the Hartford bog (Corylus). Grasses have the highest representa- suggest continuous deposition since the time tion from the 21-inch to the surface levels, and the meltwaters from near-by glaciers poured spores of ferns in places reached numbers equal rock flour into the depression, while pioneer to those of total tree pollens (Table 2). forests of spruce and some fir added pollen to- The Hartford bog is completely enclosed the inorganic sediments. This early forest (verlandet) and now has a dense growth of blue- (lowermost 9 feet of the Hartford profile) must berries with scattered white pine and broad- have been a strikingly simple association, be- leaved trees. The Hartford pollen profile (Fig. cause not even jack pine participated in the 6) no doubt records forest changes during much crown cover. In this respect the forest history of post-Gary time. Although the forest succes- differs greatly from that of lower Quebec, as sion at the Hartford bog is similar to the one described by Potzger (1953), and from that of indicated by the South Haven profile, hemlock the Laurentian highland regions of central is much less abundant in upper levels, and a Quebec as Potzger and Courtemanche (1954) prominent pine peak becomes evident in the found in a recent study. If jack pine is, indeed, closing foot levels. Apparently the South Haven so extremely overrepresented by pollen as it peat layer (1-inch to 30-inch level) is the cor- usually seems to be, then the Hartford moraines

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 284 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

TABLE 2.—POLLENS or SHRUBS, TREES, AND HERBACEOUS PLANTS AND SPORES or FERNS COUNTED AT SOUTH HAVEN

1 .>. «*H i Depth in inches No. of tree birc h 2 pollens ~s 1 .£ holl y 1 fern s 1 a J3 Dwar f Hazelnu t Deciduou s Chenopod s Sphagnu m < o C/5 i U Crt Clu b mosse s Osmund a £

Surfa.cc 185 2 2 2 10 i 200 1 3 14 4 4 4 2 1 7 2 175 1 8 1 1 6 7 6 2 1 3 3 200 1 3 12 1 5 8 7 5 1 15 4 150 1 1 30 2 5 17 8 10 90 24 5 200 1 1 28 2 1 4 4 6 1 8 6 200 3 3 17 7 4 13 5 4 4 15 7 150 1 33 4 1 22 17 47 8 150 3 1 14 2 4 2 8 4 31 9 195 2 25 3 5 2 5 11 7 60 10 150 4 1 1 6 13 1 1 1 2 19 26 11 200 1 3 21 2 2 3 16 6 43 12 200 17 1 4 23 4 102 13 200 4 19 3 1 127 2 23 14 200 4 22 1 1 1 15 8 1 16 150 2 11 1 1 1 11 5 19 17 200 4 10 1 1 1 20 13 18 150 2 1 4 1 1 19 1 10 20 200 5 1 1 10 1 14 13 21 200 7 2 16 1 4 1 20 1 5 22 200 5 1 9 1 2 11 1 6 23 200 1 1 1 4 2 4 13 11 24 200 4 1 3 1 2 25 200 10 4 9 31 3 26 200 11 2 4 1 1 2 1 1 27 200 2 4 1 1 1 28 200 1 1 7 1 29 200 3 1 14 1 2 1 30 200 2 13 3 1

35 150 12 2 2 36 200 3 1 1 6 2 1 2 1 2 37 150 2 1 4 2 38 150 1 16 2 1 40 150 2 3 24 1 1 3 1 1 42 150 1 7 3 1 1 46 200 1 10 1 1 3 1 60 150 4 1 5 1 1 64 150 1 3 74 100 1 4 1 79 125 4 1 82 150 1 4 98 150 4 1 1 107 (clay) 150 5 1 3

Soil horizon 150 5 1 2 60 1 8 14 1 14

TABLE 3.—POLLENS or SHRUBS, TREES, AND HERBACEOUS PLANTS, AND SPORES or FERNS, HARTFORD BOG

No. of s birc h tree fer n Depth in feet DOl- 1 mosse s B fern s lens holl y 1 I Cattai l Dwar f Ericad s Deciduou s Shiel d Clu b 1 0 OJ s u Wate r lil y Osmudn a £ 1

Surface 100 1 7 6 52 1 6 6-inch 150 1 8 2 27 1 7 18 1 1 150 1 1 19 23 1 1 16 2 2 150 1 24 2 2 79 3 150 12 1 9 2 21 1 4 150 1 53 20 1 5 2 2 64 3 5 150 3 3 16 1 7 1 2 99 4 6 150 ? 1 9 1 5 1 5 7 4 7 150 1 1 8 1 1 7 1 1 1 8 150 1 2 7 17 3 1 3 1 11 1 12 9 150 2 9 2 4 19 1 2 3 10 150 2 1 15 1 1 7 4 3 11 175 3 2 2 26 4 1 1 2 2 1 12 150 2 1 12 1 1 5 1 13 2 4 13 150 1 3 10 1 1 1 1 1 11 4 14 150 7 g 14.5 150 2 4 3 15 150 5 5 4 16 200 S 4 17 150 10 18 150 4 4 18.5 150 1 S 19 150 3 4 20 200 2 5 21 150 2 3 22 150 2 23 150 S 23.5 150 2 24 150 8 25 150 2 26 125 6 27 150 1 28 150 8 29 140 1 30 140 5 31 150 32 150 2^

must have had only a few scattered individual assume that southwestern Michigan must have trees. Moreover, it is therefore even more had a similar forest cover 7000 years ago. striking that at the 20- to 18-foot levels of the The forest history gives no indication of Hartford profile, and at the 28- to 24-inch great climatic fluctuations. Just what the levels of the South Haven peat, jack pine climatic significance of the double peaks for indicates a decided control of the forest. In a chestnut-oak and the lesser two hemlock peaks recent airplane trip into the Quebec tundra, in the South Haven profile is cannot be con- Potzger saw thousands of square miles where clusively inferred. It is also questionable jack pine constituted at least 75 per cent of the whether the decline of jack pine and the accom- forest cover. One might therefore readily panying increase of fir at the 22-foot level in

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 286 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN

the Hartford profile is significant in terms of According to Suess (1954) radiocarbon dates climatic deterioration. Another possible indi- from deposits in Steuben County, Indiana, cation of slight climatic deterioration can be interpreted as late Gary by W. J. Wayne of the read into the decline of oak and increase in pine Indiana Geological Survey, permit a late Gary for the upper 2 feet of the Hartford profile. The location of the Hartford bog between TABLE 4.—CLIMATIC CHANGE INTERRED PROM FOREST HISTORY IN SOUTHWESTERN MICHIGAN the Valparaiso and Lake Border moraines (Leverett and Taylor, 1915, PL VII) makes it Years before Forest cover Climate possible for deposition to have started in this present basin after the ice retreated from the Valparaiso position. Indeed, it appears that the rock flour Oak-pine Deterioration underlying the peat in the Hartford bog repre- 3500- Expanding oak-hem- Warmest and driest sents deposition while the ice front was not far 4000 lock-broadleaved since retreat of away, perhaps at the Lake Border position. forest ice At any rate, it is safe to infer that pollen rain 5000 Decline of pine period Continued warm- began in the Hartford bog sometime during the and ascendency of ing early retreat of Gary ice (Lake Michigan Lobe) oak and chestnut and has continued ever since. 6000 Pine maximum, near Warming climate elimination of Therefore, the Hartford pollen profile spruce and fir readily indicates forest changes during deterio- 8000 Decline of spruce-fir, Moderating cli- rating climate which theoretically ushered in increase of pine mate Valders time. While there is a marked increase 11000 Spruce-fir Cool to cold, moist of fir and black spruce with decline of jack pine at the 23- to 22-foot levels, there is no date of 12,000-13,000 years before the present. replacement between coniferous and broad- This allows only 1000-2000 years for the ice leaved genera, nor is there replacement of pine margin to melt back and readvance to the by spruce-fir over several sampling levels. Valders position (Fig. 4) and thus supports the However, deteriorating climate can well be idea that some of the glacial substages devel- read into the forest change just referred to, oped without causing pronounced changes in for then as now in northern Quebec the forest forest cover. belts must have been wide in latitudinal extent. A similar situation is evident in pollen profiles Summary of Pollen Studies at South Haven and from Tazewell and Gary regions in Indiana Hartford Bog (Potzger, 1946). In bogs from areas where the (1) The South Haven forest succession ap- surface drift is Tazewell, no replacement occurs pears to have continued uninterruptedly from between broad-leaved and coniferous genera Bowmanville-Two Creeks time until deposition to mark initiation of Gary time, nor is there of peat was abruptly ended by eolian activity anything in the profile to indicate a warming of just before the beginning of Nipissing time. the climate during the Tazewell-Cary inter- Seventy-five inches of sand within thin organic stadial. The profiles show one continuous bands beneath the 30-inch peat layer accumu- prominent spruce-fir period, and this is con- lated during a prominent spruce-fir period. The sistently spread over more foot levels in the sand represents lacustrine deposition in the Tazewell than in Gary regions. This raises the shallow waters of the Calumet and Toleston question as to how far northward ice had re- stages of Lake Chicago, and in Lake Algonquin treated during interstadial periods. If the The 30-inch peat layer represents 4000 years of South Haven profile began in Valders or early forest history beginning with the close of the post-Valders time and the Hartford bog in spruce-fir period and ending with the beginning early post-Gary time, climate could not have of the Climatic Optimum (Xerothermic), or moderated greatly during the Two Creeks oak-pine-broad-leaved forest. interstadial as expressed by the high spruce-fir (2) The correlation of geologic events with abundance in lower levels at both stations. the forest changes in the South Haven latitude

is given in Figure 4. Inference of climatic United States and southeastern Canada: Geol. Soc. America Bull., v. 64, p. 897-919 changes from forest history correlated with time Geisler, Florence, 1935, A new method of separating is shown in Table 4. fossil pollen from peat: Butler Univ. Bot. Studies, v. 3, p. 140-141 (3) The Hartford bog records continuous Goldthwait, J. W., 1906, Correlation of the raised deposition since Gary ice retreated from the beaches on the west side of Lake Michigan: Valparaiso moraine in southwestern Michigan. Jour. Geology, v. 14, p. 411-424 1907, The abandoned shorelines of eastern The 32- to 22-foot levels show a spruce-fir Wisconsin: Wis. Geol. Nat. History Survey, sci. maximum comparable to the 75 inches of sand ser., Bull. 17, 134 p. (106- to 31-inch levels) in the South Haven pro- 1908, A reconstruction of water planes of the extinct glacial lakes in the Michigan Basin: file. The slight decline of spruce with increase of Jour. Geology, v. 16, p. 459-476 jack pine at the 23-foot level of the Hartford Hough, J. L., 1952, Postglacial low-water stage of Lake Michigan indicated by bottom sediments profile, followed by a decided decrease of jack (Abstract): Geol. Soc. America Bull., v. 63, p. pine and increase of fir at the 22-foot level, may 1265. indicate initiation of Valders time, although 1953, Pleistocene chronology of the Great Lakes region: mimeographed report, Urbana, this cannot be definitely stated. The 21- to 10- Illinois, Office of Naval Research, Contract No. foot levels of the Hartford profile are correlated NGori-07133, Project NR-018-122. with the 30-inch peat layer at South Haven. 1955, Lake Chippewa, a low stage of Lake Michigan indicated by bottom sediments: The forest history is similar except that hem- Geol. Soc. America Bull., v. 66, p. 957-968. lock and chestnut are less important in the Leverett, Frank, 1897, The Pleistocene features and deposits of the Chicago area: Chicago Acad. Hartford forest association. This could be due Sci. Bull. 2, 86 p. to local differences in soil factors. — 1899, The Illinois glacial lobe: U. S. Geol. (4) Development of the Xerothermic (cli- Survey Mon. 38, 817 p. • 1929, Moraines and shorelines of the Lake matic optimum) at the Hartford bog (9- to 2- Superior region: U. S. Geol. Survey Prof. Paper foot levels), and later deterioration of climate 154, p. 1-72. (increase of pine) are events of the last 4000 Leverett, Frank, and Taylor, F. B., 1915, The Pleistocene of Indiana and Michigan and the years of this latitude. history of the Great Lakes: U. S. Geol. Survey Mon. 53, p. 316-518. Libby, W. F., 1952, Radiocarbon dating: Chicago, REFERENCES CITED Univ. Chicago Press, 124 p. Lutz, H. J., 1941, The nature and origin of layers Alden, W. C., 1902, Description of the Chicago of fine-textured material in sand dunes: Jour. district: U. S. Geol. Survey Geol. Atlas, Chicago Sed. Petrology, v. 11, p. 105-123 folio no. 81, 14 p. Potzger, J. E., 1946, Phytosociology of the primeval 1906, Description of the Milwaukee quad- forest in central-northern Wisconsin and upper rangle, Wisconsin: TJ. S. Geol. Survey Geol. Michigan, and a brief post-glacial history of the Atlas, Milwaukee folio no. 140, 12 p. lake forest formation: Ecol. Mon., v. 16, p. 1918, The Quaternary geology of southeastern 211-250 Wisconsin: U. S. Geol. Survey Prof. Paper 106, 1953, Nineteen bogs from southern Quebec: 356 p. Canad. Jour. Botany, v. 31, p. 383^01 Antevs, Ernst, 1953, Geochronology of the deglacial Potzger, J. E., and Courtemanche, Albert, 1954, and neothermal ages: Jour. Geology, v. 61, p. Bog and lake studies on the Laurentian shield 195-230 in Mont Tremblant Park, Quebec: Canad. 1954, Geochronology of the deglacial and neo- Jour. Botany, v. 32, p. 549-560 thermal ages: a reply: Jour. Geology, v. 62, p. Sharp, R. P., 1953, Shorelines of the glacial Great 516-520 Lakes in Cook County, Minnesota: Am. Jour. 1955, Varve and radiocarbon chronologies Sci., v. 251, p. 109-139 appraised by pollen data: Jour. Geology, v. 63, Stanley, George, 1936, Lower Algonquin beaches p. 495-499 of Penetanguishene Peninsula: Geol. Soc. Baker, F. C., 1910, Preliminary note on the life of Soc. America Bull., v. 47, p. 1933-1955. Glacial Lake Chicago: Science, n. ser., v. 31, p. 1937, Lower Algonquin beaches of Cape Rich, 715 Georgian Bay: Geol. Soc. America Bull., v. 48, Bretz, J H., 1951, The stages of Lake Chicago:'their p. 1665-1686 1938, The submerged valley through Mackinac causes and correlation: Am. Jour. Sci., v. 249, Straits: Jour. Geology, v. 46, p. 966-974 p. 401-429. Suess, H. F., 1954, U. S. Geological Survey radio- Chamberlin, T. C., 1877, Geology of Wisconsin: carbon dates I: Science, v. 120, p. 467-473 Survey of 1873-77, v. 2, p. 219-228. Tague, G. C., 1946, The post-glacial geology of the Chapman, L. J., 1953, Outlet of Lake Algonquin at Grand Marais embayment in Berrien County, Fossmill, Ontario, Canada (Abstract): Geol. Michigan: In Occasional papers for 1946 on the Soc. America Bull., v. 64, p. 1407 geology of Michigan, Mich. Dept. Conserva- Flint, R. F., 1953, Probable Wisconsin substages tion, Pub. 45, Geol. ser. 38, p. 82 and late Wisconsin events in northeastern Taylor, Frank, 1894, A reconnaissance of the ancient

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021 288 ZUMBERGE AND POTZGER—LAKE MICHIGAN BASIN shorelines of Green Bay: Am. Geologist, v. 8, Wright, G. F., 1918, Explanation of the abandoned p. 315-327 beaches about the south end of Lake Michigan: Thwaites, F. T., 1943, Pleistocene of part of north- Geol. Soc. America Bull., v. 29, p. 235-244 eastern Wisconsin: Geol. Soc. America Bull., v. 54, p. 87-144 DEPARTMENT OF GEOLOGY, UNIVERSITY OF MICH- Wilson, L. R., 1932, The Two Creeks forest bed, IGAN, ANN ARBOR, MICH. Manitowoc County, Wisconsin: Wis. Acad. MANUSCRIPT RECEIVED BY THE SECRETARY OF THE Sci. Trans., v. 27, p. 31-46 SOCIETY, MARCH 2,1955

Downloaded from http://pubs.geoscienceworld.org/gsa/gsabulletin/article-pdf/67/3/271/3416636/i0016-7606-67-3-271.pdf by guest on 02 October 2021