Geology of Ice Age National Scientific Reserve of Wisconsin
NPS Scientific Monograph No. 2
NPS Logo

Discussion of the Pleistocene of Wisconsin

For our purposes with respect to specific areas recommended for inclusion in the Reserve, we can do little with the history of events that preceded the Wisconsinan Stage. When our partial record is compared with the record of events in Illinois (Frye et al. 1965:43-61) for the Pleistocene as a whole and for the early Wisconsinan specifically, the differences are striking. We simply do not have recorded or have not recognized many events well displayed in Illinois. It seems likely that many fluctuations of ice margins in Illinois did not take place in Wisconsin or were so slight as to be inseparable; in, other instances events recorded only by deposits had the record erased by subsequent glaciations. As Wisconsin does not have the complete record, we must turn to adjoining states for theirs in order to work out a composite sequence to which Wisconsin's record can be fitted. This has been done frequently since the concept of former widespread glaciation was established many decades ago. Unfortunately, in spite of the multiplicity of chronologies proposed, we cannot yet agree on a standard one for the Upper Mississippi Valley (Black and Reed 1965). The official chronology for Illinois by Frye and Willman (1960) (Table 1) differs markedly from the classical or standard rendition as recently presented in detail by Leighton (1960) (Table 1). The former attempts to bring order out of a proliferation of names and to correct discrepancies that developed as new data were acquired. The latter retains the classical names, introduces some more, and perpetuates some correlations and usages that should be changed. Names from both have been used as needed for historical reasons or emphasis on the local situation. For additional comments on those classifications see Wright (1964).

Part of the difficulty of correlating events from one state to another is our inability to trace deposits or other indicators of events from one point to another across major breaks. We must then turn to other means of correlation such as similarity of features, materials, or pattern of events or to precise means of dating. Fossils are locally helpful, as in loess deposits, and widespread paleosols are traditionally one of the most important, but Wisconsin lacks them in any mappable deposits of consequence. Radioactive carbon has been especially useful in recent years (Libby 1961). Although not without problems, results from carbon dating can be duplicated and generally are consistent when newer techniques are used. Surely it has done more in recent years to point out discrepancies in our former chronologies and to place a truer time span on events than was remotely possible before its development.

When we examine the dates now available from Wisconsin, we see they fall into natural groups (Table 2) that may be interpreted in different ways. For example, of the four available dates greater than 33,000 radiocarbon years, the two from the vicinity of Marshfield in Wood County are of finely disseminated organic matter in silty clay on bedrock and beneath a single drift sheet that is surely Wisconsinan in age (Hole 1943). The date of more than 45,000 radiocarbon years may be interpreted to mean that the fluctuations of the Altonian ice in Illinois (Table 1) (Frye et al. 1965:43-61) were not represented in central Wisconsin from the time of existence of the pond to the advance of the ice that left the overlying till. The same interpretation is possible for the situation in St. Croix County. There the basal till with erratic wood dated at 29,000 and 30,650 radiocarbon years also seems to have incorporated peat in former ponds that is dated at greater than 45,000 radiocarbon years. The wood is thought to date the time the ice advance destroyed the spruce forest of the area; the older peat which now overlies the younger wood is thought to represent pond fillings on the surface overrun by the ice. If the different kinds of organic matter were transported by the ice only once, this would imply the area was ice-free from more than 45,000 years ago until about 31,000 years B.P. Obviously other interpretations are possible, pointing up the kinds of situations one encounters. Nonetheless, a similar situation in the same general time span seems to have existed in Ontario (Dreimanis and Vogel 1965:782-791) but apparently not in Illinois (Kempton 1963). Similarly, the spruce and willow fragments from Polk County are more than 38,000 radiocarbon years, but they tell us little about the chronology of the area. The 175-180 ft of drift overlying is so poorly recorded in the well records that almost any interpretation is possible.

Table 2. Some radiocarbon dates from Wisconsin. a

Age Laboratory
County Location Remarks

4800 ± 150L-605Bayfield46°57'N 91°00'WSand Island. Compressed peat under 14 ft of sand in water depth of 40 ft. Equivalent to C-504 at 3656 yrs. Dates low water stage of Lake Superior.
4880 ± 190Y-238Marinette47°5'N 87°38'WSewer trench, 600 ft or less in elevation. White oak. Dates maximum or recession of Lake Nipissing.
6040 ± 350 W-1139ColumbiaSW1/4SW1/4 sec. 7, T. 12 N., R. 8 E. Driftwood at 7 ft in Wisconsin River alluvium. With W-1138, dates rapid filling of alluvium of the Wisconsin River.
6070 ± 320 W-1138ColumbiaSW1/4SE1/4 sec. 7, T. 12 N., R. 8 E. Beaver-cut stump in situ at 20 ft in Wisconsin River alluvium. See W-1139.
6340 ± 300 W-1017Kenosha42°33'N 87°49'W Red oak. Dates paleosol at present lake level, under dune sand.
7650 ± 250 SM-16Vilas46°9'N 89°37'W Grassy Lake. Yellow-brown-black gyttja in south center in 6-7 ft of water; lower part of 30 ft of sediment on sand. Total organic content is minimal date for organic accumulation in lake.
10,400 ± 600M-343
Type areaInner part of log.
10,700 ± 600M-342
Type areaOuter part of log.
10,877 ± 740C-308
Type areaSpruce.
11,097 ± 600C-366
Type area Peat from soil horizon.
11,130 ± 350Y-227
Type area Spruce from soil horizon. Equivalent to C-308, 365, 366, 536, 537, and W-42 and 83.
11,350 ± 120W-42
Type area Wood from soil horizon. Equivalent to Y-141.
11,410 ± 180W-83
Type area Wood from soil horizon. Equivalent to Y-227.
11,437 ± 770C-365
Type area Spruce root from soil horizon.
11,442 ± 640C-537
Type area Peat from soil horizon.
11,500 ± 300W-698
Type area Wood from soil horizon.
Type area Log in sediment below soil. Average of cellulose and lignin.
11,850 ± 100L-607A
Type area Wood in soil horizon. Cellulose age.
12,000 ± 400A-79B
Type area Wood.
12,150 ± 400A-79A
Type area Wood from soil horizon.
12,168 ± 1500C-536
Type area Spruce from soil horizon.
12,200 ± 400W-670
Type area Wood from soil horizon.

10,676 ± 750C-630
At Kimberly Clark Paper Mill.Tree stump at depth of 10 ft in varved deposits 25 ft thick, formed in front of Valders ice. Reworked.
10,700 ± 210Tx-44
SE1/4 sec. 28, T. 21 N., R. 17 E.Spruce at depth of 14 ft below Lake Oskosh sediments. Equivalent to C-800.
10,856 ± 410C-800
SE1/4, sec. 28, T. 21 N., R. 17 E.Spruce at depth of 14 ft below Lake Oskosh sediments in Valders till. Reworked (average of 10,241 ± 650 and 11,471 ± 500).
11,640 ± 350W-1110
SW1/4 SE1/4 sec. 19, T. 23 N., R. 19 E.Tamarack in soil horizon.
Menasha. 44°12'N 88°27'W Wood in till. Average of cellulose and lignin.
44°17'N 88°25'W Spruce in Valders till. Average of cellulose and lignin.
NE1/4SE1/4 sec. 19, T. 23 N., R. 19 E.Spruce in soil horizon. Average of cellulose and lignin.
SW1/4 SE1/4 sec. 19, T. 23 N., R. 19 E.Spruce in soil horizon. Average of cellulose and lignin.

11,140 ± 300W-590
sec. 23, T. 24 N., R. 21 E.Wood in Valders till.
11,940 ± 390Y-147X
SW1/4 SW1/4 sec. 23, T. 24 N., R. 21 E.Wood in Valders till.

11,280 ± 100Y-488
44°14'N 88°27'WWood from varved clay of Greater Lake Oshkosh. Equivalent to C-419 at 6401 ± 230.
11,690 ± 370Y-237
NE1/4 SW1/4 sec. 15, T. 20 N., R. 17 E.Wood in red clay.
12,060 ± 700W-1183
SW1/4 NE1/4 sec. 11, T. 20 N., R. 17 E.Peat and spruce under till.

10,420 ± 300W-820
SE1/4 NE1/4 sec. 31, T. 19 N., R. 10 E.Peat at base of kettle.
11,000 ± 400UCLA-633
NE1/4SE1/4 sec. 31, T. 20 N., R. 10 E.Carbonate in gyttja at base of kettle (organic age estimated at 9900).
11,600 ± 300UCLA-631
NE1/4SE1/4 sec. 31, T. 20 N., R. 10 E.Wood at base of kettle.
12,000 ± 500W-641
SE1/4SW1/4 sec. 25, T. 19 N., R. 13 E.Peat at 2-4 ft under silt in last stage of Later Lake Oshkosh.
12,220 ± 250W-762
NW1/4 SE1/4 sec. 10, T. 18 N., R. 12 E.Peat at 4 ft under lake clays.
13,700 ± 300UCLA-632
NE1/4SE1/4 sec. 31, T. 20 N., R. 10 E.Carbonate in marl, 4 ft above base of kettle (organic = 12,800 ± 400).

Other Locations
11,130 ± 600W-1391JacksonSW1/4SE1/4 sec. 17, T. 22 N., R. 5 W.Wood in meander scar of Trempealeau Valley.
11,611 ± 600M-812SaukRaddatz rockshelter Sk5. 43°21'N 89°56'WCharcoal in fire bed in stratum "R".
12,800 ± 220WIS-48JeffersonNW1/4SW1/4 sec. 9, T. 5 N., R. 15 E.Spruce at base of peat mound that formed shortly after deglaciation.
17,250 ± 300GX-0457GrantNW1/4 NE1/4 sec. 26, T. 5 N., R. 2 W.Caribou bone at base of 7 ft of loess in cave and on sandy gravel. H. Palmer, pers. comm.
19,250 ± 350GrN-3624GrantSE1/4SW1/4 sec. 16, T. 1 N., R. 2 W.Alkali soluble organic matter 3.7 m. below top of loess and 2.8 m. above base.
24,800 ± 1100Gro-21 14GrantSE1/4 SW1/4 sec. 16, T. 1 N., R. 2 W.Bulk soil sample at base of loess.
29,000 ± 900W-903WalworthNW1/4 SE1/4 sec. 1, T. 3 N., R. 16 E.Spruce in overridden outwash.
29,000 ± 1000W-747St. CroixNW1/4 SW1/4 sec. 6, T. 28 N., R. 17W.Spruce in dark-gray clayey till on bedrock.
29,300 ± 700GrN-2907GrantNW1/4 NE1/4 sec. 35, T. 2 N., R. 2 W.Spruce charcoal in paleosol at base of loess, on bedrock.
30,650 ± 1640Y-572St. CroixWoodville. 44°57'N 92°18'WSpruce in dark gray clayey till on bedrock. See W-1758.
30,800 ± 1000W-901WaukeshaNW1/4 SE1/4 sec. 36, T. 7 N., R. 19 E.Spruce in overridden outwash.
31,800 ± 1200W-638WalworthSW1/4 NE1/4 sec. 17, T. 17 N., R. 1 E.Spruce in till. Equivalent to M-936 at > 30,000.
>33,000W-1370WoodNW1/4 sec. 2 T. 25 N., R. 3 E.Organic matter in fine mud below till.
>38,000W-1598PolkSW1/4 SW1/4 sec. 3, T. 36 N., R. 15W.Spruce and willow at depth of 175-180 ft.
>45,000Nucl. Sci. & Eng. Corp.WoodNW1/4 NE1/4 sec. 22, T. 25 N., R. 3 E.Organic matter in fine mud below till. M. T. Beatty and F. D. Hole, pers. comm.
>45,000W-1758St. CroixWoodville. 44°57'N 92°18'W Peat in transported pond filling above Y-572.

aExcludes younger archeologic dates, those samples well above the bottom of lake deposits, and some solid carbon dates of doubtful validity. Letter prefix denotes laboratory where sample was run. See various issues of Radiocarbon for further details on individual samples.

However, the three dates (Table 2) of 29,000, 30,800, and 31,800 radiocarbon years from spruce in drift of Walworth and Waukesha counties, the two comparable dates of spruce from St. Croix County, and the comparable date of spruce of the basal loess in Grant County are believed to represent the time of a brief ice advance, called Rockian by Black (1960, 1962) (Frye et al. 1965:43-61; Black et al. 1965:56-81) that occurred simultaneously from the Des Moines Lobe on the west and from the Lake Michigan Lobe on the east. This time is latest Altonian (Table 1) and is recorded also outside Wisconsin (White and Totten 1965). The wood in St. Croix County is in the basal till which is rich in disseminated organic matter, clay, and residual chert (Black 1959a); the wood in Walworth and Waukesha counties is in oxidized sandy till and overridden gravelly outwash. All the wood is erratic and could have been picked up and transported more than once by ice or water, so clearly other interpretations than mine are possible. Nonetheless, if the dates are correct, the deposits can only be younger—not older.

Rockian ice from the two lobes joined in the center of the state. How far the Rockian ice went into the Driftless Area of southwest Wisconsin is not known for certain; it may have covered all of it. At least the Driftless Area has been glaciated (Black 1960), and disagreement is concerned more with timing—Frye and Willman (Frye et al. 1965:43-61) suggest the igneous erratics in the Driftless Area of northwest Illinois may be Nebraskan in age. Positive evidence of glaciation (Frye et al. 1965:43-61) comes from some fragments of Precambrian igneous and metamorphic rocks and particularly Paleozoic chert and sandstone (Akers 1964) that rest on younger formations. Erratics of sedimentary rocks are especially abundant in the central and northern parts of the area (Akers 1964). Sparse igneous erratics occur in isolated kame-like deposits south of Taylor in the northern part of the area and in fresh gravel on the upland beneath thick loess at Hazel Green. Igneous and metamorphic rock erratics are also found tens of feet above the Wisconsin River, as near Muscoda, and are associated with large blocks of dolomite transported several miles from possible sources. These deposits have structures and sorting typical of ice-contact deposits. Large sand bodies in the Kickapoo River Valley have come off dolomite uplands and have glauconite (complex iron-magnesium silicate) above any known source. Anomalous rubble deposits on the upland (Akers 1964) also have anomalous clay minerals (Akers 1961). Thus, in an area of 10,000 mile2 in southwest Wisconsin, we see an absence or paucity of chert and clay residuum on bedrock, an absence or paucity of loess older than 29,000-30,000 years, and an almost complete absence of older paleosols (Hogan and Beatty 1963).1 Moreover, shale with thin seams of unweathered dolomite (Maquoketa Formation) caps East Blue Mound, with only small fragments of the silicified Niagaran dolomite scattered on the broad flat upland (Black et al. 1965:56-81). This is an incongruous situation as supposed peneplains lie below it. Ice seems the only logical agent capable of removing this material out of the area. These suggest that most of the Driftless Area was covered by the Rockian ice (Black et al. 1965:56-81). A pre-Wisconsinan age for some of the peculiar features or deposits in the area can neither be confirmed nor denied (Akers 1964; Black et al. 1965:56-81; Frye et al. 1965:43-61; Palmquist 1965); it is suggested by weathering phenomena.2

1I have recently found Altonian and Sangamonian paleosols in a small area of loess near Hillsboro.
2See Univ. Wis. Geol. and Nat. Hist. Surv. Info. Circ. No. 15 for more recent information.

Dates within the time of the Farmdalian deglaciation, which is recorded so well in Illinois (Frye et al. 1965:43-61), are limited in Wisconsin to one at the base of the loess of the Driftless Area. It does not record a break in the loess sequence—in fact none of significance from the dated paleosol at the base to the present surface has been found (Glenn et al. 1960:63-83 Hogan and Beatty 1963). Farmdalian and Woodfordian time in Wisconsin was at least partly a time of very cold climates and accompanying permafrost and periglacial phenomena (Black 1964a, 1965). No trace of trees has been found in Wisconsin from Rockian time to Twocreekan time which is dated about 11,000-12,500 years B.P.

Woodfordian time is represented in Wisconsin only by two dates in the Driftless Area. One, of caribou bone, is 17,250 radiocarbon years (H. Palmer pers. comm.). The other is a bulk sample of loess. Their significance and relationship to the prominent late Woodfordian (Cary) front or the chronology of glacial events are not known. Drift of middle and late Woodfordian age makes up the surface of more of the state than any other, yet isochronous boundaries (Alden 1918) at the front or within the drift sheet are exceedingly tenuous. Woodfordian time in Illinois is represented by tens of moraines and numerous radiocarbon dates (Frye et al. 1965:43-61). Clearly the Woodfordian is multiple, that is, it is composed of many pulsations of the ice front some with only limited movement but others with retreats or advances (in Illinois) up to 100 miles. The outermost Cary of presumed late Woodfordian age is not represented everywhere in either Wisconsin or Illinois by the same pulse. Although its border from the Plains to the Atlantic Ocean has been described and mapped for decades as the break between deposits of the First and Second Glacial Epochs (Chamberlin 1878, 1883a, 1883b:261-298), we still have much to learn about it. Without a single radiocarbon date related to the advance of that ice in Wisconsin and few to record its destruction, we have been dependent on morphology of forms and direction indicators to separate pulsations. These are applied with difficulty in many places but generally seem better than lithology or texture of the material involved in any one lobe (Oakes 1960); lithology helps to distinguish major lobes (Anderson 1957).

Post-Cary or latest Woodfordian events which are pre-Twocreekan are much less well known in Wisconsin than elsewhere. Moraines as signed to Mankato and Port Huron in Minnesota and Michigan, for example, are presumed to be present in Wisconsin behind the Cary front. However, the correlation of moraines in Wisconsin with type localities has not been done, and deployment of such ice in the state is conjectural.

Deglaciation of the Woodfordian ice in Wisconsin may be time transgressive—being slightly earlier in the south than in the north. However, the available radiocarbon dates tend to negate this. A peat mound on Cary drift in Jefferson County has spruce at the base dated at 12,800±200 radiocarbon years (Ciolkosz 1965). In Waushara County a date of 12,800±400 years B.P. was obtained on organic matter in marly gyttja (fine compact, organic-rich detritus) 4 ft above the base of undisturbed marsh deposits (Park 1965:8). Spruce at the base of the same deposit and higher on the flank of the kettle was dated at 11,600±300 years B.P. Three other dates on peat in basal pond deposits in Waushara County are 10,420±300, 12,000±500, and 12,220±250 years B. P. One in Winnebago County is 12,060±700 years B. P. The main evidence for the time transgressive deglaciation is morphologic—that is the wide spread evidence of ice stagnation in the north particularly and the more youthful lakes and other features in the north. However, the time difference may be several thousand years for all buried ice to melt out (Black 1962).

The Twocreekan interval is named from Two Creeks, where a buried soil and organic remains were recognized in lacustrine deposits along the exposed bluff of Lake Michigan (Goldthwait 1907). This is the best dated interval in Wisconsin, the latest dates yielding an average of 11,850 years B.P. (Broecker and Farrand 1963). A number of dates (Thwaites and Bertrand 1957) derived by the original solid-carbon method were as much as several thousands of years in error according to reruns by better methods. The general range of Two Creeks time from 11,000 to 12,500 years proposed by Frye and Willman (1960) seems distinctly longer than the interval represented at Two Creeks. There, only an incipient soil profile was formed under trees in which the oldest by tree-ring count was only 142 years (Wilson 1932, 1936). Several other localities in east-central Wisconsin contain the Two Creeks horizon in situ, and logs from it are incorporated in the overlying Valderan till. These also tend to cluster close to 11,800 years B. P. so the span of Twocreekan time in central and northern Wisconsin probably is considerably less than in Illinois. This is to be expected, for deglaciation through several hundred miles of latitude of an ice lobe the size of that which occupied Lake Michigan cannot be accomplished over night.

The land surface of Wisconsin generally was exposed during Twocreekan time and was covered by the subsequent Valderan ice advance only in northeastern Wisconsin (Black 1966). It was named from the Valders quarry where relations were established by Thwaites (1943). Consequently, over most of the state the effects of Twocreekan soil formation and geomorphic processes were merged and obliterated by the same processes that continued down to the present day in all but rare situations. One such is the deposits in the rock shelter under the Natural Bridge in Sauk County (Black 1959b; Wittry 1959). Twocreekan deposits were recognized in the shelter from an interpretation of the geology and confirmed by radiocarbon dating. Man was associated with the shelter, leaving his wood fires for dating purposes. The climate in northeastern Wisconsin at the time was perhaps similar to that of today in northern Minnesota (Roy 1964).

Distribution of the Valderan ice is only now being reevaluated (Black 1966). Whereas it was formerly thought to extend across northern Wisconsin (Leverett 1929, 1932) and correlate with red clayey till in eastern Minnesota, this is clearly incorrect (Wright and Ruhe 1965:29-41). Unfortunately we have no radiocarbon dates in Wisconsin directly reflecting either its rate of advance or retreat. Valderan ice occupied the eastern part of Lake Superior and the northern part of Lake Michigan. Parts of both those lakes must have been open water from the latter part of Woodfordian time to the present. This interpretation differs somewhat from that of Hough (1958), but other differences are also appearing (Bretz 1964, 1966; Hough 1966). Unquestionably the full history of the Great Lakes is complicated and beyond the scope of this book. It seems that Valderan ice advanced to the vicinity of Milwaukee3 after having retreated north of the Straits of Mackinac during Twocreekan time. However, only the Valderan ice of the Lake Michigan Lobe entered the State. It possibly reached its climax about 10,500 years B.P. and by 9500 years B.P. was largely gone from the state.

3This is now being questioned.

The close of the Valderan substage has been accepted as the close of the Pleistocene and assigned arbitrarily at about 5000 years B.P. when sea level returned to approximately its present level (Frye and Willman 1960), but others draw the boundary at 10,000 years ago and elsewhere. Wright (1964) would prefer to use the pollen-zone boundary that marks the end of the boreal spruce forest. He would define the boundary in dated pollen-bearing lake sediments resting directly on Valders drift near the type locality of that drift in Wisconsin. Such studies unfortunately have not been made at the type locality although some have been made nearby (West 1961; Schumacher 1966). Broecker et al. (1960) would use the abrupt change in several climatic indicators at about 11,000 years B.P. as the boundary between the Pleistocene and Recent, but this is obviously complicated because of the subsequent Valderan glaciation in the north-central United States. Clearly we have as little agreement on the termination of the Pleistocene as on its beginning although the differences in time involved are markedly less. Agreement on the time of termination is not likely to be reached immediately. It is further complicated, for example in Canada, by the splitting up of the continental ice sheet into several masses that continued to fluctuate independently of each other (Zoltai 1965) for some time after the Valderan ice disappeared from Wisconsin.

After the Valderan glaciation fluctuating climates are recorded in pollen sequences in Wisconsin (West 1961; Schumacher 1966), in physical changes such as the rapid filling of part of the Wisconsin River Valley near Portage about 6000 radiocarbon years B.P. (Frye et al. 1965:43-61), or in other drainage changes (Dury 1964, 1965; Palmquist 1965). These and other facets like faunal and floral changes are beyond the scope of the present discussion.

The full story of the Pleistocene is a fascinating one, yet so long and involved that a single textbook can no longer do justice to it. We have come a long way since the last summary of the glacial geology of the central states was presented (Alden 1932). The new INQUA volume (Wright and Frey 1965) will stand for years as the best summation for the entire United States yet many portions of it are already outdated.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 1-Apr-2005