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

Introduction

About 10 miles west of Madison, in the vicinity of the town of Cross Plains (Figs. 1, 2, 46, 47) is the terminal moraine of the late Woodfordian or Cary ice advance. At Cross Plains and for a number of miles north and south the late Wisconsinan ice sheet ground to a halt on the southwestern Wisconsin uplands, marking the east boundary of the Driftless Area with a young moraine. The moraine, part of the Johnstown Moraine, extends southward in a broad curve through Adams and Sauk counties into Dane County. It follows an irregular looping course across the Baraboo Range and is partly obscured in the Wisconsin River Valley near Sauk City. Its minutely irregular course was controlled by local topography in the deeply dissected Driftless Area in the vicinity of Cross Plains from where it trends south-southeasterly to Verona, Brooklyn, Evansville, and Janesville. Its name comes from its prominent front and abrupt reentrant angle near the town of Johnstown, east of Janesville.

Fig. 46. Part of Wisconsin State Highway map, showing locations of areas 5-7.

Fig. 47. Part of the U.S. Geological Survey Topographic Quadrangle—Cross Plains showing the main Johnstown Moraine (heavy ticked line), some retreatal moraines (lighter ticked lines), a brief advance beyond the main front (light ticked broken line), some glacial lakes (diagonal lines), and outwash (dots).

The deployment of the Johnstown Moraine of the Green Bay Lobe was one of the first major glacial phenomena in the state to be worked out (Fig. 16). The moraine was described by Chamberlin (1883a, 1883b:261-298) as the terminal moraine of the Second Glacial Epoch and cited as the most important discontinuity in the Pleistocene epoch in Wisconsin. Alden (1918) in his detailed reconnaissance of southeastern Wisconsin clearly defined and described the moraine and its associated features. His paper still stands as a model today. Later workers have published information on local areas of the front, but none in the vicinity of Cross Plains.

The Cross Plains area was selected for inclusion in the Reserve in part because it contains a typical portion of the Johnstown Moraine on the uplands and a typical proglacial stream in Black Earth Creek Valley, and is close to a center of population. More importantly it is the only place known to me where the terminal moraine rests directly on well exposed, weathered dolomitic bedrock and where small marginal proglacial lakes, a marginal drainageway, and a subglacial drainageway may all be seen in a small area. Most of the length of the terminal moraine in southern and central Wisconsin fronts on broad outwash plains, in large proglacial lakes, or against older drift. There the relation of the moraine to its adjacent features is clear, but the observer must visit a large area with map in hand to appreciate it. In contrast, the various glacial features associated with the moraine in the vicinity of Cross Plains are more varied and yet as definitive as one could hope to see, all preserved in a neat little package. The area is one of increasing urbanization, and preservation of parts of the front and its associated phenomena can only be assured in the Reserve.

General Description of the Moraine

Alden (1918:212-213) described the distribution and topographic relations of the Johnstown Moraine between Verona and the Wisconsin River as follows:

Crossing an ancient valley at Verona, where it is cut through by Badger Creek, the moraine continues northwest up a second old tributary to Sugar River. For 1-1/2 miles north of the line of the Chicago, Milwaukee & St. Paul Railway the glacier occupied the valley and left its moraine crowded against the west slope, being separated there from by only a sharp narrow ravine 35 to 40 feet in depth. One side of this ravine, which was probably kept open by the southward flow of the glacial waters, is of nearly bare Lower Magnesian limestone [Prairie du Chien, Ordovician, Fig. 3]; the other is formed by the abrupt front of the moraine. Through the next 1-1/2 miles the moraine front rises abruptly 60 to 80 feet from a flat terrace to a well-marked ridge crest, back of which a belt one-half to 2 miles in width, marked by gentle sags and swells and several ponds, extends to a very indefinite inner margin. Near the north line of Verona Township the moraine crosses the old valley and ascends 120 feet to the crest of the Trenton-capped [Platteville, Ordovician, Fig. 3] ridge beyond. Wells indicate thicknesses of 18 to 80 feet for the moraine in Verona Township, the average of 16 measurements being 46 feet. In sec. 5 the moraine is cut through by a narrow ravine 80 to 100 feet in depth, whose lower slopes and bottom expose the St. Peter sandstone and Lower Magnesian limestone.

Outside the moraine near the town line the limestone crest of the ridge is covered only by thin clay soil and scattered boulders which probably came from the moraine or the ice front itself. Here for the first time, tracing it from the southeast, does the moraine reach the Driftless Area and mark the limit of glaciation for this part of the State. For 85 or 90 miles northward from this point no earlier glacier is known to have extended farther west in Wisconsin than the Green Bay Glacier of the later Wisconsin substage; and from this moraine front westward to the Mississippi, a distance of 75 to 80 miles, no unmodified glacial drift has ever been found. The relations of the Green Bay Glacier and its deposits to this thoroughly dissected erosion topography are instructive. For 2 miles in southwestern Middleton Township the ice front lay along the east slope and crest of the rock ridge and deposited its moraine there and in the heads of the ravines which cut the western slope. It is remarkable that no outwash deposits lead down these ravines away from the moraine. Possibly most of the water drained backward down the east slope of the ridge beneath the ice. In sec. 30 the rock ridge swings westward about a mile across the line into Cross Plains Township. The ice did not press forward to the head of the valley thus extended but deposited its moraine across the valley in such a manner as to leave the upper part an enclosed basin 60 to 80 feet in depth. The west front of the morainal dam is bordered by a narrow flat terrace deposited in a temporary lake which occupied the basin. From this terrace the slope rises abruptly 30 to 40 feet to a narrow crest marked by parallel ridges and sharp kettles and many bowlders from which a long gentle slope drops down eastward to an indefinite inner margin. Mr. Voss's well near the top of this slope, at a point about 30 feet lower than the crest, reached rock at a depth of 130 feet.

North of this valley is a high limestone divide between the Sugar River basin and the Black Earth Valley. The rock ridge, before it was covered by the drift, had a relief of 140 to 200 feet on the south and of 300 feet or more on the north. In overriding this ridge obliquely a notch or reentrant three fourths mile in depth was developed in the glacial margin. This indicates that the extreme frontal slope of the ice rose about 200 feet in the first mile from the edge. The moraine ascends the south slope, its crest reaching, in the SE1/4 sec. 18, T. 7 N., R. 8 E., the highest elevation thus far attained, 1,239 feet above sea level. Just how much drift there is at this point is not known, but rock is exposed in the slope and reached in wells 80 to 100 feet lower beneath 20 feet of drift within half a mile. An average of the thicknesses of drift penetrated in the moraine in Middleton Township by 10 wells is about 60 feet. Continuing northwestward the front of the moraine runs for a mile along the crest of the south bluff of the Black Earth Valley as a small marginal ridge 150 to 200 feet above the bottom of the partly filled valley to the north, where it blocks the heads of two ravines. In the NW1/4 sec. 13, T. 7 N., R. 7 E. (Cross Plains Township), it drops down the slope into the valley where, in crossing obliquely, its relief is lost in the general filling of moraine and outwash deposits.

In the NE1/4 sec. 11 a narrow marginal ridge 15 to 20 feet high extends up the north slope of the valley and thence across the heads of ravines and intervening ridges one-fourth mile or less back of the crests of the bluffs, which rise abruptly 100 to 140 feet from the flat floor of the partly filled valley. The surface of this drift is thickly strewn with erratic bowlders, but outside its margin not a piece of drift is found in the thin clay soil. Within the marginal ridge, drift marked by slight sags and swells mantles the rock ridge, but the moraine is not bulky and its inner limit is poorly defined.

The converging of valleys from the east and northeast at Cross Plains led to the ice front crowding forward slightly into the narrow opening between the heads of opposite salients. Here the moraine is pitted with slight sags through a width of a mile and has a relief of 20 to 40 feet above the flat outwash plain to the west. The contrast between the craggy bluffs capped with Lower Magnesian limestone outside the moraine and the smoothly rounded slopes within it is very striking in the vicinity of Cross Plains.

Northwestward from Cross Plains to the Wisconsin River valley, a distance of about 10 miles, the limit of glaciation is generally plainly marked by a narrow marginal ridge with plentiful bowlders. This ridge is a few rods in width and rarely more than 20 feet high, but it is traceable continuously across a greatly dissected topography of five rock ridges 200 to 250 feet in height and four intervening valleys. After leaving the valley at Cross Plains the surface shows, for the most part, only the smooth undulating contours of ground-moraine topography. It looks as though the bulk of the morainal deposit was not formed at the limit of the advance in this part but is represented by the morainal belt which leaves the marginal ridge just north of Cross Plains and thence northward lies about 2 miles farther east.

For a mile north of Cross Plains the ice pressed against the east slope of the rock ridge fitting snuggly into the ravines, as shown by the little marginal ridge which encircles their heads. The crossing of this first ridge causes a reentrant of nearly three-fourths mile, the ice having pushed forward in the next valley. Two to three miles farther northwest a reentrant of one-half mile resulted from overriding a ridge 250 feet high, the ice pressing forward in the valley west of Marxville. So also on the broad low tract near Wisconsin River the ice extended about a mile farther west than on the narrow crest on the south, which rose 300 to 400 feet higher.

The deposit which separates from the marginal ridge north of Cross Plains merges with it again in the broad valley west of Roxbury. The limits of this morainal belt are very indefinite, its presence usually being indicated only by sags and kettles which pit the surface. A mile west of Martinville, however, in adjacent parts of secs. 11, 12, 13, and 14, T. 8 N., R. 7 E. (Berry Township), some of the most strongly marked morainal topography within the region occurs. One bulky ridge which appears to be drift has a relief of 160 feet on the west; and its crest stands 1,249 feet above sea level, or about 400 feet above the rock bottom of the partly filled valley at Marxville, 3 miles to the northwest. The average thickness of drift penetrated by 16 wells in this morainal belt in the towns of Cross Plains and Berry is about 55 feet. These depths range from 25 to 95 feet, several of them not reaching the base of the drift.

The above statement is a clear description of that part of the Johnstown Moraine in the vicinity of Cross Plains. When the conditions and methods of study imposed upon Alden (1918) in his reconnaissance of all southeastern Wisconsin are taken into account, his accomplishments and insight into the Pleistocene geology of the region are nothing short of remarkable. His map, published at a scale of 4 miles to the inch, could not show all the details portrayed by the new topographic quadrangles at about 0.4 mile to the inch and 10-ft contour interval nor by aerial photographs. To this day, only details of the story need be changed. Alden (1918:209-217) clearly recognized that not all of the drift in the Johnstown Moraine was deposited during the one substage, that the thickness of the drift varied markedly from segment to segment of the moraine, and that the outermost front of the Johnstown Moraine was not everywhere synchronous nor representative of equal periods of time. Alden (1918:220-222) also demonstrated that the bulk of the pebbles and stones in the moraine were derived from rocks that crop out in the vicinity and that only 5-20% were derived from Precambrian igneous and metamorphic rocks from northern Wisconsin, Upper Michigan, or Canada. Keewenaw copper nuggets from Upper Michigan and one diamond, presumably from Canada, are among the least common constituents. Dolomite, chert, and sandstone of the local formations (Fig. 3) are most abundant.

Details of the Recommended Areas

Two areas are recommended for acquisition—a northern and southern. The northern area is the north central quarter section of sec. 13, T. 7 N., R. 7 E. (Fig. 47, 48) and the southern contains portions of secs. 4 and 9, T. 6 N., R. 8 E. (Fig. 47). Several possible waysides can be found in Fig. 47, where excellent views of the Johnstown Moraine and other features may be seen. The northern area (Figs. 47, 48) contains an excellent and typical part of the terminal moraine (Figs. 49, 50), drained proglacial lakes, a marginal drainageway (Fig. 51), a subglacial channel, and weathered dolomitic bedrock with large erratics on it (Figs. 52-55). The southern area contains a representative portion of the terminal moraine breached by a drainageway that followed a preglacial bedrock valley, kettle ponds, and an outwash apron (Fig. 56).

A portion of the Cross Plains topographic map is reproduced in Fig. 47, showing the outer edge of the Johnstown Terminal Moraine, some of the fronts established during retreat (and a position occupied briefly beyond the main front), and some of the marginal lakes and outwash as interpreted by me largely from aerial photographs. The two ponds in secs. 24 and 25 are now separated by Mineral Point Road (County Highway S) (Fig. 57). They are the remnants of a former single proglacial lake that filled the basin to about 1155 ft in elevation. The lowest pass from that basin to the west into the headwaters of the Sugar River is about 1175 ft; no evidence that the former lake ever drained through it has been found. Instead it seems to have drained northward across a bedrock ridge of the Platteville limestone (Ord., formerly Trenton of Alden 1918) at about 1155 ft into the adjacent proglacial lake at the same elevation. That lake was short lived, being held in by ice that only temporarily filled the valley 0.4 mile east of the radio tower in the extreme northeast corner of sec. 24. (The gravel pit shown in Fig. 47, in the SE1/4SW1/4 sec. 18, is actually a small quarry in the Platteville-Galena group). Water from the two lakes to the south flowed northwestward marginal to the ice from the vicinity of that pit, past an outcrop of the St. Peter sandstone on the southwest side of the valley, into another small proglacial lake in sec. 13, at an elevation of about 1090 ft. The terminal moraine lies on the northeast side of that valley, although large foreign erratic boulders may be seen to the southwest of the intermittent stream. Water from the large proglacial lake in sec. 13 briefly flowed across the bedrock spur shown in the center of Fig. 48, through the drainageway indicated (Fig. 51) leaving bare weathered dolomite of the Prairie du Chien Group exposed in ridges (Figs. 53, 54) between bifurcating distributaries as the water plunged westward from the steep face. The bare dolomite is solution etched into bizarre forms (Figs. 54, 55). Large foreign erratic boulders are scattered on the dolomite (Figs. 52, 53). That drainageway apparently was occupied only for a short time by the overflowing lake waters which soon began flowing down the Wilkie Gorge and under the ice. Water from about 2.3 miles along the front of the Cary glacier thus flowed northward along the front, from one proglacial lake to another, until finally cascading to the lowland. At first it flowed into the small drained lake basin shown in the western part of Fig. 48 and thence along the margin of the ice in Black Earth Creek Valley (Fig. 47). Shortly thereafter it cascaded down Wilkie Gorge beneath the ice.

Fig. 48. North-central part of sec. 13, T. 7 N., R. 7 E., after U. S. Geological Survey Topgographic Quadrangle—Middleton.

Fig. 49. End moraine on upland in southeast corner of Fig. 48, looking westward. An outwash area an drained lake are in the extreme left part.

Fig. 50. Cary end moraine in center blocking a small drained lake basin behind it in the west-central lowland of Fig. 48. View south-southeastward.

Fig. 51. View northeastward of drainageway in center of Fig. 48. Prairie du Chien dolomite crops out in foreground; the Johnstown End Moraine blankets the opposite side of the channel.

Fig. 52. Basic igneous rock erratic resting on Prairie du Chien dolomite at west edge of Wilkie Gorge, about where end moraine crosses (Fig. 48). View eastward.

Fig. 53. Basic igneous rock erratic resting on eroded Prairie du Chien dolomite with weathered dolomite in the ridge behind and a distributary channel of the drainageway (Fig. 48) on the right. View eastward.

Fig. 54. Detail of Prairie du Chien dolomite ridge in Fig. 53.

Fig. 55. Solution effects of Prairie du Chien dolomite on top of ridge between distributary channels west of artificial fill (Fig. 48).

Fig. 56. Breached end moraine in background and outwash apron in foreground, looking east-southeastward from possible wayside shown south of County Highway PD in sec. 9 (Fig. 47).

Fig. 57. Proglacial pond on north side of County Highway S as viewed from the Johnstown Terminal Moraine (Fig. 47).

The amount of material deposited directly by the ice in this part of the Johnstown Moraine varies markedly from point to point. Alden (1918:218) records the log of a well at the home of Mr. Voss (NW1/4 sec. 30, T. 7 N., R. 8 E.) believed to be that which is 0.2 mile east of the road junction at 1166 ft on County Highway S east of the two ponds. The well penetrated 75 ft of clay and 55 ft of sand and gravel on top of the St. Peter sandstone. However, I found dolomite just 2 ft below the surface of the gently dipping slope of the outwash apron of the Johnstown Moraine, in the extreme southeast corner of sec. 24, and dolomite crops out 0.50 mile north, in the same ridge. Thus the thickness of till in the moraine seems to be no more than 40 ft at its crest which lies on the westerly rim of a preglacial valley. The moraine is even thinner to north and south from County Highway S. The amount of fill in the basin of the proglacial lake bisected by County Highway S probably is several tens of feet although no subsurface exploration has been attempted. Erratics have been found on the west side of the basin in the vicinity of the farm house. Whether carried (here by ice rafting or by glacial ice directly is not known.

The gully crossing the drained lake basin in the southeast comer of Fig. 48 exposes 8 ft of silt on 7 ft of clean, poorly sorted sand and gravel. The base of the section was not seen. The upper silt resembles loess but contains more clay and sand and is believed to have been deposited in the former proglacial lake on top of deltaic sediments and outwash. The axis of the former valley occupied by the proglacial lake in the south part of sec. 13 lies to the east of Wilkie Gorge, about in the position of the town road that descends to the north along the east margin of Fig. 48. That axis is choked with glacial debris. Wilkie Gorge exposes the Prairie du Chien dolomite (formerly Lower Magnesian limestone of Alden 1918) up to the vicinity of the town road crossing the southern part of Fig. 48.

The end moraine to the east of the gorge is only about 20 ft thick. The moraine north of the drainageway (Fig. 48) is the same order of thickness. The upper part of the Prairie du Chien dolomite, as indicated by oolitic chert and sandstone layers in dolomite, crops out on the south side of the drainageway up to 1080 ft, and locally on the flanks of the spur at about the same elevation to the north.

Black Earth Creek Valley contains many tens of feet of glacial outwash (Dury 1964:11) whose bottom has not been reached in the vicinity of Cross Plains. The gravel pit operations 1 mile southeast of town expose at least 50 ft of coarse gravelly outwash. This outwash built up in the valley choking the mouths of tributaries downstream and forming lakes or swamps in them. Whether it is all late Woodfordian in age is not known.

The late Woodfordian or Cary ice quickly retreated slightly from its maximum position in several places north and south of Cross Plains. Only a few of the retreatal moraines are indicated in Fig. 47. One is crossed in the valley by the town road that trends north along the east margin of sec. 13 (Fig. 48). It forms a conspicuous ridge trending northwesterly on top of and west from the prominent bedrock-supported "island" in Black Earth Creek Valley, 3 miles southeast of Cross Plains. That "island" formed a distinct barrier to ice flow and to the later melt water which passed westward both on the north and on the south. Small kettles with ponds are conspicuous on its north side. Many tens of feet of lacustrine sediments are found in the south channel. Loess on top of outwash and on the moraine in Black Earth Creek Valley obviously means that part at least postdates the withdrawal of the Cary ice from its extreme position at the Johnstown Moraine (Dury 1964:13).

Glacial Lake Middleton (Alden 1918, Fig. 47) has had a long and complex history which is not clearly understood. Other shorter-lived lakes occupied parts of Black Earth Creek Valley, of which only one small one is shown in Fig. 47. The complex relationship of the Milton morainic system, thought by Alden (1918) to be a retreatal phase of the Cary, to the Johnstown Moraine will not be discussed here. Even though it affected Black Earth Creek Valley directly, its ice did not reach closer than 2 miles to the Johnstown front. Its effects on the history of the recommended sites are indirect, i.e., through loess formation or climatic modification.

The southern area recommended for inclusion in the Reserve includes the SW1/4 sec. 4 and the northcentral 1/4 sec. 9, T. 6 N., R. 8 E. Because of a bedrock ridge in the northern part of the area, the Cary ice left a prominent, but not thick, moraine with steep outer face, and small kettles in the drift behind the front (Fig. 47). Relief of the kettles is only a few feet. The steep outer face of the moraine rises 100 ft above the bedrock spurs on which it fronts, but only part is drift. A deep drainageway, cutting through the moraine in the center of sec. 5 exposes St. Peter sandstone and Prairie du Chien dolomite below the moraine, as does the marginal drainageway extending southward from it. Bedrock is also exposed in the road-cuts and in the gravel pits in the outwash area shown in Fig. 47. Drilling between the gravel pits disclosed bedrock at a depth of a few feet. Thus the moraine and outwash are only a relatively thin veneer mantling a stream-dissected bedrock topography. The top of the moraine rises over bedrock highs and drops in the preglacial valleys.

Dating of the outwash has not been done. Alden (1918, Pls. 1,111) shows his boundary of older drift encompassing the outwash area and terminating against the Johnstown Moraine at about the center of sec. 5.

One of the deeper preglacial bedrock valleys may be traced from the gap in the moraine, south of County Highway PD, northeastward to Five Points and Morse Pond, and thence northeasterly to the marked steep-walled valley northwest of the WISC radio tower. Dolomite and sandstone of the St. Peter and Platteville-Galena Formations crop out on both sides of the valley which is occupied partly by kettle lakes, the largest being Morse Pond. Typically along the Johnstown front, areas of outwash are localized by preglacial topographic lows in the bedrock. There the glacier could maintain its thickness with a lower surface elevation, which in turn concentrated surface runoff at those points. The highest bedrock ridges have negligible outwash, as is seen between the two recommended areas or between the southern area and Verona where another preglacial valley is found. The one at Verona is larger and topographically lower than the one in the southern area (Fig. 47) and consequently contains far more outwash. The same principle applies to the Rock River Valley at Janesville where still lower topography concentrated many times more runoff and several hundred feet of outwash, not all of which can be attributed to the Cary ice.

The contrast of drift-mantled surfaces behind the Johnstown front with only a thin loess mantle beyond the front can be seen easily from a car. The topographic map (Fig. 47) shows somewhat more irregularity of topography in front of the moraine than behind, but the effect of glaciation close to the front has been one mainly of filling in the lower areas rather than of eroding the higher areas. Drift on many of the uplands behind the front is only a few feet thick, but in valleys it is tens of feet thick. The main valleys in front of the moraine have also been filled with tens of feet of outwash and loess-derived colluvium. Hence, only valley sides and tops of hills beyond the front show the paucity of cover. Erosion by frost action, gravity movements, and surface runoff apparently was far greater during glacial times than now, and temporary features like pinnacles and castelated spires of easily eroded sandstone are fairly commonplace within the Driftless Area. Small sandstone pinnacles may be seen 1 mile west of Pine Bluff, on the south tip of a spur of St. Peter sandstone. A larger pinnacle is in a roadside park on Highway 92, 1.5 miles northwest of Mount Vernon. Devils Chimney 2.3 miles southeast of Mount Vernon (New Glarus quadrangle) is still larger (Fig. 58). Both of these pinnacles are thought to be related to a former glacial lake that occupied the West Branch of Mount Vernon Creek, but the details of these and other pinnacles in the area are beyond the scope of this discussion. It is hoped that some may ultimately find their place in the Reserve.

Fig. 58. Devils Chimney a pinnacle of St. Peter sandstone as viewed northeasterly.

The problem of whether the Driftless Area was ever glaciated is not of immediate or direct concern to the two recommended areas. They provide no evidence either for or against the concept. The solution effects in the Prairie du Chien dolomite could have been accomplished in the 13,000 years the rock presumably has been exposed, but it has not been proved that a longer time was not utilized. I suspect that the solution phenomena have been exhumed by the glacial melt water and were formed during an earlier weathering cycle. The thin young loess cover and lack of residual materials on the bedrock outside the front prove that any older loess and residuum have been removed or were never deposited. The timing of this event cannot be dated except indirectly in other areas. No evidence of older drift outside the Johnstown Moraine between the town line north of Verona and the Wisconsin River has been recognized, yet no old accumulation of residuum nor old loess has been recognized either. Why? Their removal from the hills and completely out of the drainage network as well is not explicable by normal runoff such as is experienced today. This problem will be discussed further in Chapter 11.