388 THE JOURNAL OF GEOLOGY. 



While the ice was retreating and supplying the loess, the land 

 thus uncovered and relieved from the ice weight had been grad- 

 ually rising, until it had attained approximately its present height 

 in Wisconsin, Iowa, and southern Minnesota, before the formation 

 of the moraines. This altitude has endured, excepting minor 



studies to conclude that between the deposition of the early till in southeastern Illinois, 

 with its accompanying loess, and that of the till and attendant paha or eskers of loess 

 in northeastern Iowa, there intervened a very long and diversified history of glacial 

 recessions and re-advances, including at least one prolonged interglacial epoch. A 

 summary of these views in relation to the glacial succession in Ohio is well stated by 

 Mr. Frank Leverett in this Journal of Geology, Vol. i, pages 129-146, Feb.-March, 

 1893. From my early study, "Modified Drift in New Hampshire" (Geol. of N. H., 

 Vol. 3., 1878, chapter i., pp. 3-176, with maps and sections), and from my later work on 

 the Glacial Lake Agassiz, I am strongly impressed with the conviction that the depo- 

 sition and ensuing erosion of the drift, both till and stratified beds, as the loess, went 

 forward very rapidly. What these authors have ascribed to interglacial epochs, one or 

 more of them of great length, seems to me to be more probably referable to geologically 

 very short stages of fluctuation of the mainly waning ice-sheet. 



Professor Salisbury, in the report cited, shows that there were two successive de- 

 posits of till, and a corresponding division of the loess, on and near to the boundaries 

 of the drift ; these seem to me probably due to two closely consecutive stages of ice 

 advance, instead of the long time interval which he thinks to be indicated. Again, in 

 the report on northeastern Iowa, to which reference was given, Mr. McGee clearly 

 shows, chiefly by the forest bed intercalated between two sheets of till, that likewise 

 there the ice advanced twice, with a considerable intervening time, which he supposes 

 to have been far longer than the Postglacial epoch. To my mind, however, the forest- 

 covered borders of the Malaspina glacier or ice-sheet in Alaska leave no doubt that 

 forest beds enclosed in till may be due to oscillations of the ice-front within distances 

 of no more than a few miles or even less than one mile, and that they may have 

 required no longer interval than a few tens of years or at most a century, sufficient for 

 the forest growth, between the times of ice retreat and re-advance. 



When the depression of the ice-loaded land brought it down to so low altitude 

 that the borders of the ice-sheet began to be melted more rapidly than they received 

 increase by snowfall and onflow from the thicker central portion of the ice, a general 

 recession of the glacial margin ensued. On the southern part of the drift in the 

 Mississippi basin no continuous moraines were accumulated, and I attribute their 

 absence principally to the attenuated condition of the ice there and its lack of a steep 

 border. During the glacial retreat, wherever the wavering climate caused the mainly 

 waning ice-border to remain nearly stationary during several years the vigorous outflow 

 of the ice to its then steep frontal slope brought much drift, forming belts of irregular 

 morainic hills and ridges, and leaving many hollows which enclose lakes. The fluctu- 

 ations of the general glacial retreat seem to me to have been due principally to varia- 

 tions of snowfall, some long terms of years having much snow and prevailingly cool 

 temperature, "therefore allowing considerable glacial re-advance, while for the greater 

 part other series of years favored rapid melting and retreat. 



