78 BOTANICAL GAZETTE [AUGUST 
and others. A certain degree of aridity seems to have prevailed, 
not only during the withdrawal of the ice, but up to the period 
which resulted in the formation of the Bloomington morainic sys-- 
tem. The Bloomington period of peat formation was stopped by 
the upper seam of clay. This suggested correlation appears to 
be correct, for the upper clay layer can be esniaeiete closely with 
that part of the Valparaiso-Kalamazoo morainic 
system which passes northeast of Canton through Portage County. 
The principle members of this group of moraines show west of here 
a marked differentiation of glacial lobes and a shifting of the lines 
of axial ice movements. The glaciers, as shown by the studies 
particularly of LEvERETT and others, encroached again over the 
surface of land that had been vacated by the earlier recession of 
the ice border. This readvance, the limits of which are marked by 
a morainal belt reaching from eastern Illinois and extending north- 
ward into Michigan to the vicinity of Kalamazoo and Battle Creek, 
has usually been designated the late Wisconsin stage. It covers 
a time of drift deposition reaching to the series of generally weak 
moraines which are included in the Lake Border-Defiance system. 
During the time which elapsed while this ice front receded, and 
which may tentatively be called the Mississinawa glacial substage, 
the third tier of macerated peat was formed and probably also 
some of the superposed layers of fibrous plant remains. There is 
hardly any feature in the structure of the Canton deposit so con- 
spicuous as the fibrous layers of peat, which rest on and in places 
grade into the underlying third basal bed of macerated organic 
material. The Carex and Phragmites plant populations, from 
which these layers of relatively coarse fibrous peat are derived, 
appear to have grown at ground water levels much lower than. 
those which prevailed at later glacial substages. The uppermost 
beds of fibrous peat of more recent development contain an admix- 
ture of aquatic plant débris. They do not represent in their texture 
the features which would be characteristic of a gradual decrease 
in available ground water coincident with the closing of water 
basins by vegetation. 
In the absence of more definite correlations, these three primary 
series of peat layers, namely, the several basal layers of macerated 
plant remains, the middle bed of coarsely fibrous peat, and the 
