255 
The Origin of Sphagnum Atolls. 
layer which is impervious to the upward diffusion of the calcium 
carbonate. Hence from the nature of the succession it seems 
unlikely that shade or lime can ever be responsible for the trench 
of an atoll. 
From the observations on occurrence of atolls it has been 
noted that they are numerous in sand areas. In fact none have 
been reported or observed in clay areas. The soil of Crow Wing 
County near Gull Lake is notably deficient in lime, and in such an 
area it is impossible to assume that lime in the run-off water of the 
bank has any effect on the outer zone of a hog. The fact that the 
Sphagnum and heaths of the Hubert atoll join with the vegetation 
of the bank where the fire was absent is proof that there is nothing 
in the run-off water of the bank to interfere with the growth of 
these plants. 
The fact that fire is the cause of the marginal trench has not 
been apparent because the bog assumes its atoll character only 
when the water-level is high enough to obscure or conceal the fire 
scars. Moreover very dry seasons are rare where bogs are formed, 
especially in clay soil areas. Since the water-content of bogs is 
always high it seems almost incredible that they should burn. 
Fire in an undrained bog usually burns very slowly, and the heat 
dries out the surrounding medium which is then consumed. If the 
burning, starting from the edge of the bog, is left undisturbed it 
consumes the organic matter as far down as the bottom of the 
parent pond. Apparently the centripetal progress of the fire is 
checked when it reaches the peat that directly overlies the water. 
When the water rises again to the normal level the trench appears 
and the hog seems mysteriously isolated from the bank. 
The nature of the succession in the trench depends on the 
depth of the water. It is possible for the depth to be sufficient to 
initiate a prisere. However, in the hogs studied the succession of 
the trenches is typically secondary. The Ballard and Anderson 
atolls are now for the most part in the Carex-Calamagrostis 
associes which is being slowly followed by Sphagnum and Atidro- 
meda-Chanice daphne of the prisere. In the deeper places of the 
ditch of the Glenwood atoll the succession is beginning with Typha 
which will probably be followed by a Carex-Calamagrostis associes, 
and this by Androtneda-Chamcedaphne. The outer margin of the 
Hubert atoll is in the Carex-Calamagrostis associes, which in the 
shallower places is already being invaded by representatives of the 
Andromeda-Chamcedaphne associes. 
