MORAINES OF RECESSION 457 
sibility of drawing the line just there and saying that it could not 
have been 10,000 years or even more. At the present day the 
combined weight of the facts, theories and analogies available 
to us seems to me to lean toward the conclusion that an oscil- 
lation period of between 5000 and 10,000 years would be the 
most satisfactory. If we take the Fort Wayne moraine to be 
five miles wide and the period of oscillation to be 5000 years, 
and if we suppose the period to be divided in two halves equal 
in time, one of advance to Fort Wayne over a space of fifty 
miles and the other of retreat from Fort Wayne over a space of 
100 miles, then we may say in round numbers that it took 700 
or 800 years to build the Fort Wayne moraine. If we take the 
period of oscillation to be 10,000 years, then it took twice as 
long. 
THE SUPERPOSITION OF THE OSCILLATIONS UPON A GREATER AND 
MUCH SLOWER CLIMATIC VARIATION. 
That there was a periodic oscillation of the ice-front of some 
kind or other needs no better proof than the simple fact of the 
existence of the moraine series itself. We have seen that by its 
influence upon climate precession would tend to produce a to and 
fro oscillation of the ice-front. We have seen also that there 
is a considerable amount of reliable evidence showing that the 
oscillations which took place did, in fact, have the character 
ascribed to precession. But if terrestrial climate had suffered no 
change from any other cause than that which produced the oscil- 
lations, then the ice-front would have gone on playing back and 
forth for an indefinite time over the same narrow strip of ground. 
But the moraines are distributed in a great series from south to 
north with intervals between them that are remarkable for their 
regularity, when due allowance is made for the influence of 
topography. This arrangement seems to be explicable only on 
the supposition that besides the oscillations there was another 
greater, slower change of climate in progress—many times 
longer in duration than the period of oscillation. Whatever 
their cause, the oscillations were obviously superposed upon this 
