40 
As the flood recedes, relict pools on the bottom-lands and along 
the margins of the permanent backwaters are formed, in which the 
conditions favoring sporulation or other means of providing for 
resuscitation are to be found. The emerging bottom-lands thus be- 
come the seed-bed for starting a new cycle of diatoms whenever flood 
conditions return. Jn the river, on the other hand, the conditions 
for sporulation are not so favorable, and the current tends to carry 
away such resting stages as may be formed. The observed facts 
regarding the distribution of diatoms and the examination of the 
conditions under which these pulses occur thus alike yield corrob- 
oration of the view that floods are potent factors in determining the 
occurrence of diatoms in fluviatile waters, especially where backwaters 
are extensive. 
The nature of the action of floods is in some respects similar to 
that of the overturning of the water which occurs in lakes when the 
point of maximum density, 39.2°, is passed in either direction. Inlakes 
of some depth the vertical circulation of so large a volume of water 
results in a stirring up of the bottom deposits containing the resting 
stages of diatoms, so that they are brought again into increased 
light and to better aeration. Whipple (94) has emphasized the 
importance of this overturning in starting the growth of diatoms. 
In our shallow waters this physical phenomenon is of less impor- 
tance than in the deeper waters of the lake or reservoir. The vol- 
ume in circulation is smaller, though some compensation for this 
may exist in the possibility of repeated overturnings with fluctua- 
tions in temperatures at the critical stage. The existence of cur- 
rents, the movements of fish, and the roiling effect of strong and 
long-continued winds upon our shallow backwaters, combined with 
the fact that much of the seed-bed area of overflow is dry land at 
the time of the autumnal overturning, all serve to minimize the 
effect of this overturning in our waters upon the growth of diatoms 
in the plankton. The spring overturning occurs early in March, 
and in 1896, 1898,and 1899 a slight pulse not exceeding an increase of 
100 per cent. follows the overturning within an interval of a fortnight. 
The vernal pulse is about two months later than the overturning, 
and the relation of this to the overturning does not seem to be inti- 
mate. The autumnal overturning occurs towards the middle or end 
of November, and in 1895, 1896, and 1898 the hiemal pulse of 
December follows close upon it, within two, or at most three, weeks. 
