107 
7, 1897, was at 80°. A pulse of 48,000 on November 22 at 40° gives 
evidence of considerable range in adaptation to temperatures. 
In Table I. the seasonal distribution of D. globulosa is given in 
full. It differs from that of previous years mainly in the fact that 
the summer pulses do not here have the amplitude reached in other 
years; for example, in 1896 (252,000) and 1897 (1,240,000). It is 
characterized by considerable irregularity caused by somewhat 
abrupt pulses at irregular intervals. A comparison of these occur- 
rences with the hydrographic conditions (Pt. I., Pl. XII.) indicates 
that in the colder months increase in numbers in the plankton at- 
tends flood waters only, as, for example, in January, February, late 
October, and November. In the summer, pulses may also come 
with floods. For example, that of 252,000 on May 25, 1896, ap- 
peared on the upward slope of the June rise of the year, and that 
of 80,000 on June 28, 1897, came with the belated June rise of that 
year. On the other hand, some of the minor fluctuations appear 
on declining floods, and the maximum one of our records, that of 
Sept. 7, 1897, came in the midst of the most prolonged period of 
ssaple low water (Pt. I., Pl. XI.) found in the six years of our 
operations. From these facts it is evident that floods are efficient 
in increasing the number of D. globulosa in the plankton, and that 
the amplitude of the pulses to which they contribute is much greater 
in the warmer months (above 60°) than in the colder ones—as a 
result, perhaps, of the greater numbers present in their normal 
habitat, the shores and bottom, and also as a result of their readier 
flotation at this season. In so far as their presence is due to floods 
they are adventitious. On the other hand, it is very probable that 
they become temporarily eulimnetic in habit during the summer 
months. The evidence for this lies in their greater numbers in a 
period which is predominantly one of greater stability. Thus in 
1898, in the 22 collections between May 1 and October 1, the average 
number present is 9,731, while in the remaining seven months of 
colder weather the number is only 5,200. Additional evidence arises 
from the fact that pulses of unusual magnitude have occurred quite 
independently of any factor such as flood or other disturbance which 
might cause their adventitious introduction into the plankton. 
Thus on Sept. 7, 1897, there is a symmetrical pulse whose rise and 
decline occupy four weeks, as shown in the following table. The 
total change in river levels in this period of four weeks (Pt. I., Pl. 
