386 Birge—The Crustacea of the Plankton. 
be if the deeper water could be utilized. It is not impossible 
also that one factor in determining the small number of the 
periodic species of Crustacea in lake Mendota may be in the fact 
that the upper water is so completely occupied by the perennial 
forms as to leave little chance for the development of other 
species. Fourth, the Crustacea are not excluded from the deeper 
water of the lake by the low temperature of the water, as is 
proved by the occurrence of the same species in the far colder 
water of other lakes in the same district. The exclusion is due 
to the accumulation of the products of decomposition in the 
lower water, which remains entirely stagnant after the thermo- 
cline has been formed and is never exposed to the action of sun 
and air. This water in lake Mendota acquires an offensive smell 
and a disagreeable taste, though in. neither respect does it go 
as far as certain waters mentioned by the Massachusetts Board 
of Health (Drown, ’90, p. 553.) It is always clear and bright 
to the eye. 
The products of decomposition of the algae and Crustacea of 
winter and spring remain stored in the deeper water, and un¬ 
doubtedly the addition of this store of nutritive material to the 
water of the lake as the thermocline gradually moves downward 
is one of the factors which occasions the enormous increase of 
the vegetable plankton in the late summer and autumn. 
Autumn — -October, November , and December. 
The summer conditions of distribution end with the breaking 
of the thermocline and the resulting establishment of the fall 
homothermous period. This occurs at different times in different 
years. The date depends on: First, The rapidity of cooling 
of the surface; Second, The summer temperature of the bottom; 
Third, The amount and direction of the winds, especially of gales. 
In 1895 and 1896, the “turn over” came in the last week of Sep¬ 
tember ; in 1894 the distribution of the Crustacea shows that it did 
not come until the first week of October, and it was equally late 
in 1897. In the year 1894 no observations were made in the first 
half of September, but the distribution in the latter part of 
September of that year closely resembles that in the early part 
