568 
BUI.LETIN OF THE BUREAU OF FISHERIES. 
Inspection of table xii shows that the mean temperature Of the epilimnion and 
the thermocline are not very variable in the six major lakes. In the first region Tm®-4 
equals 15.66°, as the mean of 12 observations, ranging from 15.2° to 16.8°. For the 
thermocline the mean is 9.56°, ranging from 7.8° to 10.6°. The differences in the 
amount of heat stored by a lake in these regions are due much more to the thickness of 
the stratum than to its temperature. For instance, in Seneca Lake in 1910 the tem- 
perature of the thermocline was 19.4°, and in Skaneateles Lake in 1910 it was 19.7°. 
But its thickness in Seneca Lake was 15 meters, and only 9 meters in Skaneateles; and 
the wind-distributed heat in the epilimnion of the former lake was therefore over 75 per 
cent greater than that of the latter. 
It appears from the table that a very large part of the wind-distributed heat is in 
the epilimnion. The upper 10 or 15 meters of a lake, even 60 kilometers long and nearly 
200 meters deep, contain 50 to 70 per cent of the heat, or even more. If to this stratum 
is added that which lies immediately below it and has derived its heat from it, it appears 
that the upper 20 meters contain 70 to 80 per cent, or even 90 per cent, of the wind- 
distributed heat. (See table xin.) This limitation of the heat to the upper strata is 
responsible not only for the sharply-defined thermocline, but also for the general 
uniformity in the amount of wind-distributed heat in the heat budget of the several 
lakes. A large percentage of the heat is always near the surface in summer. During 
the period of light winds and summer weather, when heat is furnished far in excess of 
the capacity of the distributing agents, so much is lost in any case that there is enough 
for any lake to get the maximum possible supply, and any deficiency is likely to be due 
to the distributing agents, and not to lack of supply. Still more, any loss caused 
by a short cool period during the summer may be quickly repaired, since no violent 
wind is needed to distribute the heat to water near the surface. Indeed, such a cool 
period may well be the indirect cause of the gain of heat. It allows the heat already in 
the lake to be distributed to a greater depth, while the surface will rapidly renew its 
supply during the succeeding warm days. 
The heat distributed to the hypolimnion is extremely variable both in quantity 
and in the per cent it constitutes of the total heat. If the sums of the calories found 
in the epilimnion and thermocline are compared for the 12 observations, it will be 
found that the mean is 22,200. The mean departure of each observation from the 
mean is about 8 per cent; the maximum departures are -t- 19 per cent and — 16 per cent; 
and the range is about 35 per cent of the mean. In the case of the hypolimnion the 
mean amount of heat is 4,900 calories. The mean departure of each observation is 
28 per cent of this sum, the maximum is -f-8o per cent and — 50 per cent, with a range 
of 130 per cent of the mean. Not only so, but the difference in the same lake in suc- 
cessive years is even more striking. The hypolimnion of Seneca Lake received 8,800 
calories per square centimeter of surface in 1910, and only 3,250 calories in 1911. Keuka 
Lake had 6,300 calories in 1910 and 2,700 calories in 1911. The mean of the six lakes 
for 1911 was hardly more than half as great as that for 1910 (6,400 and 3,400 calories), 
and the largest amount in 1911 (4,800 calories in Skaneateles Lake) was below the 
