Birge—Work of Wind in Warming a Lake . 351 
meter of the five, there will be six observations to be added 
and divided by six, and each will be employed once in de¬ 
termining the mean. If the mean temperature for each 
meter of the stratum is taken as the mean of the readings 
at its top and bottom, the mean for the five meters will be 
the sum of these means divided by five. But in reaching 
this result four readings of the six have been used twice, 
all, that is, except the readings at the top and the bottom 
of the 5 m. stratum. If the temperature falls rapidly in 
the stratum, the two methods of computations may give 
results that differ considerably, as the following example 
shows. This is taken from Beasley lake. 
TABLE 3—BEASLEY LAKE. 
Depth, m. 
Temp. 
Depth, m. 
Temp. 
5 
21.5° 
5-6 
18.5* 
6 
15.5 
6-7 
13.35 
7 
11.2 
7-8 
10.2 
8 
9.2 
8 9 
8.2 
9 
7.2 
9-10 
6.85 
10 
6.5 
mean 11.85 
- 
mean 11.42 
The value of 1-D for 11.85° is 0.000448 and for 11.42° 
it is 0.000411, so that the value is increased about 9%, by 
the different method of computation. This is an extreme 
case and the difference is usually much less. The following 
example comes from the thermocline of lake Mendota. 
TABLE 4—LAKE MENDOTA. 
Depth, m. 
Temp. 
Depth 
Temp. 
9 
20.3° 
9-10 
18.8° 
10 
17.3 
10-11 
16.3 
11 
15.3 
11-12 
14.95 
12 
14.6 
12-13 
14.35 
13 
14.1 
13-14 
13.9 
14 
13.7 
mean 15.65 
mean 15.9 
In this case computations based on the larger mean give a 
result about 4.2% larger than that derived from the smaller 
number. 
