340 
Ohservatiom on TtTinites 
TABLE VI. 
Termes Natalensis Haviland. 
Correlation of the Means of Small Soldiers and Large Workers from 27 nests.* 
Small Soldiers 
Unit = 0-04mm. 
- 
— - 
rH 
0 
CM 
0 

27 
231—324 
1 
1 
225—228 
2 
2 
239—332 
1 
2 
1 
4 
232—236 
1 
0 
2 
3 
237—240 
1 
1 
2 
241—244 
2 
0 
1 
1 
1 
5 
245—248 
1 
2 
3 
349—353 
0 
253—256 
1 
1 
2 
257—360 
0 
261—264 
1 
1 
265—368 
1 
1 
269—272 
1 
1 
273—276 
0 
377—380 
2 
2 
Large Workers 
Unit = 0-04 mm. 
1^ 
>-i 
1 
©1 
i 
1 
Si 
3i 
1 
1 
^> 
1 
©i 
1 
5>i 
1 
1 
^? 
"o 
i 
0 
Si 
■"O 
®> 
1 
~* 
®! 
©J 
1 
00 
'0 
©( 
n 
(9) Standard Deviation of the Means. 
The standard deviation of the means of the large soldiers for i-andom samples 
taken from 20 nests is 16"90, for small soldiers fi'om 27 nests 15"29, and for 
large workej-s from 27 nests, 13 04. It is thus seen that the means of different 
nests vaiy very considerably, and reference to Table I. will show that this variation 
is not entirely or even mainly due to the fact that the material was not all collected 
at the same time of the year. We have here to do with the effect of inheritance, 
or with that of the slightly different environment of each nest. It will be noticed 
that the standard deviations of the means for the three castes are all of the same 
order of magnitude, although that for the large workers is the lowest. 
(10) Correlation of the Coefficients of Variation. 
The variability of the members of a nest varies in different nests ; thus the 
standard deviation of the coefficients of variation of random samples of small 
soldiers taken from 27 nests is 4'4<2 units, and of large workers 3"40 units. The 
coefficient of correlation between the coefficients of variation of small soldiers and 
large workers in 27 nests is "453; but the probable error is + '103, which is so 
great that it is scarcely safe to conclude that we have this correlation very closely. 
* The smallest number of iudividuals from a nest was 32. 
