4 A Third Cooperative Study of " Vespa Vulgaris" 
The indices determined were : 
H=D/A, 
I = EjD, 
J= G/D, 
K = B/A, 
L = G/A, 
M = F/A. 
The correlations selected for comparison were those of the previous memoirs, i.e. : 
A and D, G and F, I and K, 
A and G, G and D, M and J, 
A and E, B and E, L and J, 
B and F, I and L. 
Further the correlation of all the absolute lengths measured on the two wings 
was made as before. 
(3) On the Means of the three Series. 
Taking the absolute sizes we see a remarkable accordance between the series 
of spring and of autumn queens of the same species from districts some 20 miles 
apart. On the average the autumn queens appear to be slightly larger (T0074 
right wing and 1"0038 left wing), but no stress whatever can be laid on such 
a slight difference, nor on the fact that in the case of two characters the autumn 
series appear to be smaller. It would require far larger numbers than we 
have dealt with to give such differences real significance. In the case of the 
indices also, it is impossible to assert that there is any sensible differentiation 
between the queens of the general population in autumn and in spring. But the 
determination of the characters for the autumn population has been of great value, 
for we conclude that the differentiation in absolute size is not due to (a) local race 
differences — the queens of Gerard's Cross are identical in size with the queens of 
Cbeckendon and Woodcote, nor (c) the selective action of the winter in differ- 
entiating autumn from spring queens. Even one portion of our alternative* 
(b), individuality of the nest arising from heredity or nurture, seems excluded, 
because the general change of environment from Gerard's Cross to Checkendon 
would we suppose be greater than between two nests at any one place. We shall 
be able to test this point on further material in our possession, but it seems 
reasonable to accept as a working hypothesis that the difference between the nest 
queens and the general population queens in absolute size was the result of 
heredity in the individual nest. In the case of the indices there are no average 
differences on which, for our limited data, we should feel inclined to lay the least 
stress. Lastly, we do not find either in absolute measurement or in the indices any 
persistent differences of a sensible character between right and left wing for the 
three series. If the right side is differentiated from the left — as it certainly 
appears to be in the case of bone measurements in man — there are no differences 
of such a marked character that they are sensible on samples of the present size. 
* See Biometrika, Vol. vn. p. 52. 
