60 A Second Cooperative Study of " Vespa Vulgaris " 
(6) Goeffijcients of Correlation of Characters in the General Population and 
Nest Queens. 
The results are given in Table IV. Examining first the index correlations 
of the population queens we see that as in the case of the nest queens they 
are all slight and of small significance. Yet they are now all positive, and do 
TABLE IV. 
Coefficients of Correlation for Characters of Nest and Population Queens. 
Population Queen 
Nest Queen 
Ratio P. Q.IN. Q. 
Characters 
correlated 
Eight Wing 
Left Wing 
Right Wing 
Left Wing 
Eight Wing 
Left Wing 
A and B 
•478 + -038 
•579+^033 
•558 + ^040 
•619 + ^037 
•86 
•94 
A and C 
•917 + -008 
•894+^010 
•838 + -018 
•857 + ^016 
1^09 
1 ^04 
A and E 
•531 ± •035 
•647 + ^029 
•238+^056 
•259f053 
2 •as 
2^50 
G and F 
•413+041 
•396 + ^042 
•164 + ^058 
•051 + ^059 
2^52 
7^76 
G and D 
•672 + -027 
•781 + •oig 
•677+ -032 
•694f031 
•99 
1^13 
B and E 
•506 +^037 
•617 + ^031 
•233 + ^056 
•280 + -055 
2^17 
2^20 
B and F 
•677 ± ^027 
•671 ±^027 
•ISO ±-057 
•139 ±^058 
3-76 
4^83 
Mean 
•599 
•655 
•413 
•414 
1^95 
£•92 
/ and K 
•123+ ^048 
•054 + ^049 
- ^021 + ^059 
•037 ±^059 
M and J 
•050 ±-049 
•137 + ^048 
-•1.38f058 
-•185 + ^057 
L and J 
•097 + ^049 
•132±^048 
-•307 ±-051 
-•186±^057 
I and L 
•137f048 
•203 ±^047 
•039 ± ^059 
•194 ±^057 
Mean 
•102 
•131 
-•122 
-•035 
not take the somewhat erratic values found in the case of the nest queens. It 
may be doubted whether, considering their probable errors, any weight whatever 
can be laid on these correlations. The small variation of these indices and their 
low correlations demonstrate, however, that they might be of considerable service 
for classificatory purposes. If we turn to the correlations of the absolute sizes, 
we see at once a marked difference. With the exception of one pair to be referred 
to especially below, in which the nest correlation is higher for both wings, and 
one pair, where in the right wing they are sensibly equal, the correlation is 
higher, and very markedly higher in the general population, than in the nest 
queens. Now this is precisely what the theory of correlation would lead us to 
expect ; if we select a small element of a total population we thereby reduce the 
correlation. Now let us omit the correlations of length and breadth of wing, 
where clearly some special cause has been at work. We find for the mean corre- 
lation of population queens for the remaining six pairs on both sides "64. The 
