Sex Ratios in Pigeons. 
497 
Furthermore, a high degree of correlation exists between the time of 
hatching of the two eggs of a set. The data are arranged in the usual 
fashion in Table VI, and the calculated coefficient gives a correlation 
of + 0.728 +_ 0.029. The class interval used—i. e., 12 hours — is rather 
wide, and it is probable that the apparent correlation has been exag¬ 
gerated somewhat for this reason; that there has, however, been no 
sensible selection of the data by choosing day and night periods is 
indicated by a comparison of the means already given with those 
obtained when the records are treated separately for these periods 
as is done in Table VII. 
Table VII .—Mean time of hatching of eggs A and B after laying of B. 
Egg A. 
Egg B. 
Number of 
records. 
Mean 
Number of 
records. 
Mean. 
All records combined.... 
118 
16.42 + 0.05 
118 
16.89 ± 0.04 
Day hatchings (8 A. M- 
8 P. M.). 
45 
16.42 ± 0.07 
45 
17.00 ± 0.05 
Night hatchings (8 P.M.— 
8 A. M.). 
73 
16.42 + 0.06 
73 
16.82 + 0.05 
Although there are in both cases more eggs hatched in the night 
« 
period than in the day* (the numbers happen to be the same for both 
A and B), the mean is unchanged for A, and in B the difference is 
not great enough to affect the results of the correlation seriously. 
We are justified in concluding, therefore, that not only is the mean 
interval between hatching of all the A eggs as compared with B a 
half day, but that for any particular set of two eggs there is a strong 
probability that B will hatch in the next half-day following the 
*This may perhaps be accounted for in part by the fact that sometimes the last day records were 
made at 6 P. M., instead of 8 P. M. (see footnote, p. 495). This does not seem sufficient to account 
for the entire difference, but we have at present no other explanation to offer for the excess of night 
hatchings. 
