208 THE VARIATION OF ANIMALS IN NATURE 



They found (p. 6) that certain linear measurements of the 

 autumn queens are on the average 10-12 per cent, and certain 

 indices 18-22 per cent, more variable than in the spring 

 queens. They also found that there is a slightly higher 

 correlation between the parts of the wing in the spring, as 

 opposed to the autumn queens. According to the principle 

 that selection increases correlation, they argue that ' the only 

 reasonable assumption to make is that there has been a direct 

 selection of correlation as well as selection round a type ' (p. 4). 



We assume that the authors infer that these differences in 

 variability and correlation were due to some selective agency 

 at work during the winter. What that agency was they do 

 not discuss. They say (p. 6) that the ' fitness for survival of 

 the queen during the period in which she is seeking winter 

 quarters, hibernating and starting to form a new colony, 

 seems to depend more considerably on the ratio of the parts 

 of the wing than on their absolute size.' The only further 

 light cast on this matter is the authors' analogy (I.e.) between 

 the wing of an insect and the parts of an aeroplane, reliability 

 in the latter being due to minute details comparable to those 

 of the insect wing ! 



This case is similar to that of Philosamia (p. 206), and 

 we should rather expect that the cell-characters of the wing 

 were correlated with some physiological character determining 

 survival rather than that it was of actual utility. We are 

 somewhat doubtful as to the value of inferences based merely 

 on the reduction of variability. To assign the latter to selection 

 on purely theoretical grounds seems to us dangerous, and we 

 think other causes reducing variability might be operative. 

 There is no proof that the characters ' selected ' are heritable. 



(15) Kellogg and Bell [1904) : observations on the variation of 

 various species of insects. 



The authors point out that the variation in various insects, 

 in spite of exposure for a season to all kinds of rigorous external 

 factors, is just as great as at the beginning of the season, and 

 none of the types of variation is eliminated. This is very 

 well seen in the ladybird (Hippodamia convergens) and in the 

 Honey Bee (Apis mellifera). 



' Determinate variation ' (i.e. statistical change in the 

 constitution of the population) is seen in the pattern of the 



