NATURAL SELECTION 207 



less variable in 6/8. The reduction of variability is, 

 of course, assumed to show that ' selection ' has been 

 operative. Thus there is definite selection for ' type,' 

 but very little for ' variability.' In the females there 

 is selection shown both for ' type ' and c variability.' 

 (ii) Imaginal stage. — Ten characters were examined, (a) 

 o* : Selection for ' type ' occurs probably in only 3 

 characters, possibly in 2, and is not shown in the 

 remaining 5. Selection for ' variability ' is certain 

 in 1 character, probable in 2, possible in 4, and 

 absent in 3. (b) ? : In the females selection for 

 ' type ' was certain in 4, probable in 2, possible 

 in 1, and absent in 3. Selection for ' variability ' is 

 reversed (survivors are more variable) in 7, possible 

 in 1, and absent in 2. 



Crampton points out that the actual characters cannot 

 possibly be of service. He thinks that the basis of selection 

 is ' the proper co-ordination of functional and structural 

 elements.' If we understand him correctly, he means that 

 the deviations eliminated are indices of a structural noncon- 

 formity and lack of developmental harmony. This is some- 

 what vague : but the fact remains that survivors and elimi- 

 nated are statistically different (significantly). There are 

 certain ambiguities which require explanation — e.g. why there 

 is selection for variability in the females and not in the males 

 at the pupal stage, and why there is less selection in males at 

 the imaginal stage than at the pupal stage. 



As this work was conducted on rigorous statistical principles 

 and the numbers were fairly high, it is to be accepted as proving 

 that the survivors at each stage differed structurally from the 

 eliminated. The failure to find a basis for selection in the 

 characters studied is not necessarily a limitation. 



(14) Thompson, Bell and Pearson {igi 1) : variation and correlation 

 in Vespa vulgaris. 



These authors undertook a study of the means, variation 

 and correlation of certain wing-characters (dimensions of wings 

 and of individual cells) in the general populations of autumn 

 and spring queens of the Common Wasp. Their object was 

 to study the influence of hibernation on these characters. 



