NATURAL SELECTION 211 



Finally, if all the birds had been left out of doors, probably 

 all would have died, and the real selective agency was human 

 interference (i.e. the bringing of the birds into the laboratory). 

 It is most unfortunate that Bumpus did not investigate the 

 actual cause of death in each case, and for this reason (coupled, 

 of course, with the actual paucity of individual variates) we 

 hold that the quite clearly established ' significant ' differences 

 are suspect. 



( 1 7) Weldon (igoi) : comparison of earlier and later whorls of 

 the shell of Clausilia laminata. 



A series of measurements of the earlier and later whorls 

 of the shell (made on sections) shows that ' the mean 

 spiral of the young generation is sensibly identical with 

 that of the parental generation [earlier as opposed to later 

 whorls] and is not altered by any process of selective 

 destruction.' 



As, however, the variability of younger shells is greater 

 than that of adults, it is inferred that there is ' periodic selec- 

 tion ' (reduction of variation at each generation). The fact 

 that the mean remains the same is held to be an indication 

 of the effect of selection. 



We are not convinced that if a difference between 

 the early whorls and the later had been shown, it would 

 necessarily imply that the difference was due to selection as 

 Weldon suggests. It seems that any changes that might 

 have been found could have been due to environmental causes. 

 As for the reduction of variability in the adult stage, we 

 think that this might possibly have been due to greater 

 plasticity of the young, as well as to selection. 



(18) Weldon (1904) : shells of Clausilia itala. 



The same type of measurement was undertaken on the 

 shells of 100 young and 100 adult C. itala. No difference 

 between the young and adult shells was found. Weldon 

 suggests that this might be explained in two main ways : either 

 that (1) no selection was operating, or (2) the lack of selection 

 was due to the specimens having been collected in the spring. 

 If measured in the autumn differences might have been 

 shown (?). 



