200 THE VARIATION OF ANIMALS IN NATURE 



wood, while in the birch wood few, if any, birds occur, as the 

 wood is not well grown enough to afford cover. 



This case is very summarily expressed. The number of 

 likely enemies in the two woods is not discussed in detail. 

 It is quite uncertain how the individuals whose remains were 

 found actually met their fate — i.e. whether they were killed by 

 birds or bats. There is no statement as to how many of the 

 15 pairs of wings were pale (? 14 : 1 or 8 : 7). Nevertheless if, as 

 he says, the population of the pine wood is preponderatingly 

 dark, a ' majority ' of light eliminated forms is significant. 



On its surface value this case might pass as definite evidence 

 for selective elimination. It seems to us to be open to two 

 main criticisms : ( 1 ) the lack of definitely expressed evidence 

 as to the frequency of enemies in the two woods, and (2) the 

 small number of observations and the failure to state what is 

 meant by a ' majority,' particularly in regard to the frequency 

 of each variety. 



(3) Trueman {1916) : alleged selection of ' banded shells of 

 Cepea.' 



It has long been known that birds feed on the common 

 snails C. hortensis and C. nemoralis, taking them to stones on 

 which the shells are broken in order that the bodies may be 

 extracted. Masses of shells are often found around these 

 c anvils,' and Woodruffe-Peacock (1909) suggested that it 

 might be possible to detect from the broken shells any selection 

 of a particular type, e.g. as between the banded and unbanded 

 types. Peacock's observations did not include a survey of the 

 percentage occurrence of the various types in the local popula- 

 tion from which the victims were taken, and are therefore 

 useless. 



Trueman compared his shells from ' anvils ' with a standard 

 collection, not a local one, and his conclusions are also value- 

 less, inasmuch as the local percentage of banded and unbanded 

 forms varies very much from district to district. He also fails 

 to give the actual numbers of shells obtained, expressing his 

 results as percentages, of which the following is the essential 

 result : 



' Standard ' collection Found on * anvils ' 



Unbanded 25 per cent. 38 per cent. 



5-banded 42 ,, ,, 23 



53 



