NATURAL SELECTION 201 



He claims that this shows preferential selection of the 

 unhanded. 



Expressed in this form the figures are worthless from the 

 statistical point of view. His results have been criticised 

 (on the lines already suggested) by the under-mentioned 

 authors (4). 



(4) Haviland and Pitt {1 gig). 



These writers, in addition to a criticism of Trueman's work, 

 supply the results of their own experiments, etc. 



(i) Banded and unhanded snails were tethered to pegs, 

 and the selection by birds was observed. It was 

 found that both types were taken. More of the 

 banded were killed, but the numbers were small. 



(ii) Collections from ' anvils ' were compared with the 

 local population, and it was found that there was no 

 preference as between the banded and unhanded. 



(hi) A captive Thrush was kept under observation and, 

 when offered the two types, exercised no discrimina- 

 tion. 



(i) and (hi) are of little value as evidence. The com- 

 parison of a large series of shells from ' anvils ' with the local 

 population is clearly indicative of no selection. 



(5) Jameson {i8g8) : colour of Mus musculus on sandhills. 



Jameson observed the coat-colour of mice on sandhills on 

 an island in Dublin Bay. There was evidence that the island 

 was about a hundred years old. The mean colour of the mice 

 was lighter than that of the typical form. Thirty-six specimens 

 were examined : of these 5 were as dark as the typical form ; 

 5 were intermediate ; the remaining 26 were distinctly more 

 pallid than the typical form. Jameson states that the island 

 is infested by short-eared owls and hawks, and that these 

 ' most readily capture those mice which contrast most strongly 

 with the sand and arid vegetation.' He does not say that 

 this was actually observed, and there is no statement as to what 

 types were actually seen captured. As there is no direct 

 evidence that one type is captured in preference to another, 

 this case cannot rank as one of direct evidence. 



