196 THE VARIATION OF ANIMALS IN NATURE 



The view has been expressed that ' it is impossible to con- 

 ceive that the detailed action of Natural Selection could ever 

 be completely within human knowledge ' (Fisher, 1930, p. 47). 

 The process might nevertheless be brought sufficiently within 

 human observation to provide direct visual proof. Obviously 

 the conditions for observing an act of adaptive transformation 

 are very rarely available for a human observer. The coinci- 

 dence of several propitious circumstances, that is rarely realised, 

 is required : but it will be seen that the opportunity is not so 

 rare as Fisher suggests, and that more efforts should be made 

 by field workers to locate likely situations and bring them to 

 the notice of those able to carry out the necessary observations. 



Many observations and experiments have been made 

 on animals living freely or in captivity which are claimed to 

 prove either the elimination of certain types of variant and the 

 survival of others, or the absence of selective elimination. 

 These studies are not of the same kind. 1 The problems they 

 set out to solve and the procedure adopted are not of the same 

 order, and it is necessary to show at the offset exactly what 

 they aim at demonstrating, before proceeding to detail the 

 results obtained and the criticisms that may be made as to 

 their interpretation. 



(1) In a certain number of cases the observations (with or 



without control experiments) relate to animals living 

 freely and exposed to a known or reasonably assumed 

 cause of death (Weldon, 1899 ; Harrison, 1920 ; 

 Trueman, 191 6 ; Haviland and Pitt, 1919 ; Jameson, 

 1898 ; Kane, 1896). 



(2) In six cases the observations relate to animals either 



subjected to laboratory or other experimental condi- 

 tions or experimentally exposed to natural enemies, the 

 cause of death being known or assumed (di Cesnola, 

 1904, and Beljajeff, 1927; Poulton and Saunders, 

 1899, and Moss, 1933; Boettger, 1931 ; Lutz, 

 1915; Davenport, 1908; Pearl, 191 1). 



(3) In two cases the animals observed were living under 



natural conditions, but the cause of death was un- 

 known (Crampton, 1904 ; Thompson, Bell and 

 Pearson, 191 1). 



1 Studies comparable with some included in i— 19 below are also to be found in 

 our section dealing with Protective Resemblance and Mimicry (pp. 232-265). 



