40 THE VARIATION OF ANIMALS IN NATURE 



were subjected to different temperatures and it was found that 

 in ' warm room ' temperature there was an increase in tail-, 

 foot- and ear-length. The offspring of these were born and 

 reared in normal temperature and had longer tails, ears, and 

 feet than the progeny of animals kept in ' cold room ' tempera- 

 ture. This was found in three out of four lots. In the fourth 

 lot the relations as regards tail and foot were reversed. F 2 

 animals were not studied. For various reasons the experi- 

 ments were not very satisfactory (see Robson, 1928, p. 170). 

 It should be pointed out that similar results were obtained by 

 Przibram (1909). 



Transplantation experiments were undertaken with 

 Peromyscus (1932, p. 27) and it was found that mice trans- 

 planted from one environment to another {e.g. from the 

 Mohave Desert to La Jolla) showed no change over six to eight 

 years and that there was no convergence in transplants of 

 various races under the influence of a common environment. 

 This fact and others (e.g. p. 58, the wide range in an un- 

 modified condition through a diversity of environments of 

 P. maniculatus gambeli) induced Sumner to abandon his belief 

 in the effect of climate in producing subspecific characters, at 

 least over a few generations. 



xviii. Macdougall (1927, 1930). The inheritance of train- 

 ing in rats. 



Macdougall has presented evidence to show that rats 

 trained over twenty-three generations may be definitely modi- 

 fied. The animals had to escape from a tank full of water. 

 They could attempt to escape either at a lighted platform (in 

 which case they received an electric shock) or at an unlighted 

 one (without a shock). There was evidence that the number 

 of mistakes made by the rats before they chose the exit where 

 they did not receive a shock was gradually reduced with each 

 generation. The data are not treated statistically, but seem 

 convincing. They have been criticised by Sonneborn (1931) 

 on various technical grounds, especially that there may have 

 been unconscious selection 1 or that the strength of the shock 

 varied. We believe that Macdougall has made out a good 

 prima facie case, but confirmation is required. Somewhat 



1 But cf. Rhine, J. B., and Macdougall, W., 1933, Brit. J.Psychol. (Gen. Section), 

 24, pp. 2 13-235. (The authors show that in fourteen generations selected adversely 



^4., pp. •£ 13— ^35. ^ i. 11c auuiuis snuw uiai 111 il 

 for ability, marked improvement took place.) 



