94 SECTIONAL ADDRESSES. 



South Africa seems to me the country of all others which could provide 

 the subjects for such an investigation. 



But physiological work of the kind which I have siiggested, although it 

 will show to what extent there are variations between races and species 

 of animals which fit them specially for life under definite physical environ- 

 ments, will not in general elucidate those morphological differences which 

 alone are recognisable in a museum, and which have commonly been 

 assumed to be of an adaptive nature. 



That these structural differences are adaptive even in the sense that, 

 no matter imder what circumstances they arose, they do now in fact fit 

 each form especially to its circumstances, is for the most part pure assump- 

 tion. I do not know a single case in which it has been shown that the 

 differences which separate two races of a mammalian species from one 

 another have the slightest adaptive significance. 



There is no branch of zoology in which assumption has played a greater 

 or evidence a less part than in the study of such presumed adaptations. 



The implication which lies behind any statement that such and such 

 a structure is an adaptation, is that under the existing environmental 

 conditions an individual possessing it has a greater chance of survival 

 than one which has not. 



Mr. G. C. Robson in his book 'The Species Problem,' which includes an 

 invaluable summary of a widespread literature, could only refer to some 

 eighteen papers in which an attempt was made to show by a definite 

 statement of evidence that under natural conditions the death-rate of a 

 population of animals is selective, sparing relatively those individuals 

 which are distinguished from their fellows by the possession of definite 

 structural peculiarities. 



My predecessor. Prof. Weldon, a convinced Darwinian, judged 

 rightly when he devoted many years to an investigation of this funda- 

 mental postulate of the theory of Natural Selection. A ' selective death- 

 rate ' is a term which clearly is only applicable to a population, it has no 

 meaning when applied to an individual ; thus any attempt to determine its 

 incidence and the extent to which groups of individuals possessing definite 

 characters are spared can only be carried out by a statistical method. 

 But it is very difficult to discover cases in which it is possible to collect 

 the data. Any investigation must show as a preliminary that the popula- 

 tion considered is stable and that it is neither added to by immigration nor 

 subject to emigration. The character of a sample must be determined, 

 and in the nature of the case, if for example the character under investiga- 

 tion is the efficiency of a concealing coloration, the sampling error may be 

 large and may be in the same direction as the divergence exhibited by 

 that sample of the population which have died through some external 

 cause. 



Amongst the processes so far investigated only one seems likely to 

 provide at all a general method. This is the study by Dr. Schmidt of 

 Zoarces. He showed that the unborn young extracted from individuals of 

 this fish living at the end of a long fiord did not differ significantly in any 

 of the characters he observed, number of fin rays and of vertebrae for 

 example, from adults taken some years later which could be regarded as 

 having been born in the same season and place. 



