198 COMPARATIVE PHYSIOLOGY 



female respectively make to the constitution of a new animate 

 unit. 



Bearing on Other Branches of Physiology.— For example, 

 many kinds of physiological experiment involve comparisons 

 for which controlled observations can only be carried out on 

 different individuals, e.g. effect of diet or endocrine substances 

 on growth. It is clearly established that growth-phenomena 

 and size-differences in a number of cases depend upon factorial 

 inheritance. Experiments of this kind, therefore, unless 

 based on large numbers of animals (which is often impossible) 

 and subjected to statistical analysis, are of doubtful value, 

 when the material is not known to be genetically homogeneous. 



To achieve this end it is not necessary to set about breeding 

 pure strains with reference to every characteristic it is desired 

 to study. With the aid of the conventional symbols the reader 

 can easily satisfy himself that in a cross involving one pair of 

 factors, the proportion of heterozygotes diminishes generation 

 by generation in a continuously convergent series, if any 

 system of close inbreeding is employed. After about twenty 

 generations of brother and sister mating, for instance, or ten of 

 self-fertilization, the proportion of heterozygotes is indefinitely 

 small. If a stock that is not undergoing mutation is bred from 

 generation to generation by close inbreeding, it must eventually 

 become for all practical purposes homozygous for any charac- 

 teristic. Such stocks of white rats have been reared by the 

 Wistar Institute, and are employed in physiological experi- 

 mentation increasingly by American workers (see East and 

 Jones). 



Again, the following citation, from an important physio- 

 logical memoir, illustrates how easily genetical bias enters into 

 physiological reasoning : — 



" Now the history of the surfaces in the hearts of rays 

 on the one hand and of the dogfish and angel-fish on the other, 

 differs in this significant respect. They have been laved for 

 years, or one might properly say for generations before the 

 experiment, with solutions of different hydrogen ion concentra- 

 tion." 



It is here implied that the action of a stimulus upon the 



