i86 COMPARATIVE PHYSIOLOGY 



parent predominates in the heterozygous condition, it is called 

 the dominant character (long in this case) in contradistinction 

 to the recessive (vestigial in this example). 



An immense variety of characters both in plants and animals 

 have been found to follow^ the rule of segregation. To mention 

 but a few, colour of the hair in mam^mals, duration of life in 

 Drosophila, fecundity and absence of feathers on the neck in 

 fowls, brachydactyly in man, absence of eyes and wings in 

 flies. These suffice to show what diverse types of hereditable 

 characteristics, anatomical and physiological, depend on 

 segregating hereditary factors or genes. 



However, factorial analysis, as this method of investigation 

 is sometimes called, is not often as simple as in the case cited. 

 And those who have criticised the universal applicability of 

 the gene hypothesis usually do so in the expectation of a text- 

 book simplicity in eveiy instance. When we cross two strains, 

 it may, and often does, happen that the difference which 

 distinguishes them depends on more than one gene. The 

 applicabihty of the factorial hypothesis can here be substantiated 

 by the possibility of recovering types identical with both 

 parents in the F.2 generation. Of course, as the number of 

 genes involved increases, the number of possible combinations 

 in the F.2 increases, and the likelihood of reclaiming the 

 parental types diminishes. 



There is another criterion of segregation which has been 

 successfully applied to the analysis of a phenomenon which 

 has been held up as a stumblingblock to the general validity 

 of the gene hypothesis, namely the inheritance of size. In a 

 good many cases clear-cut size differences depending on single 

 genes have been found out. Very often, however, the F.2 

 form a continuous unimodal series. If segregation took place 

 in a cross involving a large number of factors, it follows from 

 quite elementary statistical principles that the coefficient of 

 variation in the F.i should not be greater than that of either 

 parent ; but that the coefficient of variation of the F.2 should 

 be greater than that of the F.i ; and that the coefficient of 

 variation of every subsequent generation would be on the whole 

 less and never greater than that of the F.2. Furthermore, 



