178 THE VARIATION OF ANIMALS IN NATURE 



unique character of the proteins of each species. Experimental 

 embryology has shown that this unique character may be 

 maintained even in small fragments grafted into an individual 

 of another species. Perhaps some taxonomists would bring 

 forward certain pairs of very closely allied species that seem to 

 differ only in one or two unit characters. But we think it can 

 be safely said that, even in these cases, the few unit characters 

 are only indicators which the taxonomist finds convenient to 

 use. As soon as a comparison can be made on the basis of a 

 sufficiently large number of individuals studied alive as well as 

 dead, all sorts of other differences begin to appear, sometimes 

 not easy to define, yet statistically significant. Sometimes it is 

 a slight difference in habit that first suggests to the taxo- 

 nomist that there may also be undetected morphological 

 differences. 



Such considerations make it very doubtful how far the 

 abstract concept of species as mere collections of characters 

 really covers all the facts. But we may further recall that 

 many characters, which in taxonomy are conveniently con- 

 sidered as units, actually affect many different parts of the 

 body. Such are size, colour (especially 'ground colour'), 

 hairiness and sculpture. It is possible that these could be 

 reduced to unitary physiological effects, but this is unlikely. 

 As soon as we consider structure in terms of the physiological 

 processes that give rise to it, the whole idea of units becomes 

 more difficult. This is implicit in the idea of the multiple 

 effects of genes. A complete extension of this theory would 

 make every gene responsible in some degree for every part of 

 the whole, and the unit-character conception of heredity 

 would go by the board. Actually geneticists are now more 

 cautious than they were in the past in their theories as to how 

 genes affect development. As Morgan (1932a) has recently 

 stated, ' the earlier, premature idea, that for each character 

 there is a specific gene — the so-called unit character — was 

 never a cardinal doctrine of genetics, although some of the 

 earlier popularisers of the new theory were certainly guilty of 

 giving this impression. The opposite extreme statement, 

 namely, that every character is the product of all the genes, 

 may also have its limitations, but is undoubtedly more nearly 

 in accord with our conception of the relation of genes and 

 characters. A more accurate statement would be that the 



