I50 MENDELISM chap. 



ferences as to the factors involved in connection 

 with the characters which he was investigating. 

 He would then set to work to test his hypothesis by 

 devising specific tests to settle specific points. It 

 might be that the example which he had chosen 

 offered no special difficulty to a complete and simple 

 factorial analysis. In such a case he would thence- 

 forward be in a position to control the characters so 

 analysed, and to build up strains of any combination 

 that he wished. Fortunately for the investigator, 

 however, plain sailing of this sort is by no means the 

 rule. His working hypothesis is often wrecked by 

 an unsuspected snag. He is up against some fresh 

 phenomenon, and the work at once becomes more 

 interesting. For the unexpected brings with it the 

 hope of discovery. It is the object of this chapter 

 to give a few illustrations of the occurrence of the 

 unexpected, and of the way in which it has been 

 met. 



Many of the earlier experiments with animals 

 were made with fancy mice. Their many shades of 

 coat colour, the rapidity with which they could be 

 bred, and the small cost of keeping them, marked 

 them out from the first as suitable material for 

 genetic work. Among the various colours used 

 was yellow, and it was not long before Cu^not 

 showed it to be dominant to any other colour used. 

 Crossed with any other colour, heterozygous yellows 

 threw equal numbers of yellows and non-yellows. 

 But when such heterozygous yellows were mated 

 together it was found that although they gave many 

 more yellows than other colours, the proportion of 

 yellows was not as great as that expressed by the 



