BASIS OF THE PROBLEM 71 



germinal constitution of the offspring. Strains, therefore, do not 

 remain pure, as in the case of the bean, because they are con- 

 tinually crossed ; we require to know what happens when crossing 

 takes place. This study, first successfully undertaken by Mendel, 

 has been greatly extended in late years. We may first illustrate 

 in the simplest form what it was that Mendel discovered and then 

 go on to inquire what deductions bearing upon the question we 

 have to answer are to be drawn from it. 



A large number of experiments of the following kind has been 

 made. Two strains in any species are chosen ; these strains 

 exhibit opposed characters. Such characters may be tallness 

 and dwarfness, colour of the flower, shape of the comb in fowls, 

 condition of the seed, whether smooth or wrinkled, and so on. 

 We may call one character A and the other a. The two strains 

 are crossed, and in the first generation the offspring are all alike 

 and exhibit a character, A'. This character may be the same as 

 either A or a, a blend between them or something wholly new. 

 Whatever form it may take, it is produced by the interaction of 

 A and a. The members of this first generation are then interbred, 

 and of the second generation one quarter exhibit the character 

 A, one quarter the character a, and the remaining half the character 

 A'. If the quarter exhibiting the character A are interbred, all 

 the offspring exhibit the character A, and the same holds good 

 regarding the quarter exhibiting the character a ; but if the half 

 exliibiting A' are interbred, the offspring will split up in the same 

 proportion as in the previous generation, one quarter exhibiting 

 A, one quarter a, and one half A'. This result holds good for any 

 number of generations so long as interbreeding is continued. 



Into the very numerous complications which occur it is not 

 necessary to go. They are all interpreted by extensions of the 

 simple explanation which is applied in the elementary case given 

 above. It is supposed that all characters which behave on crossing 

 as above are represented in the germinal constitution by factors 

 which behave as separate units ; such characters are called ' unit- 

 characters ', and such factors ' unit-factors '. It is further sup- 

 posed that each gamete bears one unit-factor only in respect 

 of each unit-character. If the strain is pure, as in the case of the 

 two strains exhibiting characters A and a, then all the gametes 

 will bear the unit-factors for A and a respectively, and the fer- 

 tihzed eggs resulting from the crosses between them will therefore 



