Ch. VI, 8] HEREDITY AND VARIATION 311 



del was the first to obtain exact knowledge, which has been 

 confirmed and greatly extended by others. Using different 

 varieties as parents, he was able to trace the separate char- 

 acters in their hybrid offspring, and thus he discovered that 

 the rule in such cases is this, — the matter does not depend 

 upon chance, but one of the two determiners regularly prevails 

 over the other (is dominant, in his phrase), and shows its in- 

 fluence in the developing cell, while the other is latent (re- 

 cessive, in his phrase), and without visible effect. This is 

 the way in which parental characters can lie unseen and 

 latent in the body, thus in our common but erroneous phrase 

 " skipping a generation." 



There is, however, much more in the subject than this. 

 As already explained (page 285), when the adult individual 

 forms its own new sex cells, the number of chromosomes 

 and therefore of determiners, is halved by the reduction 

 division, but in such manner as to give to each new sperm 

 or egg nucleus one complete set. This set is taken partly 

 from the father set and partly from the mother set, the 

 combination apparently being made wholly at random, as 

 manifest by the fact that the different sexual cells of the 

 same individual differ greatly in the make up of their com- 

 binations (see Fig. 219). Thus it happens that every sexual 

 or germ cell contains a determiner for each character from its 

 father or its mother, but never from both, a fact called 

 technically "the purity of the germ cells." It is also true 

 that, for any given character, about as many germ cells 

 carry the father determiner as carry that of the mother. 

 Now if two individuals of the same kind breed together, as 

 imagined in our figure, and if the union of the germ cells 

 is left simply to chance, as seems to be true, then there 

 follows, so far as each single character is concerned, a very 

 remarkable and important result, which can most simply 

 be described by use of our diagram. Thus, if we center our 

 attention upon color of corolla (the circles with black, 

 dominant, and white, recessive), we find that four and only 



