MENDEL'S EXPLANATION 171 



the children of a human family, any more than 

 three heads and three tails are to be expected when- 

 ever a penny is tossed six times. The statement 

 that the recessives appear in the ratio of 25 per 

 cent, to the whole second hybrid generation stands 

 the experimental test well, when it is tested on 

 a large scale. In the case of the colour of the 

 cotyledons in the pea, I raised a generation of 

 139,837 individuals ; the ratio in which the greens 

 occurred was 24.88 per cent. The point I wish to 

 insist on here is that when small numbers of indivi- 

 duals are being dealt with small deviations from the 

 expected number do not indicate that the Men- 

 delian theory does not apply to that case, any more 

 than a close approximation is proof that it does. 

 For instance, if a brown-eyed couple marry, the 

 fact that their first four children are brown-eyed 

 does not prove that both parents are pure -bred 

 brown. The next child might be blue-eyed, which 

 would show that both the parents were hybrid 

 brown. Nor would the production by a brown- 

 eyed woman of three brown-eyed and one blue-eyed 

 child prove that the father was, like the mother, a 

 hybrid duplex ; he might be a simplex, and the 

 family would, in that case, be exhibiting such an 

 approximation to 50 per cent, duplex : 50 per cent, 

 simplex as might be expected in so small a number. 



Having dealt with this question, we may now 

 return to the Mendelian theory itself. 



There is a way of illustrating the result of the 

 random union of the germ cells, which is in some 



