HOW CHARACTERS BEHAVE 



143 



we deduce that with reference to a single character the total 

 offspring resulting from mixed breeding between two races, in 

 equal numbers where no selection is involved, will be in the 

 proportion of \ pure of one variety, ^ mixed, and \ pure of the 

 other. This is in the proportion of 25 per cent, 50 per cent, 

 and 25 per cent of the total population, or of i, 2, i. If the 

 proportion between yellow and white had been as 2 to i instead 

 of even, then the proportion of the pure and mixed kernels 

 would have been different but still definite and easily computed. 



Let us now see what would happen if this crop of pure white, 

 pure yellow, and mixed should be planted together again, eac/i 

 sort in its tnie proportion ; that is, just as would happen in 

 nature, supposing all forms to be equally vigorous and equally 

 able to withstand natural selection. We will tabulate this because 

 it gets rapidly complicated. In the table let the different combi- 

 nations planted be represented by the column headings across 

 the top, and the different kinds of pollen produced be repre- 

 sented by the headings down the side. 



Remembering that every kind of pollen will fall on every 

 kind of silk, and in definite proportions, the results are as follows, 

 the body of the table repre- 

 senting the various kinds 

 of progeny and the footing 

 at the bottom showing the 

 final and total population. 



In this table the expo- 

 nents represent the num- 

 ber of infusions of pure blood ; that is, y pollen on y silk gives 

 f" kernels, or two infusions of y, as compared to the y of yzv^ 

 which represents the first mixture of yellow and white, and so 

 on for other combinations. 



Now these facts are significant : first, we have all the combi- 

 nations possible between y and zv as the result of two admixtures ; 

 second, with all this admixture for two generations we still have 

 some white (^) and some yellow (it/*) remaining as pure as if 



