238 



PROCEEDINGS OF THE AMERICAN ACADEMY. 



A, A(B), ■B(A), and B respectively, or once more 50 per cent indi- 

 viduals A in appearance. See Table IV. Selecting again for A, the 

 parents for generation 3 will consist of equal numbers of individuals 

 A and A(B) in character. Continuing the calculation in this way, we 

 get the series of generations indicated in Table IV, and expressed 

 graphically in the diagram on p. 239, A. For convenience in compari- 



TABLE V. 



Chances in 100 of isolating a Pure A bt Random Selection from Indi- 

 viduals MANIFESTING THAT CHARACTER IN THE VARIOUS GENERATIONS 

 FOLLOWING A CrOSS BETWEEN A PuRE A AND A PcRE B. 



son, there are also given in the last column of Table IV the percentages 

 of A and A{B) individuals to be expected when A is uniformly domi- 

 nant over B. Compare Table III. Inspection of Table IV will allow 

 one to continue it to any desired extent. 



The same law governs arrest of selection in cases of alternative domi- 

 nance, as in cases of uniform dominance of A over B. As soon as selec- 

 tion by elimination ceases, the race continues in the condition at that 

 time attained, provided forms A and B are equally fertile and subject 

 to the same mortality. 



On the other hand, if the breeder has the patience to make indi- 

 vidual breeding tests, and then to select for pure individuals on the basis 



* These percentages equal the series 0, |, |, |, j|, f^, etc. 

 t These percentages equal the series 0, \, f, f , |, \, etc. 



