THE AMERICAN NATURALIST [You LII 



Since these results occur in the F 2 count, it is unfortu- 

 nately not possible to deduce from them whether crossing 

 over takes place in one or both or neither sex. 



Amongst the tails were some that had both blue feathers 

 and white feathers. These give the group shown in Fig. 

 8, which closely corresponds to the blue-tail group (Fig. 

 6). There were other tails with white feathers having 



pigment along the margins as in Fig. 15. These, when 

 classified, gave the group shown in Fig. 9, which ap- 

 parently is the same as the group of white tails (Fig. 7). 



The number of birds in the F x and in the back cross 

 are too few to give significant results when broken up 

 into the two groups of white or blue. 



The tails are not a complete index of the bird from 

 which they came, for a bird with a pure white tail might 

 have color patches elsewhere on its body; but as no rec- 



