314 



THE AMERICAN NATURALIST 



[Vol. LV 



Linkage between the two principal pairs of factors in- 

 volved in the crosses, Pp and {Re){rE), may not be 

 appealed to because the only possible linkage relation 

 would produce results diametrically opposed to those 

 with which we are confronted. Since according to our 

 hypothesis the recessive white parents were in each ease 

 of the composition pp{rE){rE) and the blue-splashed 

 Andalusian parent PP(i?e)(i?e), it is evident that linkage 

 would require the production oi p{rE) gametes by the Fj 

 blues, more often than P{rE) gametes. And similarly 

 the combination P{Re) should also appear more often 

 than p{Re). 



A complete linkage between these pairs of allelomorphs 

 would result in an ratio of 1 blue-splashed and 2 blue 

 to 1 white, the blacks not appearing. The tendency of 

 even weak linkage would be to reduce the proportional 

 number of blacks. This should be true irrespective of 

 the direction of the cross. It would further be true, that 

 unless crossing-over occurred in both sexes any linkage 

 whatsoever w^ould inhibit the production of F2 blacks 

 homozygous for P. As will be shown in a later section 

 of this paper, however, F2 blacks homozygous for P have 

 been identified. Even a casual inspection of Table IV 

 shows that a relative preponderance of blacks is a quite 

 constant characteristic. 



Crossing-over in the male fowl has been found by Good- 

 ale (1917) and in the male pigeon by Cole and Kelley 

 (1919). The latter investigators definitely state that 

 there is no crossing-over of sex-linked factors in the 

 female pigeon. Goodale states that none had been ob- 

 served in the female fowl, but that a definite test of the 

 matter would be made later. So far as the writer is 

 aware no further report has been made. It should per- 

 haps be pointed out that so far only sex-linked factors 

 have IxM'H <hvilt with, no autosomal linked groups in birds 

 havlim- -0 far bfcn reported. 



Thciv a If 110 v.. data available from crosses where p 

 and {Rf} are found in one parent and P and {rE) in the 



