144 JOHN C. PHILLIPS 



the following proportions of the sexes: In Fi, males 46, females 24; 

 in F2, males 34, females 30. 



In the first case is found practically a two to one ratio, which, 

 in view of the large number of observations, is significant. Guyer 

 called attention to an excess of males in various hybrid birds, 

 especially when the parents were from widely separated species 

 or different genera. His results are amply confirmed in a recent 

 paper by Goeffrey Smith and Mrs. Haig-Thomas on pheasants. 

 An excess of males is apparent in fertile hybrids, but seems to 

 be greater in sterile hybrids. I myself have found that in hybrids 

 between Reeves and Torquatus pheasants, Jungle fowl crossed 

 with pheasants, and Swinhoe pheasants crossed with Silver, a 

 great excess of males occurs. « 



It is therefore very interesting to find the same result obtaining 

 in a cross between two domestic races of ducks, very dissimilar in 

 size, but derived from the same wild species, and producing fer- 

 tile offspring when intercrossed. It is possible to explain the in- 

 equality of the sexes as due to a selective death rate on the part 

 of the two sexes, a larger proportion of male zygotes surviving, but 

 the supposed selective death rate, if it occurs at all, must be 

 effective very early in embryonic life, for out of 76 eggs set in 

 1912, 60 hatched, and 57 were reared to maturity, producing 36 

 males and 21 females. Most of the eggs, which failed to produce 

 young were sterile or at least discarded as such after ten days 

 incubation (exact number not noted) so that a selective fertiliza- 

 tion seems to be a more probable explanation in this case, but 

 an early embryonic selective mortality cannot be ruled out. 



In conclusion I wish to express my thanks to Professor Castle 

 for help in working out details. 



SUMMARY 



1 . The high coefficient of variability of both the parent races, 

 especially the mallards, may perhaps increase the variability in Fi. 

 Possibly the mallard race may contain an admixture of the small, 

 so called 'toy' race, which would account for some of the small 

 individuals which occasionally appear. In spite of this, however, 

 the male ducks show a very considerably increased variability 



