PEARL— SEX RATIO IN DOMESTIC FOWL. 417 



taken at its face value, that sex ratios may, in some cases at least, 

 be experimentally modified and in some degree controlled. The 

 critical value of all of this evidence is not equal. In some instances 

 it appears certain, and in more cases probable, that the data pre- 

 sented do not warrant the conclusion that the sex ratio has been 

 either modified or controlled. There is, of course, no theoretical 

 impossibility in modifying the sex ratio in an organism where the 

 chromosomal mechanism of sex determination is a definite and con- 

 stant one. We know of no hereditary character which may not, 

 upon occasion, be modified ; and in the case of sex the brilliant re- 

 searches of Goldschmidt- make it clear that not only the somatic 

 manifestation of the chromosomal sex mechanism may vary and 

 be experimentally modified, but presumably also the mechanism it- 

 self. But just because of the usual and normal stability of ger- 

 minal mechanisms it becomes the more important to be sure, on the 

 one hand, that evidence alleged to demonstrate that sex ratios may 

 be modified or controlled is sound and adequate when subjected to 

 the scrutiny of modern statistical science, and, on the other hand, 

 to learn more than we now know about the normal variability of 

 sex ratios. As a contribution in this direction it seems important 

 where possible to present and critically analyze statistical data, of 

 adequate amount, regarding the normal sex ratio of forms frequently 

 used in experimental work. 



It is the purpose of this paper to present and analyze such data 

 for the domestic fowl. The statistics here used cover eight years in 

 point of time, and represent over 22,000 individual chicks. 



The specific topics which will be discussed are these: 



1. The normal, average sex ratio in the domestic fowl. 



2. The variation in the sex ratio. 



3. The influence of prenatal mortality on the sex ratio. 



II. ]\Iateri.\l and ^Methods. 

 Before undertaking the presentation and discussion of the sta- 

 tistics it is desirable to say a word in regard to their collection and 

 analysis. The data are those which have arisen in the writer's ex- 



- Goldschmidt, R., Amcr. Xaf., 1916, and other papers. 



PROC. AMER. PHIL. SOC, VOL. LVI, BB, JULY IJ, I9I7. 



