LIFE PROCESSES IN CAPTIVE GRAY RATS 41 



atrophy has increased also, or average litter size would not 

 remain constant. It is known, from the investigations of 

 Huber ( '15) that many rat fetuses die soon after implantation. 

 Such fetuses, as well as those that die at later stages of gesta- 

 tion, are absorbed in situ and not aborted as in higher mam- 

 mals, therefore the proportion of the sexes among them cannot 

 be determined. 



Recent work in genetics has thrown much light on the cause 

 of fetal mortality, by showing that germ cells frequently con- 

 tain lethal or semi-lethal genes that either kill the embryos 

 carrying them, or so reduce their vitality that they rarely 

 survive. Such deleterious genes have been reported in mice 

 (Little, '19; de Aberle, '27; Hagedoorn and Hagedoorn, '28; 

 Chesley, '35; Kamenoff, '35) and in the rabbit (Hammond, 

 '33). An assumption that germ cells of gray rats contain 

 genes that tend to kill or greatly reduce the vitality of em- 

 bryos, particularly males, would offer a plausible explanation 

 for the low sex ratio among the newborn. 



Since the sex ratio in gray rats is significantly lower than 

 that in albinos, it may be of interest to compare the sex ratios 

 in white races of man with those for colored races, chiefly of 

 negro descent, when these races live side by side under the 

 same topographical and climatic conditions, as shown in table 

 12. This tables is based on one given by Parkes ('26), aug- 

 mented by the inclusion of stillbirths in some entries. The 

 probable errors of the sex ratios are given when the data seem 

 to be sufficiently accurate to warrant their calculation. To 

 this table has been appended the sex ratio for captive gray 

 rats (table 11) and that for the series of albino rats main- 

 tained under the same environmental and nutritive conditions 

 (King, '24: table 1). 



In all groups cited in table 12 the sex ratios for black races 

 are lower than those for white races. When judged by their 

 probable errors, differences between corresponding ratios are 

 significant for the data from Heape ('09) and from Little 

 ('20), but not for data given by Nichols ('07). Jastrzebski's 

 ('19) data, given in round numbers, cannot be considered as 



