36 HELEN DEAN KING AND HENRY H. DONALDSON 



In gray females of the first generation variability in body 

 weight did not reach its maximum at an early age, as in the 

 other two series of females (table 5), but increased gradually 

 until the 395-day period, after which it declined slightly. 

 These females were more variable in body weight at all age 

 periods, excepting at 90, 120, and 608 days, than females of 

 the tenth generation, but differences between corresponding 

 coefficients are small in most cases and therefore not signifi- 

 cant. At the end of ten generations, variability in body 

 weight had decreased but little in gray females. Comparison 

 of corresponding coefficients for males and females in the 

 tenth generation (tables 4 and 5) show r s that variability in 

 body weight was much the same in the two sexes, although 

 the males tended to be slightly more variable in early life. 



Although gray females in both the first and the tenth gen- 

 erations were more variable in body weight than albino fe- 

 males, the trend of variability w r ith age in females of the later 

 generation was more like that in albino females, since it 

 reached its maximum at an early age and then declined 

 (table 5). 



There were two sources of error in the data for body 

 weight in gray females that probably increased the size of 

 the coefficients of variation for age periods during adult life. 

 Data for a number of females carrying fetuses of sufficient 

 size to add many grams to the mother's weight were doubt- 

 less included, since in gray rats pregnancy is well advanced 

 before it is evident from casual observation: weights were 

 never taken, however, if females were known to be carrying 

 young. The data given include the body weights of females 

 that were suckling litters. Lactation is a severe tax on the 

 reserves of a female, especially if the litter is large, and the 

 mother may lose considerable weight at this time. Such 

 errors in the data cannot well be avoided if one is recording, 

 at stated times, the body weights of breeding females; but, 

 since they balance each other to some extent, they do not 

 materially affect the average body weight at a given age 

 for a large series of animals. They do, however, decidedly 



