12 BULLETIN 1090, IT. S. DEPARTMENT OF AGRICULTURE. 



shall see later, by comparing the averages of whole families significant 

 differences can be found which must be attributed to heredity. 



For the purposes of the present work, frequency of litter has been 

 measured in a somewhat rough way. Matings are entered in a table 

 under the month in which the male reaches 4 months of age (an aver- 

 age of 3.5 months) or under the month following that in which the 

 mating was made if the male was already more than 3 months old. 

 The mating is dropped from the table the month after the death or 

 disposal of the female. The number of litters produced by a given 

 group of matings, divided by their effective duration in years, as cal- 

 culated by the above method, gives the average number of litters pro- 

 duced by a mating in a year. In comparing experiments it must be 

 borne in mind that difference in the age of maturity as well as in the 

 regularity in producing litters may be responsible for observed 

 differences in frequency of litter as calculated by this method. 



The production of a given size of litter, as in the case of frequency, 

 is only to a slight extent characteristic of matings. The correlation 

 between successive litters produced by the same mating among the 

 controls was —0.011 ±0.023 and that between litters which were not 

 successive was + 0.064 ±0.014. Similarly insignificant correlations 

 were obtained from extensive tabulations among the inbred families. 

 The correlation between parent and offspring in the average size of 

 the litters produced is in all cases so small as to be of doubtful 

 significance. Here, again, the only satisfactory evidence of heredity 

 is found on comparing different inbred families raised under identical 

 conditions. 



Variations in environmental conditions have a marked influence 

 on size of litters. The average is usually higher in summer and fall 

 than in winter and spring. In the controls (first 112 matings, 588 

 litters) averages of 2.75, 2.84, 3.26, and 3.16 were found for successive 

 periods of three months beginning with January to March. This 

 decrease in the average size of litter under unfavorable conditions 

 seems to be due to a reduction of large litters to medium-sized ones 

 (perhaps by death and absorption of some of the embryos), rather 

 than to increase in the percentage of small litters. The percentage 

 of litters of 1 and 2 in these data was found to be nearly constant at 

 all seasons of the year, but the percentage of litters of three increased 

 greatly in winter and spring at the expense of litters of 4 or more, 

 litters of 3 rising from 25.5 per cent in summer to 45.9 per cent in spring. 

 In such of the inbred families as were characterized by a markedly 

 smaller average size of litter the percentage of small litters was 

 much greater than in the controls. Such inbred families under 

 good conditions might produce litters of 'the same average size as 

 the controls under poor conditions, but the distribution of litter 



