20 BULLETIN 1090, U. S. DEPARTMENT OF AGRICULTURE. 



Table 2 gives the number of jna tings in each generation of each 

 family near the end of 1915. The average number of brother-sister 

 matings in the ancestry of the young from these matings was 11.4. 



Sex has a slight influence on birth weight and slightly more influence 

 on gain to weaning. These effects, however, are so small in compari- 

 son with those of other causes- of variation, and the ratio of the sexes 

 is so close to equality in all experiments, that it has not seemed 

 necessary to make tabulations of these characteristics with the sexes 

 separate. On the other hand, in studying the growth after weaning, 

 separate averages must be made for the males and females. Similarly 

 the influence of sex on the juvenile mortality can be neglected, al- 

 though very important in dealing with the death rate of adults. 



The age and previous record of the dam have been found to have 

 slight effects on some of the characteristics studied, but so slight that 

 they can safely be neglected in comparing different experiments 

 with each other. 



Summing up, size of litter and environmental conditions, together 

 with sex, in the case of adult characteristics, are the only factors for 

 which constant allowance must be made. Tabulations have thus 

 been made separately for each size of litter and for each year. A 

 method by which the averages for different sizes of litters are com- 

 bined in a single index has been described. 



CHANGES IN THE INBRED AND CONTROL STOCKS. 



The averages for each character in each size of litter in the inbred 

 and control experiments are given by years in Tables 6 to 22. The 

 indexes for each year, calculated as described above, are also given. 

 The results are presented graphically for the indexes in Figures 1 to 1 1 . 



Turning first to the indexes, we see that there are considerable 

 fluctuations from year to year. These fluctuations are evidently 

 significant, being based on quite large numbers. There is, moreover, 

 not only a remarkable degree of parallelism between the fluctuations 

 among the inbreds (A) and the controls (B) in most respects, but also 

 a high degree of parallelism between the fluctuations of different 

 characters. In the year 1910 all characters were at a maximum. 

 There was a sharp drop in 1911, improvement in 1912, decline in 

 1913, and a marked improvement in 1914. There was a pronounced 

 decline in 1915, continued to 1918, and followed by a rise in 1919. 

 The only departure from parallelism that requires comment is the 

 much greater decline in the mortality curves in 1916 and 1917 among 

 the inbreds than in the control stock. The marked decline in 1916 

 was undoubtedly due to the severity of the winter of 1915-16 and a 

 shortage of green feed in winter and early spring. The decline in the 

 inbreds, however, is probably somewhat exaggerated, since the pens 

 which they occupied were decreasing in number during 1916 and 1917 



