22 



BULLETIN 1090, U. S. DEPARTMENT OF AGRICULTURE. 



as room was made for other experiments. This meant that an 

 unusually large percentage of the inbred young were born in both 

 years in the winter and spring months, when conditions were un- 

 favorable. Unfavorable conditions have a more immediate effect on 

 the mortality among the young than on other characters. This 

 explanation, however, is not wholly adequate to account for the dis- 

 proportionate decline in the inbreds as compared with the controls in 

 one set of characters only. It therefore seems probable that the in- 

 breds had reached a critical stage, in which a given change for the 

 worse in environmental conditions actually produced a disproportion- 

 ately great effect on the mortality. 



/906 

 100 



60 



/90& 



/9I0 



1912 



1914 



/9I6 



19/3 



/9£0 



k 



^ 60 



£ 40 



20 



X^V^ 



KS — 



Fig. 3.— The percentage raised of all young born. Inbreds (A ) and controls (B), 1906 to 1920. 

 for effect of size of litter as in Figure 1. 



Corrected 



Disregarding the fluctuations from year to year, a downward trend 

 is apparent in all characters. The trend can be measured by fitting 

 the best straight line to the graphs. This has been done by Pearson's 

 method of moments for the characters given below, for the years 

 from 1907 to 1915. It is, of course, recognized that a straight line 

 would not be appropriate for changes in percentages over a very wide 

 range. A uniform decrease in the ability to rear young would be 

 represented by a curve starting from 100 per cent as an asymptote 

 and falling away ever more rapidly for some distance. Fitting with 

 a straight line, however, brings out the fact of a decline. 



