46 BULLETix 1090, u. s. depaet:mext of ageicultuee. 



of young produced per year by raature matings, shows a probably 

 significant correlation of +0.41 ±0.12. Hereditary differences in 

 the ability to bear the young ahve are demonstrated by the correla- 

 tion of +0.51 ±0.11, ■vrhile hereditary differences in the percentage 

 raised of the young born alive are at least indicated by the correla- 

 tion of +0.32 ±0,13. The correlation in the case of percentage 

 raised of all the young is of course a compromise between these 

 figures, +0.36 ±0.13. 



In considering these results it must be carefully noted that a signifi- 

 cant degree of genetic differentiation among the families during a 

 given period of time does not necessarily imply a significant degree 

 of correlation between different periods. It is to be expected that 

 there will be a certain amount of differentiation among the lines of 

 descent, which diverged in the early generations within a family. 

 The lines of descent which are most important in the first period are 

 not always most important in the second. A striking example of a 

 change in the character of a family, No. 35, brought about by the 

 rapid displacement of one set of lines of descent by the expansion 

 of a hitherto insignificant line, is discussed later. If this family 

 and another one. Family 2, which also changed its character to a 

 very marked extent in the course of the experiment, were omitted 

 from the tables, the cori'elations would be increased. 



Table 3. — Correlations between the family averages for each character, 1906 to 1910, 

 and those for tfie same character 1911 to 1915. 



[The positive correlations indicate heredity. Based on 22 families.] 



Vitality: . 



Per cent bom alive +0. 51 



Per cent raised of those born alive +.32 



Per cent raised +.36 



•Growth: 



Birth weight +.65 



Gain +.64 



33-day weight +.68 



Fertility: 



Size of litter +.65 



litters per year +.25 



Young per yeax +-41 



