566 Prolongation o£ Pregnancy 



Figure 1 shows the prolongation of pregnancy plotted against the number 

 of young lactated. It can be seen by inspection that the relationship between 

 these two variables is linear and that there is an vinmistakable positive corre- 

 lation between them. The correlation coefficient was calculated and found 



to be: , Q 



r..2/= +0.928 



This indicates a high degree of positive correlation. It is also statistically highly 

 significant, for a ^test shows: f = 11.68, 71 = 22, F^far smaller than 0.01 

 (Fishei,'^ p. 198 and 177). 



Consequently, there is a practically perfect positive correlation between the 

 prolongation of pregnancy and the number of young lactated. The calculation 

 was now carried further by computing the line expressing the relation between 

 the delay and the number of young lactated (line of regression). The equation 

 of this line of regression was found to be: 



y =r=. 1 :20.v -[-0.30 



in which y is the prolongation of pregnancy in days and x is the number of 

 young lactated. This equation shows that the delay of pregnancy per young 

 in this strain of mice is on the average 1.2 days. In the latter experiments with 

 hormones, the expected delays were always calculated by means of this equa- 

 tion. The line of regression is also marked on the dot diagram (fig. 1). 



Figure 2 shows the relationship between the prolongation of pregnancy ()') 

 and the number of young carried (Xj). Even a superficial inspection of this 

 figure reveals that there could be little if any significant correlation between 

 these variables. The calculation of the correlation coefficient fully confirms 

 this view: 



'Tx-^y =■ —0.067 



This value has no statistical significance (f = 0.318, rz=:22, and ^ = 0.75). 

 This "total" correlation coefficient may, however, be somewhat unsatisfactory. 

 Since the delay is greatly influenced by the number of young lactated (x^, 

 as has just been shown, it is possible that this influence may entirely over- 

 shadow any effect which the number of young carried might have had on the 

 prolongation. Therefore, a better index of the correlation between the pro- 

 longation and the number of young carried is the "partial" correlation coeffi- 

 cient, from which the influence of the number of young lactated is eliminated. 

 This was found to be: 



Under these conditions the correlation is somewhat increased, but even this 

 latter value is fairly low and statistically not significant (i =0.590, n = 2\, and 

 P^o.56). Consequently, no correlation could be shown to exist between the 

 prolongation of pregnancy during lactation and the number of young carried. 

 The Effect of ProgesteroJie upon Prolongation of Pregyiancy During Lacta- 

 tion. Since during lactation there is a persistence of functional corpora lutea, 



