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



tinue rapid growth longer and are much heavier at 1 year. A tabu- 

 lation among the inbred families (1910 to 1914) gave averages of 

 888 grams and 763 grams, respectively, for the weight at a year in 

 males and females. 



A little experience is enough to demonstrate that the size of litter 

 is a very important factor in determining the variations in birth 

 weight. Single guinea pigs are not only conspicuously larger as a 

 rule but also have longer hair and are more active than those in 

 large litters. The correlation between the size of litter and mean 

 weight of litter mates came out —0.658 ±0.007 as the average of 

 determinations in 11 inbred families and in the controls. The 

 correlation between size of litter and individual birth weights is, of 

 course, somewhat less (about —0.57). The correlation between 

 mean birth weight of litters and interval since last litter was +0.533 

 ±0.013 in the cases in which this interval was less than 75 days. 

 The correlation between interval and litter size was —0.457 ±0.015, 

 as previously stated. Analysis of these and other data shows that 

 size of litter has considerably more effect on birth weight through 

 its effect on rate of growth than by merely determining a longer or 

 shorter period of growth, a conclusion contrary to that reached by 

 Minot. 1 The analysis further indicates that the rate of growth is 

 affected by the general condition of the dam even more than by the 

 size of the litter. The hereditary potentialities of the fetus itself 

 are in addition important factors in a crossbred stock. 



Hereditary differences are, however, relatively unimportant in a 

 random-bred stock. The correlation between the mean birth 

 weights in successive litters was found to be —0.051 ±0.023 in. the 

 control stock and that between litters which were not successive 

 was ±0.058 ±0.014. Nevertheless, it will be shown later that 

 hereditary variation was sufficiently great in this stock to permit 

 the isolation of marked differences among the inbred families de- 

 rived from it. 



Even when all of the above factors are constant, as within a litter 

 produced by an inbred stock, there may be considerable variation. 

 There was an average standard deviation of about 10 grams within 

 litters of any given size from 2 to 7 in the control stock, after making 

 due corrections for the small numbers. The standard deviation of 

 the mean birth weights of litters, again making the necessary correc- 

 tions, was of about the same size, declining from 13.7 in litters of 

 2 to 7.4 in litters of 7. Thus in this case variations within litters and 

 among litters means are about equally important. Only an insignifi- 

 cant part of the variation within litters is genetic in the controls as 



1 Minot, C S., 1891. Senescence and rejuvenation. Jour. Physiol, v. 2 p. 97-153. 



I 



