GROWTH OF THE GERM-CELLS 



21 



40- 



ENGLISH ARTISANS. 



WEIGHTS (ROBERTS) 



relation between the growth-rate on the one hand, and the 

 variability of the organism and the degree of correlation 

 between its various parts on the other. With regard to the 

 first, it has been found by various observers that as the growth- 

 rate rises and falls so does the magnitude of the variability 

 (which we may measure by the standard deviation (<T), or better 



by the coefficient of variability (T*- x 100, M being the mean). 



A few instances will suffice. 



The way in which the variability goes up and down with the 

 growth-rate will be seen at once 

 from the accompanying graph of 

 the alteration of these quantities 

 during the last growth-cycle of 

 the human being. The data have 

 been taken from the measurements 

 of Roberts for the weights of 

 English artisans. 



Again, the same general agree- 

 ment between these two magni- 



o 



tudes is found in the Trout during 

 the short period, the first ten weeks after hatching, for which 

 data are available, and this is true not only for the growth 

 and variability of the whole body (total length) but for those 

 of the parts as well, eye diameter for instance, length of head 

 and length of caudal fin. The graphs do not indeed run parallel 

 throughout their course, but the fall or rise in variability is 

 accompanied by a fall or rise in the growth-rate. The agree- 

 ment is perhaps as close as could be expected when it is 

 remembered that a very small portion of the total life of the 

 animal has been under observation, and that this small portion 

 possibly includes one of the points of transition from one 

 growth-cycle to another (see the graphs of the growth-rate 

 of the total length in appendix). 1 



1 The agreement between growth-rate and variability is much improved 

 by taking the mean increment and not the percentage increment as the 

 measure of the former. I mention this as I was much puzzled by finding 

 that a rise of variability at the end of the period was not accompanied 

 by a rise in the growth-rate when the latter was expressed, as I now 

 believe erroneously, by the percentage increment. 



7 6 9 10 II 12 (3 14 IS 16 17 18 19 20 

 ACE IN YEARS 



