GROWTH, VARIABILITY AND CORRELATION IN 
YOUNG TROUT. 
By J. W. JENKINSON, M.A., D.Sc. 
The growth of the organism has been made the subject of numerous quanti- 
tative investigations, and amongst those who have dealt with the growth of 
animals, including the human being, the names of Quetelet, Roberts, Bowditch, 
Minot and Boas are well known. 
From their investigations it appears that there are certain features which are 
characteristic of growth in general. 
The first of these is the fact that the growth-rate is not uniform but suffers a 
gradual diminution during the life of the organism. This is well established by 
many researches [Minot (Rabbit), Quetelet, Bowditch, Roberts, Boas (the human 
being), Potts and Minot (Chickens), Semper (Limnaea), Vernon (Strong ylocentrotus)]. 
This is also known to be true of the various parts separately, though in the parts 
the rate of growth and the change of that rate differ a good deal. Moreover there 
may be more than one period of rapid growth followed by first a rapid, then a slow 
decline in the growth-rate, during the lifetime of one and the same organism. 
In the human being for example there is one such period or cycle before birth, 
another cycle during the first few years after birth, while the growth-rate rises 
once more, only once more to diminish, about the time of puberty. Robertson has 
recently suggested that, on the assumption that some autocatalytic chemical 
process underlies the phenomenon of growth, the maximum growth-rate in each 
such cycle should occur when the total growth occurring in that cycle is half 
accomplished. It may be added that results calculated on this theory are in 
fairly close agreement with those obtained by actual measurement. 
A second feature which characterises growth is the relation between change in 
growth-rate and change in the variability of the organism or of its parts, the 
latter rising and falling with the former. While it is true that .the observed 
decrease of variability has been in some cases ascribed to a selective death-rate, 
yet Boas has urged that the relation referred to must necessarily exist, however 
much it may be obscured by changes in variability due to other causes. 
