454 Growth, Variability and Correlation in Young Trout 
With two exceptions (total length correlated with length of posterior dorsal fin 
and length of ventral correlated with length of posterior dorsal fin) there is a 
sensible diminution in the value of the coefficient. This may depend on the fall 
in the growth-rate, as Boas has suggested. 
On the other hand in passing from the third to the fifth stage there is, except 
in the case of total length and breadth of caudal fin, a rise in the value of the 
coefficient, and, as we have seen, this is not accompanied by any rise in the 
absolute growth-rate. 
Nor does there appear to be any uniform connection between a rise or fall in 
the correlation coefficient and a rise or fall in the relative growth-rate. 
Thus between Stages III and V the index values of head and of body-breadth 
increase, and the coefficient of correlation between these and the total length 
increases too. But between the same two stages the index value of the position of 
the pectoral and pelvic fins decreases, while the coefficient increases. In the case 
of the breadth of the caudal fin on the other hand the coefficient diminishes 
steadily while the index value increases. 
Boas has found that the correlation is greater between axial organs — stature 
and height sitting, or stature and head-length — than between longitudinal and 
transverse parts. 
In the Trout there is — so far — no indication of this. The coefficient value 
often is high between longitudinal parts, e.g. total length and length of anterior 
dorsal or ventral fins, or total length and length of head, but it may be low 
between such parts, e.g. total length and length of posterior dorsal fin, or length of 
posterior dorsal fin and that of anterior dorsal or of ventral fin. On the other 
hand it is high between the total length and the breadth of the body, and of the 
caudal fin. 
Between the positions of one pair of metameric organs — the pectoral and pelvic 
fins — it is high : between the dimensions of other pairs of such organs (the median 
fins) it is lower. 
Lastly, constancy through the series in the index value is not necessarily a sign 
of high correlation (e.g. the length of the posterior dorsal fin and total length). 
The main results of the investigation may be briefly summarised as follows : 
(1) During the first ten weeks after hatching there is a decline in the 
absolute growth-rate. This is more rapid in earlier than in later stages. 
This is true not only of the whole body but of its parts. The growth-rate and 
the rate of its decline are not the same in the different parts. 
(2) There is diminution in the absolute variability accompanying and possibly 
caused by this general diminution of growth-rate. At the same time there is 
a sensible increase in variability in the final stage which is not accompanied 
by any rise in the growth-rate. Also at certain other stages in certain parts the 
growth-rate may change while the variability does not alter correspondingly. 
