406 
Study of Blood Corpuscles of Tadpole 
We see that the two sets of tadpoles are again in close agreement as to vari- 
ability, but whether we judge absolutely or relatively their corpuscles are in 
length more variable than those of the adult frog. It is possible that some 
amount of this difference of variability may be due to difference of treatment, 
Dr Warren taking his corpuscles from the truncated tail and Professor Weldon 
from the heart, but the bulk of the loss of variability is probably due to the fact 
that the tadpoles were at a variety of stages of growth, and the corpuscles, as we 
have seen, are also changing with this growth. Tables A, B and C (pp. 412 — 414) 
give respectively the coirelation tables of body length and cell length, of total 
length and cell length, and of body length and total length for tadpoles of R. tempo- 
raria. The body and total lengths are in mm., but the cell lengths in working 
units, each of which = '00268 mm. 
Table III. gives the biometric constants of the total and body lengths in mm. 
We see that notwithstanding the small numbers, the relative variabilities are in 
strikingly close agreement. 
TABLE III. 
Body and Total Lengths of R. temporaria and B. vidgaris Tadpoles. 
Length 
Number 
Mean 
Standard 
Deviation 
Coefficient of 
Variation 
Frog 
Body 
71 
9-058 
1 -9443 
21-46 
Total 
69 
26-125 
5-5326 
21-18 
Toad 
Body 
29 
8-966 
1-7662 
19-70 
Total 
29 
20-415 
4-3103 
21-11 
The great range of variability here exhibited is due to the fact that we are 
dealing with different stages of growth, all clubbed together. One point, however, 
comes out from regarding the relative variabilities in Tables II. and III., i.e. that 
the influence of growth on variability of the body is nearly twice as great as it is 
on the variability of the cells. 
(3) I now pass to the subject of breadth of the corpuscle. Correlating body 
length with the breadth of the corpuscle, we find from the data in Table D (p. 415) : 
r = - -016 ± 016. 
This, as in the case of length of corpuscle, is negative, but it is insensible, having 
regard to the probable error. This direct correlation is so small that it does not 
look as if it were solely due indirectly to the correlation between length and 
breadth of the corpuscle. The latter correlation, found from Table E (p. 416), is -306, 
and if we find the partial correlation of breadth and body length for a constant 
corpuscle length, we have : r = + -062, again very slight, but possibly significant, 
or for a constant length of body, the breadth of corpuscle would increase with size 
of body. 
