Raymond Pearl 
23U 
Now it is quite clear that for any given values of Vi and the value of the 
expression under the radical will decrease as /•].. increases. In our present case 
the excess of the non-conjugant coefficients of variability over those for the 
conjugants is not great enough to overbalance the tendency of the high non- 
conjugant Vyj to lower the value of Sis- This is quite clear if we arrange for 
a single case (Series A) the value of each term in equation (i) in parallel 
columns for conjugants and non-conjugants. Doing this, we have the following 
results : 
Term 
Conjugant 
Non-Conjugant 
hs 
«-'lH'y3--2*'l3Vi(3 
26-338 
88-303609 
44-462224 
■5566 
34-879840 
132-765833 
97-885993 
27-848 
123-476544 
66-994225 
1-1780 
107-141126 
190-470769 
83-329643 
So then we have for the conjugants 
^_26;338 x 9-8937 ^ 
100 
and for the non-conjugants 
„ 27-84cS X 9 1285 
2.^ =2-o421. 
The values obtained by direct calculation from the index distributions are 
2-697 and 2"502 respectively. The values from the formula are thus in very fair 
agreement with those from the frequency distributions directly, considering that 
the constants were taken only to three places of decimals. 
It is quite clear, then, that it is the higher value of 7\s for the non-conjugants 
which brings down the value of for that group. Series G shows the same 
relation. 
We may now turn to the organic correlations. Besides the correlation of 
length with breadth, the correlation of the index with length and with breadth 
will be considered. In Table VIII. are collected all the determinations so far 
made of the correlation between length and breadth of body in Paramecium. 
The upper portion of the table is arranged on the same plan as Tables V., VI. 
and VII. to bring out the differences between conjugants and non-conjugants in 
respect to degree of correlation between length and breadth. The constant 
tabulated is the well-known coefficient of correlation, r. In the column headed 
" Table " is given the number of the table in the Appendix, from which each 
value of ?■ was calculated. 
31—2 
