Kaymond Peakl 
241 
breadth of non-conjngaut Paramccia, I have formed Table IX., which gives for 
purposes of comparison a series of coefficients for different organs and charactci's. 
TABLE IX. 
Cumparisoii of Values of the Correlation Coefficient for Various Characters. 
Organism 
Correlated Characters 
r 
Actiiiospheriuiii * 
,, 
Ai'cella + 
Paramecium 
Daphnia 1 
„ 
Number of cysts and .size of body 
„ „ nuclei „ „ 
„ „ cysts „ _ cysts ... 
Diameter of shell and diameter of opening ... 
Length and breadth, mean of all non-coujugants ... 
Body length and cell length (Hatching to 3rd moult) 
„ „ „ (3rd to 4th moult) 
,, ,, „ (After 4th moult) 
■7692 
•8540 
- •6689 
•836 
•6271 
■5505 
■3930 
•2477 
The regressions for the length-breadth correlations are sensibly linear in the 
present samples. To show the nature of the regression, Diagrams II. and III. 
have been prepared §. Diagram II. gives the regression for breadth on length in 
the case of the conjugants, and Diagram III. the same for the non-conjugants, of 
Series A. 
Series A. 
39 
43 
S 47 
c 
)^ 
1 
p < 
< 
) c 
) 
1 
t 
> 
\ 
I 
1 
\ 
\ 
I 
C 
142:5 147.5 152.5 157,5 162.5 167.5 172.5 177.5 182.5 187.5 19-2.5 197.5 
Leiu/th. 
Di.\GRA5i II. Regression of breadth on length for the conjugants of Series A. 
* Smith, G. lliometrika, Vol. ii. pp. 243, 240. 
t Pearl, R., and Dunbar, F. .J. Ibid. Vol. ii. p. 3;J0. J Warren, E. Ibid. Vol. ii. p. 258. 
§ In the regression diagrams of this memoir, a broken line links points depending on too few 
observations to be reliable. The absence of any line between two points marks a total failure of 
intervening observations. 
