66 
Variation in Chilomonas 
s 
o 
s 
\_ 
.JO 
-°'\ 
m 
■am Lm 
\\ 
readth 
^ — 
\ 
• 
10-0 n-0 120 
Breadth in mikrons. 
Fig. 5. Regression lines for Series B (Favourable conditions). • • = Regression of length 
on breadth, o o = Regression of breadth on length. 
1906). It seems to me to be a result of considerable significance that in organisms 
representing three of the important types of protozoan structure (namely the 
Rhizopoda, Flagellata, and Ciliata) the regression between size characters is 
substantially linear. Biometric work on a variety of multicellular organisms has 
shown that in such forms linear regression between size characters in the fully 
developed (i.e. adult) organism is practically the universal rule. To find the same 
thing true of Protozoa seems to me to be definite quantitative evidence that the 
factors concerned in regularity of form production, if not the same, at least operate 
in fundamentally similar ways in unicellular and multicellular organisms. 
We may next examine the index correlations. It is of considerable theoretical 
interest to know what degree of correlation exists between the length-breadth 
index and each of the characters entering into it. We shall then have a measure 
of the extent to which size of body and shape of body are associated in their 
variations. These correlations may be determined from formulae which are readily 
deduced from the fundamental theorems regarding the variation and correlation 
of indices given by Pearson (1897). The particular formulae used in the special 
case with which we have to do here are given in another paper by the present 
writer (1906), and need not be repeated. The values found for the index 
correlations of Chilomonas are given in Table VI. 
