224 
Variation and Correlation in the Earthworm 
contains six somites and ends anteriorly and posteriorly in an intersegmental groove. 
Whenever the clitellar tissue overlapped so as to cover only a part of a somite at 
one end or the other, either the left or right vertical line was dotted, according 
as the overlapping occurred at the anterior or posterior end of the clitellum. 
Thus : 6 I indicates that the clitellar tissue is present on six somites, and that it 
only covers a portion of the most anterior one of these six somites. | 6 \ signifies 
that it is the most posterior somite which is only partially covered, and ; 6 ; indi- 
cates that overlapping occurs at both ends. 
In Table VII. are given the results of a determination of clitellum form in 495 
worms, the form classes being arranged in the descending order of the frequency of 
their occurrence. 
TABLE VII. 
Clitelhim Form. 
Form Class 
|6| 
|7| 
|8I 
:7| 
: 61 
7 ; 
•8! 
|8 ; 
: 7 • 
|6; 
Total 
Frequency 
389 
72 
14 
9 
4 
2 
2 
1 
1 
1 
495 
In 95'9 7o of all the worms the clitellum ends exactly in an intersegmental 
groove at both ends. In the remaining 4'1 % it will be seen that overlapping 
occurs much more frequently at the anterior than at the posterior end of the 
clitellum. Thus in only 1"4 of all the worms is there any overlapping at the 
posterior end, including the three cases, or "6 of the whole number, in which 
overlapping occurs at both ends of the clitellum. 
So far as may be judged from the present sample of worms, then, it appears 
that (1) clitellar tissue does not extend over more than eight whole somites ; 
(2) in a majority of cases it extends over exactly six whole somites ; (3) in the 
great majority of cases clitellar tissue ends exactly on intersegmental grooves, 
both anteriorly and posteriorly, and (4) when overlapping does occur, it is more 
apt to be at the anterior than at the posterior end of the clitellum. 
Now this spreading out or " overlapping " of the clitellum beyond the limits 
set by intersegmental grooves would apparently indicate a growth in longitudinal 
extent of the clitellum after it is first laid down. It seemed very desirable to 
determine exactly whether any such growth actually does occur, because it is 
usually supposed that the clitellum is an organ of definite extent which does not 
change so far as size is concerned. The matter can be definitely settled by deter- 
mining whether large worms have extensive clitella, oi', in other words, whether 
extent of clitellum and length of worm are sensibly correlated. 
In order to test this matter we have made use of Pearson's recently published 
Method of Contingency*. This extension of the theory of correlation makes it 
* " Math. Cent, to the Theory of Evolution, XIII." Draper,^'' Company Research Memoirs, Biometric 
Series I. pp. 35, Two plates. 
