226 
Variation and Correlatio7i in the Earthworm 
table and the correlation is normal. It seemed probable that the discrepancy was 
due to too fine grouping in the table. In this particular case with the grouping 
adopted a considerable number of compartments have no observed occurrences 
falling in them. 
With the hope of bettering the results a somewhat coarser grouping was tried, 
the table then taking the form shown \\\ Table IX. 
TABLE IX. 
Length. Class unit 2*5 cm. 
o 
c- 
c+ 
D- 
i> + 
E- 
E+ 
F- 
F+ 
Totals 
|6| and 
5 
35 
44 
129 
122 
38 
8 
1 
382 
16; and 
2 
14 
23 
19 
9 
6 
73 
; 7 1 and 
7 • 
3 
1 
5 
1 
10 
• 7 i and 
|8| 
6 
4 
2 
3 
15 
\S\ and 
:8 ; 
1 
1 
2 
Totals 
7 
49 
73 
155 
135 
53 
9 
1 
482 
This gives only 40 sub-contingencies and a correspondingly larger number of 
the compartments are filled with observations. With this grouping we find the 
following values : 
= 70-615, 
= 1465, 
Ci = -3575. 
For the mean contingency from the same grouping we find : 
'»|r= -1093, 
whence = "33. 
This evidently comes fairly close to satisfying the relation Cj = Cj, and further 
manipulation of the grouping seems hardly worth while. We may conclude, then, 
that there is a very sensible degree of correlation between the size of the worm 
and the extent of the clitellum, the value of ?• being about '357 with a probable 
error of somewhat less than + '054. 
This result implies that growth occurs in the clitellum after it is formed. 
Putting all the results together we are able to form a fairly complete and connected 
picture of the morphogenetic phenomena connected with the position and extent 
of this organ. These phenomena we conceive to have probably the following 
relations : (a) the clitellum develops at such a point on the body that its posterior 
end lies a definite and, except in teratological cases, a fixed number of somites 
behind the openings of the genital ducts, {h) The clitellum when formed has a 
