CLASSIFICATION 159 



ment further, and to make one group of the remaining Oligochaeta 1 ROSA has associated 

 all the terrestrial genera (including Moniligaster) into a group ' Terricolae.' It is 

 difficult, however, to get together a number of characters of first-rate importance, in 

 which they agree to differ from the Microdrili. The following list appears to me to 

 embody all the characters which the different families of earthworms have in common : 



(1) The clitellum never commences before the twelfth segment. 



(2) Its minute structure differs from that of the Microdrili, consisting always of 

 two layers of cells. 



(3) The ova are invariably small, and with little yolk. 



(4) The sperm-ducts traverse two or more segments on their way to the exterior. 



(5) The egg-sacs are small. 



(6) The spermiducal glands, when present, have not a muscular layer interposed 

 between the two layers of the lining epithelium. 



(7) Sexual maturity seems to be more or less continuous. 



(8) Oviducal pores invariably upon segment xiv, while ovaries are in segment xiii 1 . 

 We might perhaps add the invariable three pairs of nerves given off in every 



segment from the nerve-cord. The number varies in the Microdrili. 



These characters are simply the reverse of those which I have used to define the 

 Microdrili ; and they are, therefore, naturally in some cases negative characters, which 

 are not always of such value as positive characters. For example, the fact that 

 the sperm-ducts traverse more than one segment on their way to the exterior is 

 a vague character, because the number of segments traversed is a variable one ; in 

 some earthworms fewer, in others more, segments lie between the external and 

 internal orifices of the ducts. The clitellum, again, is exceedingly variable in position, 

 much more so than it is in the Microdrili. In spite of this, I consider that the 

 characters which remain are enough to indicate a community of descent of all the 

 different families which are included by ROSA in his group Terricolae. 



The next question to be examined is the relation between these two groups. 

 Can one, for example, be derived from the other, or are they both traceable only 

 to a common stock? The facts, as to the influence of an aquatic life upon normally 

 terrestrial forms, seem to tend to the conclusion that there is hardly any evidence of 

 the characteristics of the Microdrili having been caused (except perhaps by a very 

 prolonged exposure to those influences) by their aquatic life ; and, besides, those very 

 characters are found in the terrestrial Moniligasters, which at least indicates that, if 

 they are due to environment, it is a long time in acting, since M on Uigaster is distinctly 



1 Exceptions to this general statement which have been recorded (cf. Rhinodrilus pro&oscidews) require 

 confirmation. 



