AND ZOOGEOGRAPHY OF INDIAN OLIGOCHiETA. 145 



SO that the testis sacs fuse, extinguishing the intervening 

 segment ; the spermathecaj are again reduced to one pair. 



(7) The. Origin of Terrestrial Olvjochmta. 



From an ancestor having the characters illustrated above 

 (text-fig. 19), then, we can derive the various genera of Monili- 

 gastrida3 without, as it seems to me, forcing the facts in any way, 

 and relying only on processes which can be illustrated from 

 actual examples of the genera themselves or of other families of 

 Oligochaeta. 



I was, however, surprised, and I must confess to feebng 

 pleased, to find that the diagram arrived at as above gives in a 

 very natural way a starting-point for the other families of 

 terrestrial Oligochteta as well. These all agree in having, 

 primarily, two pairs of testes in segments x and xi, and one pair 

 of ovaries in xiii ; the primary number of spermathecas in the 

 Megascolecidfe is two pairs, opening in grooves 7/8 and 8/9. For 

 the'oiossoscolecidfe and Lumbricida? it is not so easy to give the 

 primitive position of the spermatheca? ; perhaps all that can be 

 said is that the Glossoscolecinse, probably the ancestral group of 

 these two families, typically have spermathecae Avhich open in 

 front of the testis segments (Michaelsen, 17). 



The arrangements of the gonads and their funnels m the 

 higher families is thus obtained simply by the disappearance of 

 the posterior pairs of testes and ovaries respectively ; and the 

 arrangement of the spermathecae— in the Megascolecidje at any 

 rate— is arrived at by the disappearance of the anterior of the 

 three pairs of the ancestral form (the anterior pair of sperma- 

 thecge would correspond, as already explained, to the posterior 

 pair of male openings). The primary position of the spermathecss 

 in the Glossoscolecida} and Lumbricida;_ is, as has just been 

 indicated, scarcely ascertained with precision. 



The occurrence, in existing earthworms, of gonads or of traces 

 of gonads, in the segments in which, on the present theory, they 

 are supposed to have existed regularly in the ancestor, confirms 

 the view here taken. Woodward (28) has shown that m the 

 common English species the presence of additional pairs of 

 ovaries in segment xiv is much commoner than has been 

 supposed, and that additional testes are not rare. There are 

 also cases where the supernumerary ovaries are not limited to 

 one pair; Woodward (27) examined a specimen with seven 



pairs. 



Confining ourselves, however, to cases where the condition is 

 definitely that of the presumed ancestor— one extra pair of 

 ovaries in segment xiv— we ha,ve the following records :— Bergh 

 (quoted by Woodward) described three cases in species of Tmm- . 

 bricus. Woodward's own investigation shows nine cases of this 

 condition out of fifty worms examined. Beddard, in an 

 examination of a number of specimens of Perionyx excavaius (1), 



Proc. Zool. Soc— 1922, No. X. 10 



