56 BRISTOL. [Vol. XV. 



gation, and the new rings are added by the division of the 

 ultimate (3d) ring alone, or by the division of both the ulti- 

 mate and the penultimate, somewhat as new metameres are 

 added by the division of the part lying behind the last one 

 formed. There is not then a uniform growth throughout the 

 trunk, but a curve for each metamere." 



When I first found the intermuscular nerve ring I made a 

 very careful search in the 3d and 4th annuli for every trace 

 of nerves, and found that they were very weak in those struc- 

 tures, and when I came to plot down the nerve ring as it occurs 

 in the successive metameres, it became evident that the 5th 

 and 2d annuli of each somite (see PI. V) were in strong 

 and equal connection with the ist annulus which carries the 

 ganglion. 



The absence of anything like a proportional division of 

 nerves between the several annuli shows, I believe, that the 

 annuli weak in nervous elements are the younger and second- 

 ary annuli, formed by the division of the 2d and 3d rings 

 as follows : the posterior half of Clepsine 2 becomes 3 in 

 Nephelis, and the anterior half of Clepsine 3 becomes 4 in 

 Nephelis. This mode of formation of the five-ring type of 

 somite does not involve any shifting of the nephridiopore, as 

 would happen if the posterior half of primitive 3 became 5 in 

 Nephelis. 



This explanation assumes, of course, that the intermuscular 

 nerve ring is a constant feature in the leeches, and I feel 

 confident in predicting its discovery not only in the leeches, 

 but either that or its homologue in other annelids as soon as 

 they are studied with good nerve methods. There are many 

 evidences in the structure of the ring that it is an old and 

 very stable structure. The constancy through successive 

 metameres of such features as a long pedicelled cell always 

 in the same position on the ring ; the group of sensory fibers 

 separated from motor fibers near the inner ventral group of 

 bipolar cells ; the strong innervation from the central system, 

 and the definite longitudinal connectives, all point strongly to 

 the nerve ring, as we find it in Nephelis, as being very highly 

 specialized and the resultant of two originally distinct systems. 



