castle: NOUTII AMERICAN RHYNCHOBDELLID^. 33 



This point being established, the somite limits must be marked off, in the 

 rei,'ions where unabbreviated somites occur, as in Figure 4, vi.-xxiv. 



I have further shown, in the publication already cited, that in Glossiphonia 

 somite abbreviation ^ is accomplished by a series of steps which follow one 

 another in regular sequence. First, a union takes place between the sensory 

 ring and the ring which precedes it; secondly, the ring which follows the sen- 

 sory ring is reduced in size ; finally, it too fuses with the sensory ring, the 

 entire somite being then represented by a single external ring. 



If, as is not improbable, some of the "abbreviated" somites are really in 

 arrested stages of development from the one-ringed to the three-ringed con- 

 dition Cas suggested in the case of Microbdella by Moore, 1900), the order of 

 the three steps enumerated should be reversed, in their case, and described in 

 the following terms : (1) A distinct narrow ring is separated off at the pos- 

 terior end of the unianuulate somite; (2) this newly formed posterior ring 

 grows in width ; (3) another new ring is separated off at the anterior end of 

 the somite. This produces a three-ringed somite, all three rings ultimately 

 attaining an equal width. For convenience in description, however, the pro- 

 cess will be uniformly treated as one of abbreviation, as explained on page 22, 

 footnote. 



The amount of " abbreviation," as is well known, becomes greater toward 

 either end of the body. 



Bearing in mind these principles, we find that the least affected of the 

 abbreviated somites of G. stagualis are those which stand nearest to the un- 

 abbreviated somites, namely, v. (Figure 3) at the anterior end of the body, 

 and XXV. (Figure ^) at the posterior end. In the case of each of these, the 

 anterior and sensory rings of the somite are united into a single broad ring. 

 But in the case of somite v. we find the union occasionally incomplete, as in- 

 dicated by the notch (less clearly than it should be) at the upper margin of 

 Figure 3, ring 4. 



Somite xxvi. (Figure A ; Figure 34, Plate 8) is usually found in the same 

 condition of abbreviation as the somites just described. Occasionally, however, 

 its posterior ring is narrower or less distinct than that of somite xxv. 



In somites iii. and iv. the process of abbreviation to a single broad ring is 

 practically complete, although the narrow posterior ring is in favorable pre- 

 parations still recognizable as a distinct element separated from the rest of the 

 somite by a shallow transverse furrow (Figure 7, in., iv.). 



Somites i., li., and xxvii. have each been reduced to a single ring ; in 

 addition a fusion (sometimes incomplete) has taken place between somites i. 

 and II., so that they are together represented by the broad ring, 1 (Figure 7). 



Somites xxviii.-xxxiv. are not represented by annuli on the surface of the 

 body ; they form collectively the posterior sucker. 



^ As to the sense in which this term is used, see p. 22, footnote. 



