84 PROCEEDINGS OF THE AMERICAN ACADEMY 



The elimination of rinjjs lias been carried farther in the acetabulum 

 than in either of the posterior somites of the body. In a few species 

 the papillte have been wfU preserv(_'<l on the disc; and in these cases 

 their arrangement shows that papillate rings alone have been preserved 

 in this region. 



If the historical development of the leech has been marked by a 

 progressive course of abbreviation, such as 1 have descril)ed, it is evi- 

 dent that an ancestral form must have existed in which the somites 

 were more nearly alike from end to end. The embryonic develop- 

 ment confirms this view, for in its earlier phases the somites form a 

 chain of very nearly like parts. Later a few (7-S) of the posterior 

 somites become constricted off and consolidated into the sucking disc. 

 The somites at the anterior end are the iirst to arise, and hence the 

 first to exhibit specialization. Among existing species, we find three 

 in Japan which have departed less fi-om the hypothetical ancestral 

 form than has Ilirudo. They agree with Hirudo in having 26 somites, 

 but differ from it in having a larger number of complete somites. 

 Ii. each of these species we find only five abbreviated somites at the 

 anterior end; but these five are abbreviated precisely as they are in 

 Hirudo. The sixth somite embraces five rings, two more than the 

 same somite in Hirudo. This difference explains other differences : 

 for instance, the position of the first pair of nephridial pores in the 

 15th, instead of the 13th ring; and the position of the genital pores 

 between the 32d and 33d, and between the 37th and 38lh rings. If 

 the difference in abbreviation is taken into account, it is seen that 

 tlie nephridial pores and sexual orifices hold the same positions as in 

 Ilirudo. Passing to the posterior end of the body, we find that the 

 23d somite differs from that of Ilirudo. In one species, Microstoma 

 pifiram, this somite shows no plain evidence of abbreviation (Fig. o) ; 

 in another, M. edentirhmi, tlie 2d and 3d rings of this somite are 

 often less plainly divided than the following rings ; and in tlie third 

 species, M. acranulatnm, these two rings are as plainly consolidated, 

 as in Macrobdella (Fig. 4). Thus one of these species has 18 com- 

 plete somites; the second has 17 complete, and an IStli nearly so; 

 tlie third has only 17 complete. The 24th, '2')lh, and 2Gth somites 

 contain in each species the same number of rings as those of Ilirudo. 

 The 103d ring o( J/, pif/rum (Fig. 5), which is the homologue of the 

 09th in Ilirudo, shows signs of duplicity at its margins; and the 102d 

 is constantly thicker than the preceding ring. 



In the land leech, abbreviation has been carried farther at both 

 ends than in Ilirudo. Tlie nutnbcr of somites is the same in both 



