178 BULLETIN OF THE 



botli differ from the second type in Sipunculus in one important point, 

 namely, their arrangement. While they are (always?) found grouped 

 in clusters in the brains of Polychsets and Nemertines, such an arrange- 

 ment is never unquestionably present in Sipunculus, though indications 

 of a regular grouping were sometimes noticed. This may be regarded, 

 perhaps, as indicating a less highly specialized condition in the Sipuncu- 

 lid nervous system. According to Rohde and Biirger, these cells have 

 nuclei slightly smaller and more deeply stained than those of the first 

 class. I did not find any such difference between the two groups in 

 Sipunculus. The third type of cells in the Sipunculid brain shows 

 also a general correspondence to Class III. of the Nemertines and 

 Class II. of the Chajtopods. In both Chsetopods and Nemertines there 

 exists a fourth type, — tiie paired " giant cells " of the central nervous 

 system, with their accompanying "giant fibres." These are entirely 

 lacking in the Sipunculids. No one of the large cells has acquired any 

 uniform or considerable superiority of size over its fellows. Furthermore, 

 no giant fibres can be found in the ventral nerve cord, so that these 

 elements probably do not exist in the Sipunculid nervous system. This 

 may be regarded as farther proof of the lower grade of specialization in 

 the Sipunculids. 



The earlier investigators regarded these " giant cells " as " Bildungen 

 ganz verschiedener Art" (Spengel, '81, p. 40), but the more recent 

 writers incline toward the opinion that they are homologous throughout 

 (Eisig, '87, and Friedlander, '89). Now, either these "giant cells" are 

 neomorphic in both groups, and hence not at all homologous, or the 

 Sipimculids were separated from the primitive stem before the separa- 

 tion of Nemertines and Annelids took place, and before the differentia- 

 tion of these elements had been effected. A complete disappearance of 

 giant cells and giant fibres in the Sipunculids is hardly probable, in the 

 light of the persistence of these and all other nervous structures. This 

 would put the origin of the Sipunculids farther back than has usually 

 been maintained, and would make their relationship to the Annelids 

 somewhat distant. Of importance in this connection is the simple un- 

 differentiated condition of the ventral nerve cord, which shows no trace 

 of a metameric concentration of ganglionic cells, such as is found in the 

 Annelids. According to the researches of Andrews ('90), moreover, the 

 lateral branches lack that metameric character which has heretofore 

 been assigned to them, and I have been able to confirm this in part for 

 S. nudus. Lack of metamerism in the adult, as well as in the larva, 

 would serve to strengthen the view of only a remote relationship 



