INTERMUSCULAR NERVE CELLS OF EARTHWORM 169 



Indeed, the latter interpretation, for a number of reasons, ap- 

 pears the more probable. In the first place, there are the struc- 

 tural and staining resemblances between these cells and those 

 of the ventral cord. In the second place, the position of the 

 crescent-shaped, bipolar cells, the absence of any processes 

 passing from them to the central system, and the relations of the 

 cells to the basiepithelial nerve net are suggestive of a possible 

 role in the correlation of the movements of pairs of setae. (The 

 basiepithelial net is probably continuous with the net surround- 

 ing the setae, since the setae themselves have an epidermal 

 origin.) A third reason, perhaps a stronger one than either of the 

 foregoing, is that one finds spindle-shaped bipolar cells widely 

 distributed along the nerve tracts from within the margin of the 

 ventral chain to a position close to the epidermis. Finally, the 

 varying positions of the tripolar cells and the varying relations of 

 their processes to other parts of the metamere also suggest an 

 associational rather than a receptor function. 



The fourth type of cell described (Jig. 6), however, appears to 

 possess all the features characteristic of sensory cells, differing 

 from the epidermal sensory cells mainly in their position and in 

 the length of their peripheral processes (figs. 6, 7). In all other 

 features, such as staining reactions, shape, nuclear structure, 

 and the possession of both a main axone and one or more acces- 

 sory fibers, the epidermal and deep-lying cells are essentially 

 alike. Just what function these deep-lying sensory cells pei'form, 

 it is difficult to surmise. Their position in the circular muscle 

 layer and their possible restriction to the ventral region indicate 

 a probable role in connection with the initiation or maintenance of 

 the creeping movements of the worm. 



If we accept the evidence in favor of regarding the bipolar and 

 triangular tripolar cells as being concerned with a motor or asso- 

 ciational function and not primarily part of the receptor system, 

 w^e are necessarily required to attempt an explanation of the 

 origin of the intraepidermal nerve fibers. They are too numerous 

 to belong to the deep-lying sensory cells described, since each 

 of these so far as observed sends but one. fine process into the 

 epidermis. However, in the midbody regioir, oir which this 



