168 A. B. DAWSON 



which is in turn connected with the muscles or other effectors. 

 In addition to this, the sensory cell and the motor cell may also 

 be united with other ganglion cells of the nerve net, to provide 

 for the diffusion of an impulse from any one point. 



In the earthworm, although there is a well-defined central ner- 

 vous system, there are still vestiges of the more primitive con- 

 dition. It has been recognized for a long time that the epidermal 

 sensory cells of this animal possess, beside their main fiber, sev- 

 eral accessory processes (fig. 7) which extend into and appar- 

 entlj'- anastomose with the basiepithelial network. Extending 

 from this network, according to Dechant ('06), there are many 

 intercellular fibers which do not end freely in the epidermis at 

 various levels as usually described, but which continue outward 

 and end superficially just beneath the cuticula. They may even 

 anastomose at this level. Definite pericellular nets have been 

 described about the large unicellular glands and Dechant has also 

 demonstrated a very complex nerve net in the region of the 

 setae. In addition to these peripheral structures, there are the 

 various nerve cells scattered through and between the muscular 

 layers. 



Both tripolar cells (Kowalski, '09) and bipolar cells (von 

 Szuts, '14) have been interpreted as representing sensory cells 

 retreating from a superficial position to give rise phylogenetically 

 to the cells of the dorsal spinal ganglia of the vertebrates, and the 

 intraepidermal fibers were believed to belong to these cells. 

 However, in their structure and staining reactions, and in the ap- 

 pearance of their fibers (coarse and irregularly kinked), the tri- 

 polar and bipolar cells resemble very closely the motor and asso- 

 ciation neurones found within the ventral chain. Furthermore, 

 the relation of the crescent-shaped bipolar cells between pairs 

 of setae, with both processes extending into the basiepithelial 

 net and with no fiber extending to the central system (fig. 5), 

 argues strongly against any theory that these intermuscular cells 

 are deep-lying sensory cells. From the evolutionary stand- 

 point also it seems equally as logical to interpret them as scat- 

 tered ganglion or motor cells of the primitive nerve net which 

 have not yet been incorporated into the ventral nerve chain. 



