294 
its margins, although a few of them are found on its ventral side. 
Figures 1 and 2 show diagrams of the nerve cord of two segments. 
These diagrams were made from camera drawings of frontal sections 
by superimposing a large number of outlines of such sections. 
} N MET ERS EN FE 
5 Se (- é. : 5 3 Mo IS, u: Shears j i 2 
| \ nn ne 
ln 2a a. A 
— ee — 
Sn 
Figures 3, 4, 5 and 6 are successive sections of one cell. Figure 3 is through ihe 
nucleus, but does not cut the sphere. Figure 4 shows a small portion of both nucleus 
and sphere. Figure 5 is through the center of the sphere, It shows centrosome and 
radiations and the band of denser protoplasm near the periphery. The orientation of the 
figures on the page is in all cases the same. (vom RATH preparation.) 
As is evident from the figures, the giant cells here show no uni- 
formity of position. The study of other segments of the worm leads 
to the same conclusion. There seem to be on an average, from eight 
to twelve such cells in each segment. Each cell has one process 
(Fig. 8), which it sends into the giant fibre or neurochord. From 
the study of a large number of these processes, I conclude that in 
entering the neurochord they always pass in a posterior direction. 
The contents of the process appear identical with the contents of the 
