MUSEUM OF COMPARATIVE ZOOLOGY. 163 
the first and fourth pairs of cells, their processes cannot be distinguished 
from each other in size (pr’c. gn., Fig. 88). 
The fifth pair of cells lies farthest lateral and dorsal of all (Plate V. 
Fig. 63), forming as it were the posterior outer corners of the ganglionic 
mass. They are the largest of the five pairs of cells, on account of the 
larger amount of protoplasm which surrounds their nuclei, and they lie 
wholly without the fibrous mass of the ganglion (Fig. 63) ; in fact, they 
often project above the general level of the brain (Plate VI. Fig. 85). 
The processes of these cells are the most prominent of all, having a 
diameter twice as great as those from any other pair of cells. They 
pass directly backward into the corresponding lateral bundle of the ven- 
tral nerve cord, and for a long distance occupy the centre of this portion ; 
but farther back they cannot be distinguished among the numerous pro- 
cesses which occupy this portion of the cord. 
8. Dorsal Cells. —'The probable nervous nature of the dorsal cells has 
already been referred to, and to make this clear it is necessary to con- 
sider in detail their structure and relations to the brain. In life they 
appear spherical when viewed from above (Fig. 2), but when seen from 
the side (Fig. 3) they are evidently conical. The two constituting the 
anterior pair lie in juxtaposition at the median plane, the posterior ones 
farther apart and in contact with the posterior and lateral portions 
of the first. No particular structure can be made out in the living 
cell further than the presence near the stalk of a dark body, presumably 
the nucleus. The stalks pass ventrad and slightly posteriad into the 
substance of the brain, where they are seen to bend decidedly backwards 
(Fig. 3), and are then lost to view. Nothing further was determined 
from the living animal, since the possibility that they might be nerve 
cells did not force itself upon me until much later. 
In preserved specimens the shape of the cells is much altered. They 
are usually shrivelled and distorted; or, again, they often contain a 
huge vacuole on one side (Plate VIII. Fig. 98). Unstained specimens 
cleared in clove oil serve only to confirm what is seen in the living ani- 
mal, and show quite distinctly that the stalks of the cells are not con- 
nected either with the cesophagus or the external cuticula at any point. 
There is also clearly apparent in the cell a fine network, which takes its 
origin from the stalk (compare Fig. 99), and fills the whole cell with a 
mass of minute meshes. If now we examine sections through this 
region, the peculiar character of the cells becomes more apparent 
(Fig. 99). Each is surrounded by an extremely fine membrane, which 
is continued on to the stalk as a delicate superficial layer hardly recog- 
