162 BULLETIN OF THE 
In Figure 75 (Plate VI.) the left cell of this pair is represented ; the 
corresponding right cell has fallen out or is aborted in this series. The 
general position of these cells may be said to be midway between the 
first and third pairs, as well in height as in antero-posterior relation. 
They may in some cases lie nearer the median plane than the following 
pair, as well as farther from it in other cases. The extremes of varia- 
tion in both directious are shown in Figure 94 (compare the right with 
the left). It was very difficult to follow the processes of this (second) 
pair, and I can only say I think they pass into the dorsal commissure, 
and through that to the opposite side of the body ; but their further course 
could not be made out. It is close to this pair of cells that the stalks 
of the anterior dorsal cells enter the brain. (Compare Figure 94, pd. 
cl. d., the anterior of the two stalks.) 
The third pair of large cells may properly be called the commissural 
cells, on account of their intimate connection with the dorsal commis- 
sure. They are pear-shaped cells, and lie on the extreme upper surface 
of the brain, and near the median plane, as may be seen in cross sections 
(Fig. 77). Their position with reference to the commissure is somewhat 
variable. Sometimes they are located well to one side in the brain mass 
(left side, Fig. 78), but again they are found well up on the commissure, 
even so far that the apex of the cell reaches the median plane, and the 
entire cell is dorsal to the cesophagus (Plate VIIT. Fig. 99, cl. coms.). 
As has already been intimated, the processes of these cells cross through 
the commissure to the opposite side of the body. After leaving the 
commissure, they bend at once sharply to the rear, and may be followed 
some distance. They were ultimately lost to view near or alongside of 
the fifth pair ; not because they are in any way connected with those cells, 
but rather because the size of the latter tends to obscure the neighbor- 
ing processes. I believe that the processes pass one into each lateral 
bundle of the ventral nerve cord, but this point could not be estab- 
lished with absolute certainty. 
The fourth pair of cells (Figs. 63 and 94, el. gn. IV.) makes its ap- 
pearance several sections back of the third. They are intermediate in 
size between those already described, and possess nearly spherical nuclei 
(Plate VI. Fig. 82, xl. gn. [V.). They occupy the dorsal portion of the 
ganglionic mass near its posterior end (Fig. 63), and are situated only a 
short distance from the median plane. Their processes pass sharply ven- 
trad and toward the median plane, where they ultimately come to lie 
near the processes of the first pair of cells in the central unpaired por- 
tion of the ventral nerve cord. In spite of the difference in size between 
