MUSEUM OF COMPARATIVE ZOOLOGY. 165 
ance, especially of the latter, can be easily understood : the nucleus has 
been shrunk by dehydration in the course of preservation. This shrink- 
age was prevented in a measure at certain points where the strong 
threads of the network were connected with it. It is important in this 
connection to call attention to the fact, that in those cases where the 
cells were unusually small the nucleus was most nearly regular in shape 
(Fig. 66). No particular attention was paid to this point when study- 
ing the living animal, but in the sketches made at the time I find that 
the nucleus, which in general size and position agrees with this struc- 
ture, was drawn with a regular oval outline (Fig. 2). This may well 
be its shape in life. 
The stalks of these four cells present a uniform appearance. They, 
or at least their initial portions, stain more deeply than any other tissue 
both in hematoxylin and in carmine solutions, and hence are easily 
traced so far as the stained portion extends. Under a high power the 
stalk exhibits in places a faint longitudinal striation, and sometimes 
shows lines of minute vacuoles between the striations. The method of 
termination in the cell has already been described. From the cells the 
stalks pass directly into the brain, those of the first pair entering just 
lateral to the second large ganglion cell on either side, and those of the 
second pair just antero-lateral to the fifth pair of large ganglion cells in 
the brain (Plate V. Fig. 63, pd. cl. d.). The processes, which in trans- 
verse sections appear to be directed toward the median plane, show in 
longitudinal section (Plate VII. Fig. 95, pd. el. d.) a backward tendency 
also. They may in this way be followed for a very few sections; in the 
last, in which they are prominent, one sees a splitting or branching of 
the process in various directions, but beyond this the parts can be 
traced at most a couple of sections. The more compact character of the 
mass and the large number of other fibres make it difficult to say 
whether any part of the process extends farther, or whether the whole 
is split up at this point into fine fibrille. The place where the splitting 
begins is surrounded by a considerable number of small ganglionic cells 
(Plate. VI. Figs. 78, 82). 
Bürger (’91, p. 639) describes two pairs of large ganglion cells and 
two subordinate pairs in the brain. His description is not in all points 
clear, and to judge from appearances the figures do not correspond to 
his interpretation. Of the structures which he calls “giant cells,” the 
anterior pair is my third or commissural pair; his posterior pair corre- 
sponds to my fifth pair, Of his subordinate cells, the pair which lies 
close to the commissure corresponds perhaps to my second pair, and 
VOL. XXIII. — NO. 3. 12 
