102 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY. 
a centrosome. It is not confined to this class of cells, but in good prep- 
arations occurs with such frequency that it may be said to exist in all 
the cells of the ventral cord (Fig. 13). The nucleus is always eccen- 
tric, and frequently, though not always, flattened. There are often two, 
three, or more centrosomes in a single cell. In one instance there were 
ten. In the cells whose cytoplasm is granular the centrosome does not 
appear as distinctly as it does in the others. However, when the stain- 
ing has been successful, there appears at the centre of the cell a mass of 
granules which are larger and stain more deeply than those of the rest 
of the cell. The granules of this mass are arranged in the form of hol- 
low spheres, the contents of which are destitute of all granules excepting 
the single round body at the centre. 
6. NeERvE FIBREs. 
a. Giant Fibres. 
There are three giant fibres which traverse the ventral cord through- 
out its entire length (Plate 2, Fig. 18, Plate 5, Fig. 31) ; the pair of 
extremely large ones, which lie one on the outer side of each of the 
paired connectives, and the smaller unpaired one lying in the median 
connective. All these have the same peculiarities of structure. With 
the methods employed they stain very lightly and appear almost homo- 
geneous. On close examination, however, the section of the fibre is seen 
to be made up of a small number of polygonal areas marked off by an 
indistinct network (Plate 1, Fig. 3). This network apparently owes its 
existence simply to the presence of discrete masses of protoplasm, the 
boundaries of which give the appearance of a network. In longitudinal 
sections the giant fibres show the same structure, except that the polygo- 
nal areas are elongated in the direction of the axis of the fibre. When 
these fibres are stained in methylen blue, the stain is precipitated at the 
borders of the areas, producing a finely granular network in a homoge- 
neous field of blue. 
The paired fibres may be traced forward into the circum-cesophageal 
connectives to a point between the anterior cirrus ganglion and the 
commissural ganglion, where they divide into a number of small branches. 
The branches cannot be distinguished from other large fibres of the con- 
nective, but they appear to pass through the commissural ganglion to 
the optic ganglion. The fibres which connect the commissural and optic 
ganglia are processes of the cells of the optic ganglion, but since I was 
unable to trace a fibre continuously from the optic ganglion until it 
