STUDIES ON THE NERVOUS SYSTEM OF CRUSTACEA. 487 
ganglia, and there arboresce, as in fig. 1 on the right side. In 
Astacus, however, I never succeeded in demonstrating the con- 
nection of any of these fibres with ganglion cells. Many cells 
were stained, as in the ganglion on the left side of fig. 1, but 
their connections could not be made out. In the lobster 
embryo I was fortunately able to fill up this gap in my previous 
work. Fig. 2 represents the stomatogastric system of such an 
embryo, and the element A is seen to start from a cell in the 
commissural ganglion (which in the embryo forms practically 
a part of the brain) and enter the ventral nerve. 
Other fibres of the ventral nerve come from the cesophageal 
ganglion. Figs. 3 and 4 represent enlarged views of this ganglion 
drawn directly from single preparations. Two kinds of cells 
are found in it, both of which send fibres to the ventral nerve, 
the one kind being unipolar, the other bipolar. The bipclar 
cells (figs. 1 and 4, C,, C,) are spindle-shaped cells sending off 
a fibre from each end. One fibre runs upwards to the azygos 
nerve (azy.), the other downwards through one of the ventral 
nerves. Some of these cells lie in the portion of the ventral 
nerve which adjoins the ganglion (figs. 1, 4, C,) rather than in 
the ganglion itself. The unipolar cells (figs. 1, 3, 4, D, E, M) 
give off a single branch which, however, generally soon 
divides, one of the branches bifurcating again almost imme- 
diately, so that there are really three fibres having their origin 
in the one unipolar cell (figs. 3, 4, D,, D,, Ds). Of these three 
fibres, one runs upwards to the azygos nerve (D,), whilst the 
other two run downwards through the ventral nerves to the 
commissural ganglia (D,,D,). It is probable that by means of 
these elements the gastric ganglion and the two commissural 
ganglia are put into direct communication. In fig. 4 a uni- 
polar cell (£) is drawn, which gives off one branch which does 
not divide but passes directly into the right anterior nerve. 
It is not improbable, however, that this is due simply to in- 
complete staining, and that the element is in reality of the 
same type as that already described (D). 
In fig. 3 is a cell (F) which appears to be closely applied to 
a single fibre. This probably represents an intermediate stage 
