114 THE VOYAGE OF H.M.S. CHALLENGER. 
cords has the appearance of a row of ganglia connected by bundles of nerve fibres. 
The size of these ganglia is not quite the same over the whole cord, the foremost beg 
slightly larger than those placed more posteriorly. As to their shape, I observed the 
following two different types. In some of the ganglia the cells are placed on both 
sides of the bundle, which passes through it, and these ganglia have a very regular 
rhomboidal form. The other type is represented by those ganglia in which ganglion 
cells are to be observed only on one side of the nerve bundle, and which accordingly 
show a triangular form. The triangular ganglia seem to be more numerous in 
Colossendeis, the rhomboidal form in Nymphon; in both genera, however, the stout 
ganglia, which are placed in the front part of the cord, and in the first place the 
comparatively large ganglion (figs. 6 and 8 y) observed by Dohrn are of a distinct rhom- 
boidal form. 
The form of the ganglion is, of course, determined by the number of nerves which 
branch off from it. The different ganglia give off besides numerous smaller nerves, 
one (in the triangular form) or two (in the rhomboidal) stouter nerves. These run in 
the foremost part from the one ganglionic bundle to the two others, and form nerve- 
rings (fig. 6, at a a™ &.), of which the secondary cesophageal ring (figs. 6 and 8 a) 
observed by Dohrn is the first and the stoutest. In Nymphon robustum 1 observed 
five or six of these nerve rings, but in Colossendeis they are still more numerous. 
With regard to the three stout proboscideal nerves, which have been observed already 
by Semper and Dohrn, and which, according to the latter author, terminate in the three 
ganglia (the front ganglia of my ganglionic bundles), I have ascertained that they are 
connected with the ganglionic bundles in the following way :—They run superiorly to and 
quite independently of the ganglionic bundle, till they reach the last but one ganglion of 
that bundle (fig. 8 wv). This they enter, their fibres passing through it and contributing to 
the comparative thickness of that part of the bundle which unites the last but one and 
the last of the ganglia (fig. 8 y). However, it is very probable that at least some of these 
fibres extend beyond the last of the ganglia. I am not quite certain whether perhaps, 
a union of the ganglionic cord with the proboscideal nerves does not also take place 
posteriorly. As I have stated already above, the proboscideal nerve gives off branches ; 
and about the middle of the proboscis of Colossendeis proboscidea, Sab. (sp.), on both sides 
of the stouter middle nerve two thinner cords run parallel with it in its immediate neigh- 
bourhood ; these are branches of the middle nerve. Investigating the first part of these 
lateral branches, close to their origin from the main proboscideal nerve, I once observed 
(in Colossendeis megalonyx, Hoek) very small ganglia with thin nerve threads running 
along this nerve without, however, exchanging fibres; these are, possibly, the end 
branches of the ganglionic bundle. While the origin of these branches and their 
connection with the proboscideal nerves is so easily noticed, with regard to their 
termination I only observed that the bundles, when approaching the end of the proboscis, 
