PUNNETT—LAND-NEMERTEANS. 61 
The nervous system is arranged on the usual metanemertean plan. Giant neurochord- 
cells are apparently absent. The most remarkable feature is the presence of the accessory 
lateral nerve (cf. Pl. 11. figs. 5 & 6). This nerve is characteristic of certain species of 
Geonemertes (cf. Table, p. 60), and takes its origin from the posterior median surface of 
the dorsal ganglion ; whereas the lateral nerve proper is, of course, a continuation of the 
ventral ganglion. The accessory nerve is found in the same sheath as the lateral nerve, 
and on its dorsal surface. At the tail end the lateral nerves form a well-marked supra- 
anal commissure, and this also occurs in G. vodericana. 
The cerebral organ is small and opens on the ventral surface just in front of the brain 
(fig. 3). A well-marked frontal organ exists. In all specimens examined there are two 
large ocelli near the tip of the snout (fig. 2). In some specimens there are also found 
-what I take to be vestiges of another pair, but as they are small and destitute of pigment 
it is difficult to speak positively of their precise nature. 
The land-nemerteans of the world, though few in number, are not without importance 
in questions of geographical distribution; and no notice of a new species would be 
adequate without some account of its relations to the other members of the group. 
Coe has recently given a valuable synopsis of what is known about the genus G'eo- 
nemertes ; and the Table on p. 60 is an amplification of the Table which accompanies 
his memoir. From this Table it is obvious that the nearest allies of G. arboricola are 
G. palaensis from the Pelew Islands and G. rodericana from Rodriguez. G. arboricola 
differs from the former in the number of ocelli, from the latter in being hermaphrodite, 
and from both in its colour-pattern. In most features, however, it resembles them 
closely. With G. graffi* it also agrees in the presence of the accessory lateral nerve 
and in the absence of an excretory system. Moreover, all these forms are striped. 
It is not impossible that G. nove-zealandie may turn out to belong to the same 
group when more is known about it. G. australiensis resembles the forms mentioned 
in the presence of an accessory lateral nerve, but differs in having an excretory 
system; for, as is well known, it was in this species that Dendy demonstrated the 
existence of flame-cells. The remaining two species, G. agricola and G, chalico- 
phora, form a group apart from the rest, and are characterised by the absence of the 
accessory lateral nerve, the presence of an excretory system, and the small number of 
proboscis-nerves. ‘To place these two forms in a separate genus would doubtless give 
a truer idea of their affinities; but the number of land-nemerteans is at present 
small, and it seems better to await further discoveries before taking such a step. 
T am inclined to consider that the species palaensis, graffi, rodericana, and arboricola 
have arisen independently from some marine form. The form I would suggest as 
nearest to this hypothetical form is Pvosadenoporus, which is remarkable in possessing 
no excretory system and in being only found in the Indian Ocean. It is possible 
that nove-zealandie and australiensis may also have arisen from this form. On the 
other hand, I consider that agricola and chalicophora have arisen from some form not 
far removed from the genus Amphiporus. 
* I regard G. micholitzi as belonging to the same species as @. graffi. The differences upon which Birger 
separates the two appear to me unessential (Zool. Jahr., Abth. f. Syst. Bd. ix. pp. 272-4). 
