298 DR J. STEPHENSON ON 
their anterior ends are much swollen ; their contractions are alternate, pass downwards 
and backwards, and have no time-relation to the contractions of the dorsal vessel. 
The genus Limnodrilus possesses in general two pairs of hearts; the present species 
is therefore remarkable in having only one pair. There is, however, in the sexual animal 
a special loop to the genital organs, including the seminal vesicles and egg-sac. There 
is no trace of any such vessel in the non-sexual animal as a rule; on one occasion, how- 
ever, in a specimen in which no genital organs were discoverable, a loop was seen in 
segment ix ; this specimen was perhaps just about to develop sexual organs. The loops 
develop along with the organs they supply, and soon extend backwards from segment 
ix in complicated windings; in the fully developed sexual animal they may extend as 
far as xvi,—-in one case two segments behind the posterior limits of the genital sacs, 
here appearing to be applied to the wall of the intestine; they end ventrally in segment 
x by joining the ventral vessel (fig. 12). The loops are contractile, and it is a fascinating 
occupation to watch the wave of contraction wandering along the complicated windings 
of the loop through segment after segment. It seems not impossible, therefore, to 
compare these genital loops with the second pair of hearts of other species of — 
Inmnodrilus ; they would be hearts developed only at sexual maturity, greatly increased — 
in length, and modified for their special function. I have no actual note of their dorsal 
origin ; it is, however, presumably from the dorsal vessel, since the supra-intestinal does 
not exist behind segment viii. 
The ventral vessel (fig. 12) is formed anteriorly at about the level of septum + by 
the union of a pair of vessels, one on each side, coming from the front end of the 
animal. After receiving the lateral loops of seement vil it becomes smaller, and in 
the next part of its course it appears to be variable. In the living animal it was — 
seen on one occasion to be continued as a fair-sized vessel to the junction of the hearts 
in ix, and so onwards after being thus reinforced. In three other cases it became 
very narrow; but close observation at times when it was filled with blood showed 
that it was here also continued to join the angle between the hearts, or to one or 
other heart immediately in front of the junction. In other cases no connection could 
be made out, nor could any such channel be discovered in either of two series of 
transverse sections; in these cases the vessel ended on the intestine in front of septum 
S, sometimes bending to the right, towards the anterior end of the channel in the 
right wall of the intestine (v. swp.) before disappearing. 
The condition therefore resembles in some degree that described in Bothrioneurum 
and Lophocheta (cf. Bepparp, 3, p. 70), where the ventral vessel of the anterior 
segments is continued backwards on the intestine as a ‘subintestinal’ vessel, while 
there is a fine channel of communication between this subintestinal vessel and the 
point of union of the hearts. The specimens in which the connection between anterior 
and posterior parts of the ventral vessel is altogether wanting show a condition similar 
to that described above for Branchiwra sowerbyi ; and with this, again, may be compared 
Tubifex costatus (Heterochxta costata), Clap. (10), the only difference being that in 
