558 Records of the Indian Museum. [Vor. XI, 
in Tiarella singularis, Schulze 1876, with its three distal rings of 
cnidoblast batteries and in Margelopsis stylostoma, Hartlaub 1903, 
which has been shown by Bedot (1911, p. 211) to be the free bud 
of the preceding species. 
It isa striking fact that annulated tentacles should be com- 
mon in the free-swimming medusoid generation, and should occur 
in the hydroid generation only in a few species, which, with the 
exception of Asyncoryne, are outstanding on account of their free 
or partially free habit. The connection of habit and structure ap- 
pears to be no coincidence, and, on the evidence before me, I 
would suggest that the arrangement of large cushion-shaped cells 
in prominent rings is an adaptation to a creeping or free-swimming 
life. Not only would the greatly enlarged surface area, due to the 
rings, add to the resistance offered by the organism to the sur- 
rounding water, and so check the rate of sinking, even were the 
organism immobile, but the very large vacuolar spaces, which the 
nodal cells of Annuleila gemmata contain (and which in absence of 
direct evidence I assume to occur in the similar cells of other 
species), may act directly as buoying agencies. It is possible also 
that these vacuoles in the tentacular rings may by their enlarge- 
ment and contraction supply in some degree the means of the daily 
vertical migrations so characteristic of most hydroid medusae. 
In none of the cases mentioned above has the detailed struc- 
ture of the cnidoblast rings been investigated. But the cushion- 
shaped cells of which the rings or nodes are made up in Annulella 
closely resemble in general appearance and detailed structure the 
isolated battery-cells which stud the tentacles of Hydra. A com- 
parison of the description and figures of these batteries in Hydra 
fusca, as given by Schneider (1890, p. 332, Tab. xvii, fig. 20), 
with the description and figures of Annulella in this paper throws 
particular emphasis on this resemblance. 
The resemblance to Hydra is further emphasized by compari- 
son of the structures of the cnidoblasts themselves (see Schneider, 
1890, p. 332 and pl. xvii). In Amnnulella I have recognized only 
two types of nematocyst, but both occur in almost identical form 
in Hydra. It is true that there are differences in detail; that 
Schneider describes no connection between the basal prolongation 
of the cytoplasm of the cnidoblast and a ‘‘ ganglion mass,” that 
he mentions neither the gubernacula within the macrocnides nor 
the spiral arrangement of barbules on the ejected filament. But 
these are negative evidences and in the examination of structures so 
notoriously difficult as cnidoblasts negative evidences are of even 
less moment than usual. 
The conjunction of capitate and scattered tentacles suggests 
relationship with the family Corynidae, but the capitation is very 
slight and might be regarded as a terminal development of the 
tentacle nodes. Some cases of scattered tentacles (without capi- 
tation) occur in the family Hydridae. 
The solid multiserial endoderm of the tentacles appears to be 
paralleled in only one other genus, Tubularia. Solid endoderm is, 
