15 
to younger stages of development. As a matter of faet, they are 
so far developed that the seeorid pair of tentacular bulbs have 
been formed and have reached a somewhat considerable size, though 
still distinetly smaller than the first pair. There is, accordingly, 
every reason to believe that the gonophores of Ichthyocodium sarco- 
tretis remain attached to the hydranths, even when mature. 
Jungersen has not offered any suggestion as to the mode of 
feeding of Ichthyocodium. The mouth of the hydranths has exaetly 
the same structure as in Kinetocodium (observed in sections cut 
by mei; the polyps are exelusively found on that side of the par¬ 
asitic crustacean, which is turned towards the body of the fish. I 
can see no other possibility, therefore, but that the hydroid must 
feed on the epidermis of the fish. 
As we have seen, the medusæ (or medusoid gonophores) of 
Hydrichthys mirus Fewkes, Hydrichthys boyeei Warren, and Ichthyo¬ 
codium sarcotretis Jungersen distinetly points towards the Tiarid 
medusæ as their nearest relatives. Medusæ belonging to the family 
Tiaridæ are, however, liberated from hydroids referred to two dif- 
ferent families, Bougainvilliidæ (Perigonimus group) and Clavidæ y 
so that the knowledge of the structure of the medusa is not suf¬ 
ficient to determine the systematical position of the hydroid. The 
absence of a perisarc on the polyps might indicate that the three 
species in question should be referred to the Clavidæ rather than 
to the Perigonimus group. In this regard our new genus, Kineto¬ 
codium, is important, in so far as it demonstrates that the perisarc 
of the polyps may be entirely lost in a species, which undoubtedly 
is related to Perigonimus, as evident by other reasons (see above, 
p. 11). Now, apart from the strong development of the mesosarc 
and the muscular elements in the pedicel, the polyps of Kineto¬ 
codium bear a great resemblance to the polyps of Hydrichthys, as 
described by Fewkes and Warren, and of Ichthyocodium , as 
described by Jungersen and seen from sections made by me. 
The lack of perisarc is, therefore, no objection to the supposition 
of a relationship between the genera in question and Perigonimus. 
Indeed, I am of opinion that they may, without any risk, be placed 
within the hydroid family Bougainvilliidæ and, within the latter, in 
the Perigonimus group. 
Looking through the “Pteropoda” byJ.J. Tesch in “Das Tier- 
