Miscellaneous. 65 



of development, therefore, the Liriopides possess twelve tentacles, 

 and the Carmarinides eighteen. Whilst the radial tentacles are 

 hollow, supple, and endowed with great mobihty, the interradial 

 tentacles are short and rigid, and scarcely possess any but a pendu- 

 lum-like movement, like the tentacles of the Trachynemidae. 



The development of all the Geryonidse presents a series of very 

 interesting metamorphoses. The young individuals, on escaping 

 from the e^^, are very different from their parents, and, before at- 

 taining their definitive form, have to pass through various phases, 

 some of which have occasioned the formation of separate genera. 

 The metamorphoses of one species only (Liriope catharinensis) had 

 hitherto been studied by F. Miiller ; but M. Hackel now makes 

 known those of another Liriope and of a Carmarina. In all the 

 species, whether their fundamental typical number be four or six, 

 the tentacles are at first equal in number to the typical number, 

 then double, and afterwards triple in number ; then in the course of 

 development the number falls to double the typical number, or even, 

 in many species, to this number itself. The first radial tentacles, 

 which are only rudimentary, generally disappear as soon as the se- 

 cond have attained a certain length. On the other hand, the solid 

 and rigid interradial tentacles persist, in some species, until the 

 commencement of sexual maturity, and in others even through the 

 whole life. 



Besides this mode of reproduction, M. Hackel has observed an- 

 other, very strange one in the Mediterranean Carmarina hastata. 

 Certain individuals of this species, both males and females, contain 

 in their stomachs a sort of spike, formed by an agglomeration of 

 medusiform buds. The number of Medusae on one of these spikes 

 may be as many as eighty-five. A more careful examination showed 

 M. Hackel that the axis of the spike is formed by the lingual cone, 

 to the surface of which the medusiform buds are attached by the 

 middle of the dorsal surface of the umbrella. This cone, therefore, 

 acts the part sometimes of a tactile tongue, sometimes of a gemmi- 

 parous organ. 



It is still more remarkable that all these buds are octuply radiated, 

 whilst all the adult Carmarince and all the larvae which issue from 

 them are sextuply radiated. It is therefore impossible that tiiese 

 buds should ever become Carmarince. Moreover the whole organiza- 

 tion of these buds removes them from Geryonidae, to give them place 

 aniong the ^EginidaB ; and in fact these young buds become deve- 

 loped into an ^ginide abundant in the Bay of Nice, and described 

 by Hackel under the name of Cunina rhododactyla. This Cunina, 

 which is destitute of the long stomachal peduncle, is a sexual indivi- 

 dual as well as the Carmarina from which it is produced. 



The two families Geryonidae and iEginidae must therefore hence- 

 forward be united into one (Phyllorchidae, Hackel). Hackel, more- 

 over, shows that the differences which separate them are not so great 

 as has been supposed. The iEginidae alone, among the Craspedota, 

 have passed as being destitute of a marginal canal ; but this exception 

 disappears, the author showing that this canal also exists in them. 



Ann, a^' Mag, N. Hist, Ser. 3. Vol. xix. 5 



