MUSEUM OF COMPARATIVE ZOOLOGY. 131 



mantle vessel. The four radial tubes of the bell, and the appendages to 

 the lateral pair, have been well figured and described. Leuckart seems 

 to liken rudimentary offshoots of the lateral vessels to mantle-tubes. I 

 do not think these offshoots more than very distantly comparable 

 with that special pair of vessels, which arises from a tube medially 

 placed in the bell, connecting the junction of the radial system with the 

 stem cavity of the animal. Such mantle-tubes, for instance, as are 

 to be found in Agalma, Gleba, or other genera, do not seem to have 

 been observed in the nectocalyx of Apolemia. I think, however, that I 

 have found in the bell of Apolemia a structure homologous to the man- 

 tle-tubes in the Physophoridse, and represented in the Calycophore by 

 the somatocyst. 



The mantle-tubes in Apolemia are difficult to make out, but seem to 

 differ only in their size from those in Gleba. Radial tubes in these 

 two genera, however, differ very greatly ; for while in the one they reach 

 a development hardly equalled among Siphonophores, in Gleba, where 

 the cavity of the bell is very shallow, and the nectocalyx itself is more 

 of a bract than a swimming-bell, the chymiferous tubes have a mini- 

 mum development. So rigid is the nectocalyx of Gleba that the walls 

 admit of little motion, and most of the propulsion is done similarly to 

 that of Circe and other Trachynemidse, by a movement of the velum, 

 a crescentic-formed vail surrounding the opening into the shallow bell 

 cavity. As a consequence, the radial system is quite diminutive in size. 

 Nowhere among Siphonophores better than in the genus Gleba do we 

 find a nectocalyx (PI. III. Figs. 4, 5), when fully grown, so closely resem- 

 bling a bract, and it seems to me that a better proof of the homology of 

 the central tube of the bract or covering scale with the mantle vessel 

 of a nectocalyx could hardly be desired. 



Apolemia has a float and a true Physophorous nectocalyx,* while Gleba 

 has no float, and is radically different from the Calycophoridse, although 

 its multiplicity of nectocalyces is a true characteristic of the Physopho- 

 ridse. Therefore I think that the Hippopodidse should make one of 

 the three great groups into which the Siphonophorse may be divided, 

 and be considered an equal division with the Physophoridse and Calyco- 

 phoridse. 



* I figure (PI. I. Fig. 1) a fragment of an Apolemia, without nectocalyx or float. 

 I have already published a representation of the nectocalyx of Apolemia. Proc. 

 Bost. Soc. Nat. Hist., Vol. XX. 



