TRANSACTIONS OF THE SECTIONS. 109 



The generative elements are formed iu oval sporosacs, developed cue on each of 

 the tliree primary canals at the spot where the wider base passes into its narrower 

 continuation. The ova may be seen within them iu various stages of development ; 

 they increase considerably in size before the commencement of segmentation, always 

 showing up to that period a large and distinct germinal vesicle with germinal spot, 

 and with a distinct nucleolus in the interior of the germinal spot. The development 

 of the ovum proceeds within the sporosac to the segmentation of the vitellus and 

 the formation of the planula, which now breaks through the outer walls of the 

 sporosac and remains for some time adhering to their external surface. The planula 

 differs remarkably from the typical hydroid planvda. It remains of a nearly sphe- 

 rical form, ne\ev acquiring cilia, and possesses little or no power of locomotion. 

 The gastric cavity, however, is fully formed. The author was imable to follow the 

 ova in their further development. 



The little jMedusa now described departs iu several important points from the 

 typical hydroid ^ledusa. From this it differs in the tei-nary disposition of the pri« 

 mary radiating canals, and iu the irregular non-symmetrical arrangement of those 

 which are subsequently formed. Among the very many specimens examined, the 

 author never found any in wliich the canals had become regular in their disposi- 

 tion, even in those which had discharged the contents of their sporosacs, and had 

 evidently attained the term of their existence. It diiiers also from the typical 

 Medusa in the form and non-ciliated condition of the planula ; and still further in 

 the fact tliat while the generative elements are borne on sporosacs, developed on 

 the radiating canals, the marginal bodies are ocelli and not lithocysts. 



4. Circe inveiiens (nov. sp.). 



Among the hydroid Meduste captured in the towing-net were two or three spe- 

 cimens of a species referable to the genus Circe of Mertens. It measures about 

 half an inch in its vertical diameter and about a quarter of an inch transversely. It 

 is cylindrical from its base upwards for about two thu-ds of its height, and then 

 contracts abruptly and arches dome-like towards the truncated summit, which is 

 surmounted by a solid cone of the gelatinous umbrella substance. From the 

 summit of the umbrella-cavity, a solid somewhat fusiform extension of the roof 

 hangs down in the axis of its cavity for about two thirds of its depth, and at its free 

 end carries the manubrium, which extends nearly to the codonostome. The margin 

 of the umbrella carried eighty very short and but .slightly extensile tentacles, which 

 were connected at their bases by a very narrow membranous extension of the 

 margin, with rather irregular free edge. Lithocysts are situated at irregular 

 intervals upon the margin. There are about sixteen of them ; they consist each of 

 a minute spherical vesicle with a single large spherical concretion. There are no 

 ocelli. There is a moderatel}^ wide velum. 



The radiating canals are eight in number. They spring from the base of the 

 manubrium, run up tlie sides of the solid process which hangs from the summit of 

 the umbrella, pass from this to the walls of the umbrella, and then run down to- 

 wards the margin in order to open into the circular canal. 



The generative elements are borne in pendent sporosacs, which spring from the 

 radiating canals close to the summit of the umbrella-cavity. 



The motion of the Medusa takes place by means of sudden jerks, reminding us of 

 the way iu which certain Diphyidte dart through the water. 



The Medusa possesses also a very singular habit of partial inversion. This takes 

 place along the line which separates the dome-like poilion of the umbrella-cavity 

 from the lower cylindrical portion, and consists in the withdrawal of this dome-like 

 summit and the lower portion of the cavity. "When thus iuverted, the little animal 

 presents a drum-shaped form, with the manubrium hanging far out of the 

 codonostome. 



Alexander Agassi z considers the genus Circe of Mertens synonymous with 

 Trachyncma, Gegenbaur, and points out that the name of Circe has been already 

 UHcd for a genus of MoUusca. He further removes it from among the_ true hydroid 

 MedusEe, and, regarding it as closely allied to the uE(iinid(e, places it along with 

 tliose in the HaplostomcBf Agassiz, a suborder of the Biscojihora. 



