MUSEUM OF COMPARATIVE ZOOLOGY. 243 



cations passes around the bell parallel u-ith the margin, and joins a 

 corresponding bifurcation from the undivided tube of the hemisphere op- 

 posite that in which the bifurcation first described takes place. Before 

 these bifurcations join, however, each sends a loop downward, which 

 approaches the neighborhood of the bell margin, but eventually returns 

 to the bifurcation. 



The single ovum lies in a thin-walled sac,* which hangs from a point 

 directly under the apex, and when ripe fills the whole bell cavity, some- 

 times projecting a little through the opening. The free gonophore is 

 propelled in the water by violent contractions of the walls of the bell. 



The male gonophore, like the female, is often found free in the water 

 in which the Agalvia is confined. When attached to the stem it is 

 found in clusters at or near the base of a taster midway between two 

 adjacent polypites. In many live specimens of Agalma some of the 

 attached male gonophores will be found to have milk-white contents. 

 Like the female, the male gonophore is commonly transparent. 



The bell of the male gonophore is more elongated and larger than the 

 female. It measures 2.5 mm. in length and .4 mm. in greatest diame- 

 ter. At tlie apex of the bell there is a short peduncle by which it is 

 attached to the adult. Through this peduncle there extends a tube, — 

 the peduncular tube. There are are four thread-like simple radial 

 tubes which have a direct course in the bell walls and unite with a 

 circular marginal vessel. Each male gonophore has a narrow, thin ve- 

 lum. The bell walls are capable of quick contractions when free from 

 the axis. 



* Whether the egg in the gonophore is surrounded by a memhrane by which it is 

 held there, or not, no one lias clearly proved. I think such is the case. In the first 

 place, the homology of the gono])hore with the gonophores of other genera which 

 have an ovisac would seem to point to such a condition in Agalma. In my figure of 

 the egg just escaping from the gonophore a structure was observed in the bell cavity 

 which called to mind the ruptured walls of such a sac. After the ovum was cast, 

 there is reason to suppose that the " sac " is retained in the gonophore. Metschnikoff 

 speaks of the egg of Agalma as " nienibranlos." Haeckel says the egg-cell of Crys- 

 tallodcs, " wie bei den iibrigeu Siphonophoren ist ganz nackt," and that oi Physo- 

 phora, " wie die Eier aller iibrigen Siphonophoren sind dieselben durchaus hiillenlos." 

 '* Ilippopodius gleba," writes Metschnikoff (op. cit., p. 46), "ist die eiuzige mir be- 

 kannte Siphonophore, deren Eier mit einer freilich ausserst diinnen Membran iiber- 

 zogen sind." Hijjpopodius ami Vogtia, according to Kolliker, have ovisacs in which, 

 when in the gonophore, numerous ova are contained. I have also observed in a 

 Uudoxia which resembles E. Lemmii, tliat here also we have numerous ova in an 

 ovisac in the female gonophore, and there are many other similar observations on 

 record. 



