OF THE HYDROMEDUSiE. 379 



to bo continuous with the endodermal portion of the edge of the peristome. Fol says, 

 (_!.!, page 484) that he found it difficult to trace, in his larger embryos, any visible Continu- 

 ity between the endoderm cells of the tentacles and the wall of the .stomach, but in our 

 species there is no such difficulty. The endodermal origin of the axial cells of the ten- 

 tacles of hydroids and medusae is such a firmly established fact, that the presence of 

 tentacles at this stage is, in itself, a proof that the digestive cavity is present, and would 

 in the present condition of embryological science compel us to regard the central struct- 

 ure as an endoderm rather than an ectodermal sub-umbrella. The tentacles, fig. 8, f 

 now rapidly elongate and their tips become enlarged and crowded with lasso-cells. 

 Fritz Muller, Haeckel, Fol and Metschnikoff figure at this stage peculiar hook-like ap- 

 pendages which project beyond the enlarged tips of the tentacles, but I have not ob- 

 served anything of the sort in our species. 



The larva shown in fig. 8 is a very interesting one, for it is in all essential points 

 a hydra with a gelatinous deposit between the ectoderm and the endoderm. It has a 

 mouth, a peristome, and solid tentacles, but no bell cavity; and if the thick umbrella 

 were absent, and the endoderm and ectoderm in contact, it would be almost exactly like 

 the floating, solitary actinula of Tubularia. It swims through the water, and its ecto- 

 derm is probably ciliated, and I think that comparison will convince any one that the 

 hydra-like stage is actually represented in the life-history of the Geryonidae; and that the 

 Actinula, the Geryonid larva, and the Polyxenia larva shown in Metschnikoff's PI. 3, 

 fig. 11 (51) are modifications of the same type, a free, solitary, swimming hydra with 

 solid tentacles. It is well known that the solid tentacles of the Geryonidae are transitory 

 larval organs, and that the persistent radial tentacles of the adult, PI. 42, fig. 1, are 

 hollow, like those of ordinary Anthomedusae and Leptom.edu sse. This difference is in 

 perfect harmony with the view that the larval tentacles are hydra tentacles, while those 

 which persist are medusa tentacles. 



According to Fol, who has given a very careful account of the changes which 

 now follow, from the study of an embryo which is much larger than ours, and, 

 therefore, more convenient for study, the periphery of the mouth area now thickens to 

 form a circular rim, from which the ectoderm of the tentacles is derived; while the rim 

 itself becomes the free edge of the umbrella, and gives rise on its inner side to a circular 

 fold of ectoderm, which becomes the veil. The ectoderm cells of the peristome, between 

 the rim and the mouth, become the epithelium of the sub-umbrella, which meets the en- 

 doderm around the mouth, where the line of demarcation between the two layers can be 

 clearly sen. 



Metschnikoff's account is like Fobs in all essentials, as he also says that the periphery 

 of the mouth area becomes the free edge of the umbrella, and gives rise to the tentacles 

 and velum, while the area between the velum and the mouth becomes the epithelium of 

 the sub-umbrella; but Lankester states (45) that there is a "substantial disagreement" 

 between Metschnikoff's statement (52, page 20), that " Der Centraltheil der Scheibe 

 stulpt sich dagagen weiter in's Innere ein, urn die aussere Bedeckung der Schirmhohle 

 darzustellen," and Fol's account. The two authors studied different genera, and we 

 should not expect to find an exact agreement in every point, but I fail to discern any 

 reason for questioning either of them, and certainly do not perceive any difference re- 



