502 MEDUSA OF THE WORLD. 



occur at the non-attached, upper pole of the larva. 4 perradial stomach-pouches develop, 2 

 being partially ectodermal and derived in some measure at least from the invaginated ectoderm 

 of the oesophagus, the other 2 pouches being entodermal and derived from the entoderm of the 

 stomach of the larva. These perradial pouches are separated by 4 interradial septa which 

 finally become perforated near the margin forming a peripheral ring-sinus. 4 funnel-like 

 cavities lined by longitudinal muscle-fibers sink downward into these 4 interradial septa from 

 the hypostome of the larva, and tentacles develop around the margin of the hypostome. Finally 

 the larva splits by transverse clefts into a series of discoidal, free-swimming, ephyra larvae, 

 and the original corona of tentacles is cast off and dies while a new set of tentacles and a 4- 

 cornered mouth may develop at the upper end of the remaining part of the larva. In some 

 species, and under certain conditions, this strobilization may be monodiscus and give rise to 

 but a single ephyra, while in others from 10 to 30 ephyrae may arise from a single scyphostoma. 

 Details of this peculiar process of delamination and regeneration, and of the larval stages in 

 general, will be found in Korschelt and Heider's text-book of the embryology of invertebrates, 

 part I, 1895, pages io'2 to 122, and throughout the text of the present work. See especially 

 Cyanea arctica, Auretlia aurita, Pelagia, Chrysaora, Cotylorhiza lubcrculata, and Cassiopea 

 xamachana. 



With reference to the relationships between Hydromedusae, Scyphomedusae, and actinians, 

 Goette's announcement that the oesophagus of the scyphostoma of Scyphomedusae is formed 

 of invaginated ectoderm seemed to suggest a close relationship between the scyphopolyp and 

 the actinozoa. Hadzi, 1907, however, re-affirms the work of Claus and supports the view that 

 the oesophagus of the scyphostoma is lined on the inside with entoderm and is not invaginated 

 but evaginated from the primary stomach. All of the stomach-pouches are therefore ento- 

 dermal according to Clans and Hadzi, whereas, 2 of the primary stomach-pouches are at least 

 partially ectodermal and 2 wholly entodermal according to Goette. 



We know, however, from Conklin's study of the development of Linuche, and Hyde's 

 research upon Aurellia, that gastrulation in one and the same species may occur either through 

 invagination or by ingression, and judging from the mass of evidence which has been accumu- 

 lated we can not doubt but that the mouth of the scyphostoma of one and the same species 

 may be formed either by invagination, or by a simple breaking through of both entoderm and 

 ectoderm in which the entoderm takes the more active part, or even by an evagination of 

 entoderm. The three processes may thus be closely related and the manner of operation 

 dependent upon which layer takes the initiative and assumes the more active part. It is there- 

 fore misleading to attach any deep morphological significance or to draw any sweeping con- 

 clusions in respect to the phylogeny of the Scyphomedusae from this process. Indeed, long 

 ago our faith in the rigid application of the germ-layer theory has become so shaken that the 

 discussion between the Goette and the Claus schools has lost most of its significance. 



According to the Claus-Hadzi view the Scyphomedusae are more closely related to the 

 Hydromedusae than they are to the actinians. Certainly the histological characters of the 

 germ layers of the scyphostoma resemble the hydropolyps rather than the actinozoa, although 

 both are so closely similar that no great weight can be attached to this fact. Herouard, 1909 

 (Comptes Rendus, Paris, tome 148, page 1225), supports the opposite view from histological 

 evidence, for he finds that the cellular investment of the pharynx of Scyphomedusse and scypho- 

 stoma larvae is similar in character to that of the gullet of Anthozoa, thus indicating a genetic 

 relationship between the Anthozoa and Scyphomedusae. 



According to Goette the 4 primary stomach-pouches are actively evaginated from the cen- 

 tral stomach-cavity and are of mixed ectodermal and entodermal origin. According to Claus- 

 Hadzi, however, they are wholly entodermal, and are only passively separated one from another 

 by the infolding of the 4 entodermal foldings in the stomach-wall which constitute the tseniolae. 



It is extraordinary that this contention should have persisted so long without a final settle- 

 ment of views. 



The table on the opposite page will serve to present the question more clearly. 



The scyphostoma larva of Nausithoe bears a remarkable superficial resemblance to 

 hydroids, being elongate, branched, and incased by a horny perisarc. It infests sponges. 



In the majority of the Scyphomedusae the sexes are separate, but in Chrysaora and in 

 certain Rhizostomae the medusae are hermaphroditic. 



