BIGELOW: MEDUSAE AND SIPHONOPHORAE. 391 



two generations, one with, the other apparently without canals. 

 Which of two characters is the more important in classification is 

 necessarily more or less a matter of opinion : and this is especially true 

 of a group so imperfectly known as the Narcomedusae. What is most 

 needed now is not so much a continued discussion of family relation- 

 ships, as an attempt to fill in the gaps in our knowledge of the actual 

 genera and species concerned. When these are sufficiently described, 

 and the parentage of the various young Narcomedusae, on which so 

 many species-names have been based, has been determined, a sound 

 phylogenetic classification will follow automatically. 



Our knowledge of certain species, and groups of species, of Narco- 

 medusae, already fulfills this requirement; consequently, they are 

 universally recognized, though called, perhaps, by different names, 

 and differently located in classification, by different authors. There 

 are, to begin with, several species with interradial gastric pockets, 

 and with otoporpae,^ which fall into four genera according as they 

 have two developed and two suppressed tentacles (Solmundella) ; 

 four developed tentacles (Aegina) ; four developed and four suppressed 

 tentacles (x\eginopsis) ; or eight developed tentacles (Aeginura of 

 Maas and Bigelow, Cunoctona of Vanhoffen). And while these four 

 genera are grouped in one family by Maas (1904b) by Mayer (1910), 

 and by me (1909a), in two families by Vanhoffen (1908a), they are 

 universally recognized, with the limits outlined above. 



Equally well defined is a species-group without gastric pockets, 

 but with very conspicuous peripheral canals (at least in one genera- 

 tion), and with otoporpae. These are all located in the Solmaridae 

 by Mayer (1910, one genus Pegantha) and by me (p. 394, two genera, 

 Polycolpa and Pegantha): Vanhoffen (1908a) gives them the dignity 

 of a separate family, Peganthidae, with three genera. 



It is in the cases of the various Narcomedusae with perradial gastric 

 pockets ; and of those with neither pockets nor peripheral canals, that 

 we find the greatest confusion and disagreement. Species with 

 perradial gastric pockets may or rnay not have a peripheral canal- 

 system, or otoporpae. According to the scheme here followed they 

 form, together, the Cuninidae, with three genera, Cunoctantha and 

 Cunina with otoporpae, Solmissus without otoporpae. The first 

 two are so closely allied (differing only in the number of tentacles) 

 that they may finally be united. But until an intergradation is 

 actually observed, it is wisest to retain both. 



1 Ectodermic ridges running over the exumbrella, from the otocysts, bearing nematocysls and 

 perhaps with a sensory function. 



