102 THE MEDUSAE. 



these structures may be variable in any genus, for a renewed examination 

 of the original specimens has shown that my description was erroneous in 

 this respect. Maas, arguing from this conclusion, believes that the status 

 of Aclauropsis and of Vallentinia is doubtful ; the former he is inclined 

 to unite with Gonionemus, the latter with Cubaia. I see no reason 

 to doubt the diagnoses of these two genera, given from new material, by 

 such a careful student as Browne, and therefore retain them. The char- 

 acters of greatest taxonomic value within the family are, as all recent 

 students agree, the presence or absence of centripetal canals ; the presence 

 of one or of two distinct series of tentacles ; the arrangement of tentacles, 

 whether uniform or in groups; and (although Maas dissents from this) the 

 presence or absence of sucking discs on certain of the tentacles. As to 

 which of these are of greater, which of less value, it is difficult to determine. 

 It seems to me that considerable weight is to be laid on whether the ten- 

 tacles are in a single series or in two series ; this character being, as we know 

 from the studies of Mayer ( : 04 ), one intimately connected with ontogeny. The 

 same is likewise true, though in less degree, of the presence or absence of 

 centripetal canals. Mayer has given taxonomic weight to the conformation 

 of the gonads, whether" papilliform or sac-like ; but this is a character 

 which changes much with growth, and one which I believe to be of minor 

 importance. 



The genera may be tabulated in the following artificial key : — 



A. With one kind of tentacles. 



1. Tentacles without suckers; no centripetal canals. 



Tentacles not in groups. Aglauropsis (? Maeotias, Ostroumoff). 

 Tentacles in groups. Gossea. 



2. Tentacles with suckers. Gonionemus. 



B. Willi two kinds of tentacles. 



1. Without centripetal canals. 



Four perradial tentacles alone with sucking discs. Vallentinia. 

 Two series of tentacles; the numerous primary tentacles all 

 with sucking discs. Cubaia. 



2. With centripetal canals. Primary tentacles with sucking discs. 



Four radial canal s. Olindias. 

 Six radial canals. Olindioides. 

 We musl bear in mind that the tentacles of group A correspond (as Mayer 

 : "1 has shown) to the primary tentacles of group B. 



