776 PROCEEDINGS OF THE ACADEMY OF [DeC, 



the second pair of the quartette. It often happens that in the 

 adult medusa two members of a quartette, in opposite quadrants, 

 are retarded in making their appearance. In Aurelia Claus 

 established the theory that while the larval tentacles seem to come 

 in fours after the earliest stage, the first four tentacles appear first 

 two, then two more, as is the case in Gonionema {v. Claus, 

 1892). Goette (1887) has examined a great number of specimens 

 of the younger stages of Aurelia, and has come to the same gen- 

 eral conclusion as Claus with regard to the primitive paired condi- 

 tion and the significance of this in the philogeny. Haeckel 

 (1881), however, regards the appearance of two tentacles in 

 advance of the second two as an accidental and insignificant occur- 

 rence ; he takes four for the primary number. While this tendency 

 to a paired origin of the tentacles disappears after the earliest 

 stages in Aurelia, Gonionema exhibits this tendency in frequent 

 instances during the whole life of the animal. Its occurrence in 

 the appearance of the sense-organs is of the same significance, 

 because, as will be pointed out below, these organs are modified 

 tentacles. PI. XXXIII, fig. 19, shows this condition in the sense- 

 organs, quadrants A and C having five, while in quadrants B and 

 D only four are developed. It is true that other variations than 

 these do occur in the appearance of the tentacles and sense- 

 organs in the adult, and of the tentacles of the larva. Polyps 

 with three, five or six tentacles are not uncommon (PL XXXII, 

 fig. 13). It is noticeable that departures from the normal number 

 correspond very closely in polyps and adults. This would be 

 expected from the evidence that the larval tentacles are permanent, 

 and that they determine the position of the four radial canals in 

 the normal medusa, or of the three, five or six in aberrant speci- 

 mens. This inference seems a likely one from the fact that in the 

 adult medusse the tentacles which, from their larger size, are pre- 

 sumably of the first cycle are always, normally, located at the ends 

 of the radial canals, The inference is that five-parted medusae 

 were five-ten tacled polyps. This is borne out by comparison of the 

 relative numbers of each kind of variation among medusae and 

 among polyps. Hargitt (1901) has tabulated the number of 

 medusae that have come under his notice having three, five and six 

 radial canals ; and he finds that about five per cent, ai'e irregular 

 in this regard, i.e., vary from the normal four-parted condition. 



