654 MR. R. T. GtTNTHER ON [June 18, 



them as a subfamily of Eucopidae. With this conclusion my own 

 observations are at variance, and we do not consider that Seitaro 

 Goto's own drawing (pi. ii. fig. 15) makes his statement evident. 



With the aid of Dr. Cunnington's material I have again been able 

 to confirm my original observation that the axial cells of the 

 sense-organs of Limnocnida are derived from the endodervi of the 

 circular canal. I have repeated the same obsei-vation in the case 

 of Olindias mulleri while at Naples, and in the latter observation 

 I believe I have the support of Sir Ray Lankester. 



In both cases I have not been able to distinguish any boundary 

 line between ectoderm and endoderm in the youngest stages of 

 the sense-organs, although a distinct mesogioeal lamina appears 

 between the two layers when the sense-organ approaches full 

 growth. A further nidication of the endodermal nature of these 

 cells is afforded by an observation of Seitaro Goto himself, who 

 admits that the central cells of the organ " stain the same colour 

 as the endoderm, while the lining epithelium and the investing 

 cells stain like the ectoderm." 



In conclusion then, Limnocnida is to be regarded as a Tracho- 

 medusa related to the Olindiadfe and to Limnocodium, both of 

 which it closely resembles in important respects, but it difiers 

 from all other known Trachomedusfe in that the gonads develop 

 vipon the walls of the maiiubrium. Olindioides and Gonionema 

 difier from Limnocnida in the presence of centripetal canals and 

 in the fact that their tentacles are provided with adhesive discs 

 for use in creeping. In Limnocodium the vesicles enclosing the 

 sensory bodies are elongated and extend into the velum, in 

 Limnocnida they do not. Limnocnida too is characterised by its 

 power of reproduction by budding from the manubrium. 



The consequential changes in the scheme of classification cannot 

 be regarded as more than tentative, for the details of the struc- 

 ture of many of the Haeckelian genera are still unknown. A 

 most important character for taxonomic purposes is undoubtedly 

 that of the structure of the sense-organs — whether they project 

 freely beyond the margin of the luiibrella as sensory clubs, or 

 whether they are sunk and enclosed in vesicles. On these lines, 

 Browne has divided the Petasidpe into the subfamilies Petachnidse 

 (with sensory clubs) and the Olindiadfe (with sensory vesicles). 

 It is to the division of Olindiadte without adhesive discs on the 

 tentacles that I would provisionally refer Limnocodium and 

 Limnocnida. 



TRACHOMEDUS^. 



Sense-organs with endodermal axis ; gonads usually radial ; 

 development without a fixed hydroid stage. 



[The radial position of the gonads is believed to have been 

 derived from a manubrial position. Limnocnida is believed to have 

 " thrown back " to the older condition. Limnocodium has never 

 been proved to pass through a fixed hydroid stage.] 



