522 PROF. GILBERT C. BOURNE ON EITE 



E. beautemjysii, described in detail by Faurot (10), and JE. iciUeyana have 

 the same number and arrangement of micromesenteries and tentacles, but 

 differ from E. claparedii in the following respect : there are three tentacles 

 in each of the dorso-lateral megacoeles, the middle one of the group of three 

 being the largest and obviously the primary tentacle. There are two 

 tentacles in each of the lateral and ventro-lateral megacceles, and in each 

 case the larger and more centrally placed primary tentacle is on the ventral 

 side of the chamber, not on the dorsal as in E. claparedii. Evidently the 

 precise order of the appearance of the micromesenteries and tentacles in 

 the megacoeles is liable to some variation within the limits of the genus ; and 

 it is just this variation that negatives any attempt to homologize the micro- 

 mesenteries that first arise in the lateral and ventro-lateral megacceles of the 

 Edwardsise with the fifth and sixth couples of the first cycle of Hexactinian 

 mesenteries. Were the order of appearance always the same as it is in 

 E. claparedii (and possibly in E. mamiuillata) the homology would be 

 justified ; for in this species the primary tentacle remains in what in a 

 Hexactinian w^ould be an exocoele, and the new secondary tentacle is 

 formed in what in a Hexactinian would be an endocoele, and thus far 

 the order of succession would be exactly analogous to what has been 

 recorded by Faurot (10) for lli/anthus partlieiwpaus and by Haddon (13) 

 for Halcampa clirysantliellam. But the different order of succession in 

 E. xoilleyana forbids our extending this homology to the whole genus. In 

 this species (and also in E. tiinida, as I shall show further on), in both 

 the lateral and ventro-lateral megacoeles, the primary tentacle remains in 

 what in a Hexactinian would be the endocoele, and the newly-formed 

 secondary tentacle is in the exocoele. Further, when new micromesenteries 

 and tentacles are added, as is the case in E. timida, they arrive on the 

 dorsal side of the primary tentacles (text-fig. 1), and therefore in what in a- 

 Hexactinian would be the endocoeles. Thus a fundamental rule of the suc- 

 cession of mesenteries in Hexactinians would be violated. The conclusion is 

 that the mesenteries in question in the Edwardsife are not homologous with, 

 and certainly not homogeneous in the sense of being derived by descent from, 

 the fifth and sixth couples of the first cycle of Hexactinian mesenteries. 



In E. rakaiyce there are twenty mesenteries (eight niacromesenteries and 

 twelve micromesenteries) and twenty tentacles, disposed in a very regular 

 manner, as is shown in text-fig. 1. In this species it is sufficiently obvious 

 that the eight larger primary tentacles form the inner cycle, and that it is 

 the primary tentacle that occupies the central position in the group of three 

 occupying each of the dorso-lateral, lateral, and ventro-lateral megacoeles. 

 It follow^s f rom this arranoement that the tentacles do not alternate as in 



o 



E. claparedii and E. iciUeyana, but that there are four pairs of contiguous 

 outer cycle tentacles — one member of each pair on either side of the dorso- 

 lateral and ventro-lateral macromesenteries. In this species there is no 



