NEW SPECIES OF EDWARDSIA FROM NEW GUINEA. 521 



the only exocoelic tentacles until three cycles of endocoelic tentacles are 

 completed, when fresh tentacles are formed in the remaining exocoeles, 

 making, with the six primary tentacles, the full complement of twenty- 

 four. 



The next step is shown by my specimen of E. mammillata. This specimen 

 is probably adolescent, and the secondary tentacles are just beginning 

 to be formed. Herej as is shown in text-fig. 1, a micromesentery has 

 been formed in each lateral megacoele. These micromesenteries correspond 

 in position with the fifth couple formed in Hexactinian development, 

 and might be considered homologous with them ; but in none of the 

 specimens I have examined do they bear any trace of longitudinal muscle- 

 banners, and, as I shall show in the sequel, there is no evidence to 

 warrant our attempting to make an exact homology between the micro- 

 mesenteries of Edwardsidse and the last two couples of the first cycle 

 and the succeeding pairs of mesenteries in Hexactinise. As regards 

 the tentacles in E. mammillata, it is obvious from text-fig. 1 that the 

 two directives, and the tentacles on each side of them, are primary ten- 

 tacles and prolong the two directive and the two dorso-lateral and 

 ventro-lateral megacoelic chambers. In each lateral chamber there are two 

 tentacles separated by the micromesentery. I was not able to determine 

 the point with certainty, but, so far as I could judge, the more dorsal of the 

 two tentacles in each of these chambers was the smaller and situated nearer 

 the edge of the disc than the rest. If this be indeed the case, the ventral 

 tentacle in each chamber is the primary, the more dorsal a newly-formed 

 secondary tentacle. As a rule, the primary tentacle is dorsal, the secondary 

 tentacle ventral in the lateral chambers ; l3ut in this matter E. mg,mmillata 

 agrees with JE. clapared'u. 



Specimens of species of Edioardsia with 11-15 tentacles have been men- 

 tioned by various authors, but none of them has been studied in detail. 

 Therefore, we must pass on to forms with sixteen tentacles, of which 

 four species are known — HI. claparedii, E. heautempsii, E. loilleyana, and 

 E. adenensits. The first of these differs in some respects from the next two, 

 and, as I shall show, E. adenensis is evidently an adolescent form of a 

 species which has a larger number of tentacles when adult. E. claparedii 

 has been thoroughly described by Andres (1). Of the sixteen tentacles, 

 two are directives, two belong to each dorso-lateral megacoele, two to each 

 lateral megacoele, and three to each ventro-lateral megacoele (text-fig. 1). 

 They are alternately long and short, and Andres' figure shows clearly that 

 the more ventral tentacle in the dorso-lateral and lateral megacceles is the 

 primary, but the middle tentacle of the group of three in the ventro-lateral 

 megacoele is the primary. Andreses figure also shows the eight micro- 

 mesenteries, but he failed to realize their character and makes no mention 

 of them. 



