DENDY— HOMOSCLEROPHORA AND ASTROTETRAXONIDA 235 



in the choanosome ; in the ectosome they lie tangentially. The chamber system (so far 

 as investigated) diplodal* " (p. 177). 



Since 1888 a considerable number of species have been assigned to the " Epipolasidse " 

 and it has become increasingly evident that the group is of polyphyletic origin. It is 

 perhaps not even certain that all Epipolasids are reduced Stellettids, for it is quite con- 

 ceivable that similar reduction may have taken place in primitive Pachastrellids and given 

 rise to Epipolasids which never passed through a stellettid stage in their ancestry. Apart 

 from this question, hov^fever, it is by no means difficult to find, amongst the known 

 Stellettidge, genera, or even species, which seem to represent very closely the ancestors 

 of certain Epipolasids. SoUas himself pointed out that " If Asteropus is a reduced 

 Stellettid it is to Stryphnus that we must look for its nearest alliance " {I.e. p. 206). 

 In another part of the same work, however, he expresses the strong opinion that " Asteropus 

 has resulted from an Algol by the loss of trisenes" (p. cxlii). 



In 1905 I pointed out that the epipolasid genera Coppatias (now sunk in Jaspis) , 

 and Cryptotethya are evidently very closely related to Stelletta, and said that " Crypto- 

 tethya may be regarded as derived from some such form as Stelletta herdmani by further 

 reduction of the trisenes and by the outgrowth of the ectosome into finger-like processes " 



(P- lio). 



Most remarkable, however, is the existence of three epipolasid species, viz. 

 Rhahdastrella distincta Thiele, Diastra sterrastrcea Row and Aurora cribriporosa, n. sp., 

 each of which is represented by a closely related, tiisene-bearing species in the stellettid 

 genus Aurora. The relationships of these species will be discussed later on. 



In view of these facts it seems to me that the time has now arrived when we may 

 conveniently abandon the family Epipolasidee altogether and distribute its members as 

 best we can amongst the Stellettidse, and, if subsequent research should render it 

 necessary, amongst the Pachastrellidse also. 



I 



Genus Myriastra SoUas [1886]. 



Stellettidae with or without a distinct fibrous cortex and with only one form of 

 microsclere, a chiaster. 



This genus has been sunk by Lendenfeld [1903] in Stelletta, but it appears to ine 

 convenient to retain it, at any rate pending a much needed revision of the Stellettidee. 

 It seems impossible, however, to draw a real distinction between Myriastra and Pilochrota, 

 for all degrees of development of the fibrous cortex occur in different species of these 

 genera. I therefore propose to merge Pilochrota in Myriastra. 



■ 4. Myriastra parva (Row). 



Pilochrota parva Row [1911]. 



This species is represented in the collection by a small fragment about 6 mm. in 

 greatest diameter and of a pale, dull yellow colour. 



* On p. 141 he says "aphodal." 



30—2 



