244 PERCY SLADEN TRUST EXPEDITION 



as an abnormal form of spicule*. That this is not the case I have convinced myself by 

 examination of Mr Carter's preparations now in my possession. Schmidt [1868] certainly 

 figured a very similar spicule in his Geodia ccuialiculata, but it is not clear, from the 

 descriptions given by him and by SoUas [1888], whether or not it is associated in this case 

 with a typical sterraster, i.e. whether the species is a Geodia or an Aurora. 



It is obvious from what has been said that the genus Aurora bridges over to a very 

 large extent the gap between the Stellettidse and Geodiidse, and the fact that Hentschel 

 described as a geodiid [Isops) a species which I feel constrained to place in the genus 

 Aurora, affords eloquent testimony to the close relationship of the two families f. If asked 

 exactly where we ought to draw the line between the two, I should say that further 

 minute anatomical investigation is needed before the question can be answered. Pro- 

 visionally we may take the typical, hilum-bearing sterraster as a distinctive feature of 

 the Geodiidge. The sterrospheraster occurs in both families and cannot be regarded as 

 distinctive of either ; indeed we cannot even regard it as affording the basis of a generic 

 separation from Aurora, because of its close relationship to the typical spheraster, the 

 gap between the two being completely bridged over by the adult spheraster of Aurora 

 aurora and A. reticulata and the developmental stages of the sterrospheraster. There is 

 just the possibility that the spheraster of Auro7'a aurora is not a true spheraster but 

 represents a case of convergence, and that true spherasters do not develop in the same 

 manner as sterrospherasters, but this does not seem very likely, and in any case more 

 information is wanted before we can settle the point. 



The close relationship between the Geodiidse and Stellettidse was recognised many 

 years ago by Czerniavsky [1879], who proposed the sub-genus Stello-geodia for a species 

 [Geodia stellosa) which he regarded as intermediate between the two, with the following 

 diagnosis: — " Membrana sarcodea superficialis, corticem tegens, stellulas minimas 

 numerosas breviradiatas continens. Parenchyma prseter globules siliceos Stellas majores 

 numerosas pauciradiatas continet." The author evidently regarded the presence of 

 euasters in addition to the " siliceous globules " as the distinguishing feature of his 

 sub-genus, but such euasters as he figures of course occur abundantly in the genus 

 Geodia itself and are by no means distinctive. On the other hand, it seems probable 

 from his figures and description that the "siliceous globules" are not true sterrasters, 

 and that the sub-genus may be identical with Aurora. The description and figures, 

 however, are not sufficiently accurate to enable me to decide this point, and it hardly 

 seems necessary to abandon the generic name Aurora in favour of Stello-geodia, especially 

 as the type species of the latter appears to contain tylostyles, which are not met with in 

 Aurora, though possibly these spicules are only abnormal forms of the oxea. 



As already indicated above, the transition from the tetractinellid to the epipolasid 

 condition by loss of the trisene megascleres appears to have taken place at least three 

 times in the genus Aurora. At any rate there are three epipolasid species each of which 



* This is also the view taken by Lendenfeld [1910 a] of the spicules termed by him "sterroids" (e.g. in 

 Geodia va/riospiculosa), which seem to be identical with the sterrospherasters. 



t It is not impossible that several other species at present included in the Geodiidse, such, for example, 

 as Hentschel's Geodia punctata [1909], may be shown by future research to be Auroras. 



