60 Journal of the Mitchell Society [September 



Jereopsis would be related to the Astrophora, Ainpliitetliya to the 

 Sigmatophora.' 



The only conclusion that is possible is tiiat tlie triaene has varied 

 and become an amphitriaene, independently, in several groups. The 

 resemblance is analogical, one of ' convergent evolution ', a rubric under 

 which we group likenesses that are due to similar responses on the 

 part of related, but not necessarily closely related (witness the simi- 

 larities between the hydro- and scyphomedusae), organisms to the en- 

 vironment. We classify it, then, as due to a heritable change that may 

 occur in triaenes independently of inheritance, that is, reversion plays 

 no part in its appearance. We would like to know how to evoke it. 



If this and similar cases _in the sponges should ever be approached 

 experimentally, through the alteration of the external or internal en- 

 vironment, the series of known spicule forms connecting perfect amphi- 

 triaenes with perfect triaenes, in Aniphitethya )nicrosig))ia, Avould lead 

 us to expect that the change induced would be more or less cumu- 

 lative. Also our general knowledge of variation in sponge spicules 

 (cf. Wilson 1904, pp. 9-10) would lead us to expect that the herita])le 

 change would probably at first affect comparatively few spicules. 



That analogical similarities are common in sponges is generaly rec- 

 ognized, although there is not much that is definite in our knowledge. 

 Nowhere is more emphasis laid on their occurrence than in 0. Schmidt's 

 writings. As illustrating Schmidt's standpoint, the following cita- 



'^ The remarkable cliaracter of the lithistiil desma makes a strong argument for the 

 inonophyletic origin of the group, weakened in no degree by the fact that some desmas are 

 built on four-rayed, others on rod-shaped, basic spicule.s. For (see above) the variation 

 1 henomena in several species show that the latter kind of basic spicule is reducible to 

 the former. Sollas (1888, p. CXIX) has laid weight on this argument for the monoiihyletic 

 origin of the group, which lie derives from the Astrophora. The derivation from the 

 Astrophora in particular is based on the fact that the most frequently-occurring type of 

 microsclere in the Lithistida is astrose. The fairly numerous Lithistida without microscleres 

 offer no difficulty to this theory, for comparative study shows that microscleres 

 have been independently lost during the evolution of various sponges. But the few forms 

 with sigmata (Scleritoderma, Taprobane) do offer a difficulty, if we continue to lay such 

 stress, as now, on differences in the matter of microscleres. The whole question (cf. Dendy 

 1905, p. 99) can only be raised, but not answered until our critical knowledge is much 

 greater. Alternative hypotheses may be formulated as follows: (1) We may assume a 

 monoi>hyletic origin of the group and trace it back along with the Astrophora and Sigma- 

 tophora to Tetraxonida with both kinds of microscleres, asters and sigmata, in which case 

 it would certainly be astonishing that the two kinds of microscleres had mutually repelled 

 one another during the evolution of the Astrophora and Sigmatophora, and again within 

 the Lithistida during the evolution of the existing families. (2) If we assume first the 

 evolution of the Astrophora and Sigmatophora from Tetraxonida with both kinds of micro 

 scleres, and then the origin of the Lithistida from the former, it is at least conceivable, 

 as Sollas has noted (Inc. cit. p. CXX) that the occurrence of sigmata in some Lithistida is 

 a case of reversion. J'ollowing out this idea, the reversion to sigmata might concoivably 

 occur in Lithistids that had lost their microscleres. If we assume it to have occurred in 

 forms with microscleres (asters), tlien again we meet tlie strange conclusion that sigmata 

 and asters repel one another, tlie reversional variation which brings back sigmata driving 

 out the asters. (3) If we lay stress in the first degree on the difference in microscleres, 

 we are driven to conclude that" the Lithistida have had a double origin, some from the Astro- 

 phora, some from the Sigmatophora, and that the habit of forming a complex body, the 

 desma. on a free basic spicule, has been twice acquired. Doubtless the number of liypothe- 

 ses that are logically sustainable, might be increased. 



