SPONGES. 135 



Tetilla, accompanied by the complete loss of the radiate skeleton arrangement, both 

 of which changes are already partially accomplished in such forms as T. Umicola, 

 would give us a typical GeUius, and in GeUius I believe we have the starting point of 

 the entire sigmatomonaxonellid series. 



From this starting point the Sigmatomonaxonellida have branched off in various 

 directions in the course of their evolution. Total loss of microscleres has given rise 

 to forms with the spiculation composed entirely of oxeote megascleres, constituting 

 the family Homorrhaphidse of Ridley and Dendy, and the strong development of 

 spongin cement has led to the evolution of the very large sub-family Chalininae, 

 from which in turn some of the so-called Ceratosa have been derived by total 

 suppression of the spicules and their replacement by horny fibre. The arrangement 

 of the megascleres in a characteristic plumose fashion, accompanied in most cases by 

 the replacement of many of the oxea by styli and the loss of microscleres, has 

 given rise to the family Axinellidaj. The development of a new type of microsclere 

 the chela by modification of the sigma, has given rise to the great family group 

 Desmacidonidse, within which the Ectyoninae have arisen by development of spined 

 echinating spicules. 



In the Report on the " Challenger" Monaxonida, published 17 years ago, Mr. Ridley 

 and I proposed to divide the Sigmatomonaxonellida (= Halichondrina) into four 

 families, viz., Homorrhaphidse, Heterorrhaphidse, Desmacidonidse and Axinellidaj, an 

 arrangement which has been variously modified by subsequent writers, chief amongst 

 whom is Topsent. This author (59) has united our Homorrhaphida? and Heteror- 

 rhaphidse in one family which he terms " Haploscleridas," a proceeding which appears 

 to me justifiable in view of the obviously close relationship between the two, and 

 especially in view of the fact that the Renierinse and Chalininse have very probably 

 arisen independently from different though closely related genera, the Renierinse from 

 GeUius (and perhaps other genera), and the Chalininse from Gelliodes and Toxochdlina, 

 as well as, perhaps, in some cases from Renierinse. 



Topsent has also proposed the name " Pceciloscleridai " in replacement of 

 " Desmacidonidaa " a proceeding which appears quite unnecessary, although the 

 extent of his family is not quite the same as that of ours. 



I therefore propose in this Report to sub-divide the Sigmatomonaxonellida amongst 

 three families, viz., Haploscleridaa (including the Homorrhaphidas and Heterorrhaphidaj 

 of Ridley and Dioidy), Desmacidonkke and Axinellkke. The Haplosclerkbe, there 

 can be little doubt, should stand hist, and indeed occupy the position of a parent 

 group from which the other two families have descended. 



Family : HAPLO.SCLEIUD.E. 



Sigmatomonaxonellida in which microscleres when present are usually in the form of 

 shnnata, or derivatives thereof, but never chelre. The skeleton is reticulate and 

 the fibre is typically not plumose. The megascleres are usually diactinal. 



