802 PROF. A. BENDY AND MR. R. W. H. ROW OX 



absent by suppression in a few cases where the skeleton has 

 undergone extreme modification. We are unable to indicate 

 any intermediate forms between the genus Sycetta and the 

 Homocoelidfe. It presumably arose from some homoccel ancestor 

 which formed colonies by radial budding, not unlike those of 

 Dendya, but the apical position of the nucleus of the collared 

 cells and the much more advanced type of skeleton shew that 

 the relationship to Dendya itself cannot be a close one, while the 

 interval to be bridged over between the most primitive Sycetta 

 and any Leucosolenia is a very wide one. Moreover, Sycetta 

 itself seems to be in the nature of a cul-de-sac, for the entiie 

 absence of the characteristic oxeote spicules of Sycon makes it 

 doubtful whether we can derive the latter genus directly from 

 the former, though both have probably sprung from some common 

 ancestor. If, however, Sycon derives its oxea from an ancestral 

 Leucosolenia, it is difficult to account for the absence of these 

 spicules in Sycetta, but the distribution of oxea in the Calcarea 

 is an extremely difficult problem about which we have perhaps 

 said enough in an earlier part of this paper. 



The fact that certain species of Sycon, for which von Lendenfeld 

 [1885 A] proposed his genus Homoderma, retain the collared cells 

 as a lining to at any rate a portion of the central gastral cavity 

 throughout life, certainly shews that one can draw no hard and 

 fast line of distinction between the Homoccelidte and the old 

 group Heteroccela in this respect, but the forms in question have 

 such a highly specialised syconoid skeletal system that they 

 hardly help us to bridge over the interval between the Homoccelidae 

 and the Sycettidte. 



The family Sycettidss is a very small one, the typical genus 

 being Sycon with a large number of species, while the only other 

 known genera are Sycetta and Sycandra, each with a very small 

 number of species and each representing an ofishoot which 

 probably leads no further. From the Sycettidse two lines of 

 descent appear to lead to the Heteropiidse and Grantiidas 

 respectively. 



In both these families the important step in further evolution, 

 as in the Leucascidfe and the Leucaltidas, has been the develop- 

 ment of a dermal cortex, but this coi'tex appears to have arisen 

 somewhat differently in the two cases. In the Heteropiidse it is 

 clearly associated with the out-turning of certain of the oral rays 

 of the distal tubar triradiates so as to arch over the entrances to 

 the inhalant canals. We may assume that with these rays the 

 dermal tissues of the sponge have spread over the intercanals 

 and have given rise ultimately to the special cortical spicules 

 developed in situ. The rotation of the distal tubar triradiates 

 in the manner indicated, and the pi'eponderating development 

 of the now centripetally directed oral raj^s, have finally converted 

 these spicules into the " pseudosagittal " triradiates which 



