822 W. J. SOLLAS ON THE STRUCTURE AND 



another species has come to light from the Philippine Islands, vase- 

 like in shape, similar to Siphonia expansa mentioned at p. 805, 

 Mr. Carter says : — 



" Schmidt's Corallistes poly discus (Atlant. Spongienf. p. 24, Taf. iii. 

 figs. 8 & 9) appears to me, from the form of its surface-spicules, to 

 be a different species, according in this respect with a large vase- 

 like specimen from the Philippine Islands that I have lately been 

 examining, in which, however, there is, in addition to the acerate 

 flesh-spicule, a small solid one of an elliptical form like that charac- 

 terizing Pachastrella abyssi, while the acerate flesh-spicule in all is 

 almost identical with that of Macandrewia azorica" 



Now, as further on (p. 464), the latter is considered equal to 

 Corallistes clavatella, Sdt., the close alliance of this Lithistid to 

 Discodermia, which I have before noticed, is thus corroborated; 

 and, as this would give a discoid spicule with more indented, tooth- 

 like margin, then the finding of the one in S. costata just mentioned 

 might be accounted for. 



The previous existence of a dermal skeleton of some kind about 

 the exterior of Siphonia appears probable also from the exposure of 

 the outermost excurrent canals as grooves radiating from the cloacal 

 tube of many Haldon specimens ; these grooves must once have been 

 covered in by a membrane of some sort in order to form completed 

 tubes. 



Two other forms of spicules also occur in Discodermia : — one the 

 flesh-spicule already alluded to, so minute that it has little or no 

 chance of surviving the changes of fossilization, and, indeed, is found 

 to have already disappeared in recent deciduous specimens of Disco- 

 dermia before fossilization has set in ; this consequently we do not 

 and cannot expect to find in Siphonia ; the other, not mentioned by 

 any preceding writer, is a long, straight, or curved acerate spicule*, 

 0-072 long and 0-002 diam. (PL XXVI. figs. 6 & 6 a), and tolerably 

 abundant in my pieces of Discodermia, and a characteristic spicule 

 of our Siphonia (p. 808, PI. XXVI. figs. 5 & 5 a). 



From the preceding comparison it will be seen that the very 

 closest resemblance exists between the recent Discodermia and the 

 fossil Siphonia. Not only in fundamental structure, but also in 

 general form, and in the arrangement of the canals and elements of 

 the skeleton, the two are, in a broad sense, the same. Thus the two 

 puzzles of the Cretaceous sponges are now cleared up, the Ventri- 

 culites survive in Myliusia Gray if, and the Siphonia; in Discodermia 

 polydiscus. 



* This spicule, Mr. Carter tells me, is equally common in the British-Museum 

 specimen and in that from the Philippine Islands, also in the Lithistids 

 Theonella Swinhoei and Azorica Pfeiffera ; it runs throughout the whole group 

 of the Lithistina, and may, as Mr. Carter proposes, be conveniently termed 

 the " beam spicule." Bocage figures from D. polydiscus another large form of 

 spicule (I. c. pi. xi. figs. ld& le), which, however, differs from the beam spicule 

 in being entirely superficial in position, and in having one end rounded off. 

 Owing to its superficial position it would almost certainly be detached from 

 the sponge before fossilization; and hence we cannot hope to find its representa- 

 tive in Siphonia. 



t Ann. & Mag. Nat. Hist. ser. 4, vol. xix. p. 121, pi. ix. figs. 8-17. 



