AFFINITIES OF THE GENUS SIPHONIA. 821 



nate botryoidally, or hi the form of a bunch of grapes, which unites or 

 interlocks with that of the neighbouring branch, and thus the internal 

 structure is formed, except at the surface, where the branches imme- 

 diately under the dermal layer of disks &c. terminate respectively in 

 flat filigreed or dendriform expansions, which do not intermingle 

 with those of oj>posite branches." 



Owing to the kindness of Mr. Carter, who has given me fragments 

 of D. polydiscus, I am able to represent the " staple spicules " 

 (PI. XXVI. Tig. 2) of the preceding paragraph side by side with those 

 of Siphonia, so that any one who so wishes may judge of their re- 

 semblance for himself. In the same Plate (figs. 9 & 9a) will be ob- 

 served some of the " disciform spicules " from the dermal skeleton of 

 Discodermia, specimens of which, judging from all analogy, we ought 

 to find in connexion with Siphonia, more especially as Mr. Carter 

 has found such spicules entangled in the interior of rolled, dead 

 fragments of Discodermia dredged up on board the ' Porcupine ' 

 near Cape St. Vincent (I. c. p. 463). No endeavour on my part, 

 however, to discover these, either on the surface or in the interior of 

 my specimens of Siphonia, has met with success, which may arise from 

 the extreme thinness and thus perishable nature of the disk. The 

 nearest approach is the form represented on PI. XXVI. fig. 11, a single 

 specimen of which was found imbedded among the staple spicules just 

 on the edge of one of the excurrent canals of /S. costata. This form, 

 however, more nearly resembles the dermal spicules of other Lithis- 

 tids, ex. gr. C. clavatella (0. S.), and is most likely a stray waif 

 washed in. Still it is quite possible that spicules with entire (PI. 

 XXVI. fig. 10) and with divided (PI. XXVI. fig. 11) margins may 

 exist together in the dermal membrane of the same sponge ; and I 

 believe that an instance of this has already been found. 



True disciform spicules (PL XXVI. fig. 10), both with simple and 

 sinuated borders like those of D. polydiscus, have been met with by 

 Mr. Carter* abundantly scattered loose in the spicule-bed of the 

 Haldon Greensand, and have been named by him Dactylocaly cites 

 Vicaryi ; and I have myself obtained them in great numbers asso- 

 ciated with pieces of Lithistid network of the Discodermia type from 

 the Upper Chalk of Trimmingham, in Norfolk. 



The coating of disciform spicules forming the dermal skeleton of 

 Discodermia is readily separable from the internal network w T hich it 

 covers and conceals ; the connexion between the two is, indeed, of the 

 very slightest kind ; and hence, if Siphonia ever possessed a similar 

 coating of disk-like spicules, it would in all probability lose it after 

 death and during decay; and the disk-like spicules separating from 

 the sponge and from each other would be buried in the surrounding 

 silt or sand, to be discovered subsequently as isolated spicules like 

 those which have occurred to Mr. Carter in the Haldon beds and to 

 myself in the Chalk. 



Here I must quote the last paragraph in Mr. Carter's description 

 of Discodermia polydiscus (Annals, t. c. p. 464) ; for it is evident tha 



* Ann. & Mag. Nat. Hist. ser. 4, vol. yii. p. 123, pi. vii. figs. 1,2, kG. 

 Q. J. G. S. No. 132. 3 u 



