288 Dr. J. E. Gray on the Arrangement of Sponges. 



addition to the fusiform, clavatc, or capitate spicules of the 

 former suborder. 



This suborder consists of the family Esperiadse (Proc. Zool. 

 Soc. 1867, pp. 502, 532), except the genus Carteria, which 

 proves to be a sexradiate spoijge. Some genera have only 

 the bihamate spicules ; but in general they have bihamate and 

 contorted spicules. 



In the 'Ann. & Mag. Nat. Hist.' (/. c.) I divided this sub- 

 order into four families — Esperiadse, Desmacidonidte, Hama- 

 canthida3, and Gelliadse. 



Suborder III. Sexradiatospongia. 



Sponges with spicules of the sexradiate type, in combination 

 with the simple spicules of the Leiospongia. 



This suborder agrees with the Hexactinellidas of JMr. Carter, 

 so well described in the 'Ann. & Mag. Nat. Hist.' 1873, xii. 

 p. 357, excluding the genus Hyalonema, and adding to it the 

 genus Carteria, which Mr. Carter partly describes and figures 

 under the name of Hyalonema. 



In the paper in the ' Proc. Zool. Soc' for 1867 I formed the 

 order Coralliospongia, and in the 'Ann. & Mag. Nat. Hist.' 

 1872, ix. p. 447, I considered "the hexaradiate spicules the 

 essential character of the order" (seep. 449). It has since 

 been found to contain a series of sponges which belong to the 

 quinqueradiate type and are placed separate by Mr. Carter, a 

 correction which I gladly acknowledge and adopt. 



The radiating spicules which form the skeleton of the body 

 are frequently of large size, and are modified in form accord- 

 ing to the position which they occupy in the body. The lateral 

 rays in perfectly developed spicules are generally equal ; but 

 at other times one or more of the rays are only partially de- 

 veloped. The shaft or centi-al axis is frequently elongate, 

 much longer than the lateral rays ; but when the spicules are 

 on the surface of the sponge the outer end of the shaft is more 

 or less imperfectly developed, sometimes reduced to a single 

 tubercle, and the other end of the shaft on the inside of the 

 sponge is frequently much lengthened. The two ends of the 

 shaft are almost always simple ; but the lateral rays, especially 

 of the quinqueradiate type, are frequently forked. 



The sexradiate spicules which are scattered in the sar- 

 code, and consequently free from each other, are always of a 

 much smaller size than the skeleton-spicules, and of uniform 

 structure, and nearly all of the same size. Mr. Carter has 

 named them rosettes. They differ in the form of their six 

 rays — which are always equal, sometimes simple, ending in a 



