Mr. II. J. Carter on the Hexactinellidai. 355 



rnately mixed up with the skeleton-spicules of the sponge 

 without ever being taken into the structure of the latter, such 

 is the case. 



Among the vitreous Hexactinellidas, however, there is a 

 sexradiate spicule of the skeleton-structure so minute and so 

 plentiful that it might easily pass for a " rosette," were it not 

 for the following differences, viz. that it never goes beyond the 

 simple or first sexradiate division (that is, where the distin- 

 guishing form of the rosette, viz. the rays, commence), and is 

 always more or less involved in or cemented to the vitreous 

 skeleton (PI. XIII. fig. 1). Thus it often appears in Ajjhrocal- 

 listes Bocagei and in the Farrew, as will hereafter be seen, 

 where, although essentially a skelcton-spicule, it is so much 

 smaller than that which formed the basis of the original fibre 

 that I cannot help viewing it as an " afterformation." Be 

 this as it may, it is so mixed up with the rosette in Farrea 

 infundibulifbrmis, one of the "new species" to be described 

 hereafter, that it is desirable to notice the existence of such a 

 spicule in order that it may not be confounded with the 

 rosettes in any sexradiate sponge, besides those mentioned, in 

 which it may be found to occur (PI. XVII. figs. 2-4). 



From this we pass to a consideration of the " rosette " 

 itself (in connexion with our illustrations), which is essentially 

 a flesh- spicule of the kind above mentioned, but of course 

 subject to modifications. 



The simplest form appears to be that in which the six 

 arms are long and respectively terminated by two divergent 

 pointed rays (PI. XIII. fig. 2) ; or the arms may be short with 

 dual rays (fig. 3) ; or the arms short with a variable plurality 

 of the same kind of rays (fig. 4). 



(Here it should be noticed that, for the sake of perspicuity, 

 the fifth and sixth arms (fig. 2, c c) — that is, the third axis — 

 will not be introduced in the illustrations after fig. 2.) 



Then the dual ray may be straight and capitate, when we 

 get the form fig. 5, where the spines of the head are few and 

 long (fig. 5, a, b, c) ; or the arms, still short, may have straight 

 capitate rays in variable plurality (fig. 6), where the spines of 

 the head are short and numerous (fig. 6, a, b) ; or the arms may 

 be long with the same kind of rays (fig. 7) ; or the arms variable 

 in length in different rosettes, and furnished with a multitude 

 of straight capitate rays of unequal length in the same group 

 (figs. 8 & 16) : — the pappiform variety. Here the ends of the 

 arms also are conically inflated and provided with tubercles, 

 each of which supports a ray. 



Or the rays may be sigmoid, capitate, and arranged en 

 fleur-de-lis expanded generally (fig. 9) ; or the same en flevr-de- 



