356 Mr. n. J. Carter on the Hexactinellidee. 



lis contracted below and only expanded above (fig. 10) ; or 

 sigmoid, clavate, capitate, and toothed outwardly, with the 

 lower third or inward curve of the ray diminished to almost 

 notliing, arranged en fleur-de-lis (fig. 11, a, Z*) ; or sigmoid 

 and without head, subulate, with flexed exti-emity diminishing 

 to almost nothing in the lower third or inward curve, also 

 arranged en fleur-de-lis (figs. 12 & 15): — the pappiform 

 rosette with flexed ray. Here, too, the ends of tlie arms are 

 conicaUy inflated and tubercled for the support of the rays 

 generally, ending on the summit in a straight coarse spine 

 (fig. 15, c). The lowest portion of the ray being stouter than 

 that which follows it, is frequently left attached to the end 

 of the arm both here and in the foregoing form, as shown in 

 fig. 15, d. 



Then there is the rosette with elongated axis, shaft-like 

 and sti-aight spines or rays, sometimes pointed, sometimes 

 capitate (figs. 17 & 18), sparsely associated with the globular 

 forms (figs. 6 & 7) which characterize Aphrocallistes Bocageij 

 and which, with fig. 19 from Aphrocallistes heatrix^ whose 

 spines are also sometimes capitate, appear to be the transi- 

 tional forms between the globular rosette on the one hand 

 and the large spined shaft (fig. 20) peculiar to Aphrocallistes 

 heatrix on the other, the globular fonn being altogether 

 absent in the latter. 



Also the birotulate spicule (fig. 21), consisting of a straight 

 subspined shaft with eight separate recurvol blades, which 

 is the flesh-spicule of Hyalonema &c,, but not unfrequently 

 assumes the sexradiate form (fig. 22), when it becomes a 

 "rosette." There is also a four-armed variety, so that the 

 ^fcirotulate spicule may be said to be two-, and occasionally 

 four- or six-headed ; while three similar forms occur in 

 Hyalonema on which there are no heads, but where the 

 ends (and sometimes the whole surface) are covered with thick 

 short vertical spines. (See Dr. Bowerbank's Brit. Spongiad^e, 

 vol. i. pi. vi. figs. 153-157.) 



Lastly, as regards varieties of the " globular rosette," the 

 rays are sometimes once-branched (fig. 13) ; sometimes the 

 head of the ray is echinated generally with spines (fig. 13, a) ; 

 in others the arms alone have become enveloped by a globular 

 vitreous mass (figs. 23 & 24), which seems to be the transi- 

 tional state to the stellate and siliceous ball of the Geodinid^e 

 &c., also flesh-spicules. There are, again, many slight modifi- 

 cations of the illustrations themselves, but fundamentally no 

 more distinct forms in the Hexactinellidse that have as yet been 

 brought to notice than those above described and delineated. 



As regards size, there is of course not only a difference 



