the Oreensand compared with those of existing Species. 117 



are borne out in this conclusion by reference to those which I 

 have cleKneated. 



It will be observed, as before stated, that they are for the 

 most part all large — that is, the largest spicules of the species 

 to which they belonged ; and therefore, if we compare them 

 with living species, it must be with the larger spicules of these 

 species. 



There is no question, then, as to whether we shall take our 

 characters from the large spicules of the latter for this purpose 

 in preference to the small or minute ones — for, as before stated, 

 most sponges contain two sizes (that is, the large and the 

 minute — those which can be seen for the most part by the un- 

 aided eye or a low magnifying-power, and those which require 

 the aid of a very high microscopic one) — since, as also before 

 stated, the latter appear to be entirely unrepresented among 

 these fossil spicules. 



Om- characters, then, among the fossil spicules (for it should 

 be remembered that we have no entire sponges here) must, in 

 common with their living allies, be taken from the largest 

 spicules generally ; and such we shall observe to be chiefly 

 confined, in the latter, to the periphery, where their shafts are 

 provided with heads which meet together externally, and thus 

 form a shield-like surface to the sponge. 



These heads, like the heads of so many nails, present forms 

 which are peculiar to the species, and are developed inversely 

 to the shafts ; that is to say, the more expanded the head the 

 shorter the shaft, and vice versa. (The position of the shaft, 

 branch, or arm where broken off is always indicated in the 

 figm'es by a little circle, which is the axial canal, within a 

 larger circle, which is the circumference of the spicule.) 

 Hence, the shaft only differing in length or size, we must look 

 to the head for the character ; and here we shall at once see 

 that, . whether we take the simple trifid or ternate one in 

 fig. 36 &c. PI. IX. (existing species, Geodia)^ or the hex- 

 ternate one fig. 30 (that is, the dichotomous division of the 

 trifid or ternate head, ex. sp. Stelletta) , or once more divided 

 (dodecateniate^ as in Dactylocalyx Boioerhankii^ Brit. Spong. 

 fig. 53), or still more divided (polyternate, as in fig. 1 — ex. sp. 

 Dactylocalyx McA)idreivii, Bk., McAndrewia azorica^ Grray, 

 Proc. Zool. Soc. Lond. 1869, pi. v. fig. 3), or where the divisions 

 are more or less united into a disk, figs. 3, 4, 5 (ex. sp. 

 Dactylocalyx polydiscus^ Bk., Proc. cit. 1869, pi. vi. figs. 10 

 &c.), we observe from Schmidt's and Du Bocage's figures of 

 D. polydiscus, respectively copied into our Plate VII. figs. 7, 

 8, and 9, which are confirmed by Dr. Bowerbank's fig. 102 

 (Brit. Spong.), and my own actual observation of the spicula, 



