512 Ms-. E. Kiikpatrick on 



failed to reveal any axial core of spicules, tliougli often there 

 were appearances strongly suggesting the presence of such 

 objects ; but on focussing, these ghosts were seen to be optical 

 illusions. 



Though there are no axial spicules present, there is justifi- 

 cation for putting forward the theory that the sponge 

 originally had a skeleton of quadriradiates joined and 

 enveloped with cement; the spicules would be orientated 

 with their odd ray in a vertical plane and basal rays in a 

 horizontal plane. Each vertical pillar in the framework, as 

 it now stands, is simply the point whence three wings radiate 

 out, at angles of 120°. 



TheLithoninre may be compared to Chalinid Sponges, and 

 the Merlinai to Horny Sponges which have lost an axial core of 

 spicules, which presumably they had, in many instances, for- 

 merly possessed. The latter analogy, however, is not quite 

 perfect, because Merlia has proper spicules. These consist of 

 tufts of long slender tyles which pass up obliquely from the 

 floor of the uppermost " cells " to the membranous roof and 

 spread out so as to support it. These spicules are U7 /x long^ 

 1*7 [JL thick ; the heads being oval, and 5 ^ long by 2 /a broad ; 

 they are slightly curved and tajjer to a sharp point. The 

 rhaphides lie in the dermal membrane; they are 55 /u, long, 

 very slender and curving to a fine hair-like extremity. 

 Microrhaphides slender, crescentic, 15 //- long, also lie in the 

 dermal membrane, and are ))robably distinct spicules, and 

 not broken off ends of the longer rhaphides. 



Tuning-forks are only 52 yu. long, tlie shaft being 32"5 /u, 

 and prongs 19"5 ytt long. 



The skeleton of Merlia is composed of calcite. 



The occurrence of the radial fibrillar cement in the 

 Lithonina3 and Merlins led me to hope that some light would 

 be thrown on that extremely aberrant form Astrosclera 

 toilleyana, Lister (5. p. 459). A dried fragment of the 

 macerated skeleton of Astrosclera, at first sight, looks not 

 unlike a similar fragment of Minchinella. In each there is 

 a firm porous skeletal reticulum, and in each the surface of 

 the strands of the network present a finely punctate or granular 

 appearance due to the projecting of the ends of fine fibrils 

 above the general surface. 



In Astrosclera the scleres are in the form of polyhedral 

 blocks, which are formed in the interior of scleroblasts ; and 

 further the skeleton is made of aragonite. 



The possibility has suggested itself to me that these scleres 

 are of the nature of cement blocks, which have originally 

 been deposited round an axial skeleton that has disappeared, 



