160 CALCAEEOUS SPONGES. 



iu Tremacystia (VerticiUites) D'Orbignp, H., and Corynella rugosa, H., and traces of 

 it are present in some species of Eudea, Eusiphonella, and Lymnorea. The other 

 kind of dermal layer has the appearance of a compact smooth or rugose membrane, 

 resembling in aspect the epitheca of a coral, which envelopes the sides and some- 

 times the greater part of the surface of the sponge. The structure of this membrane 

 is very rarely preserved ; but, according to Dunikowski, it is mainly composed of 

 very irregular three-rayed spicules. 



The genera into which Zittel has divided this family are mainly based on diffe- 

 rences in form and in the canal-structure, and not primarily, as in the case of the 

 Siliceous sponges, on the spicular characters. This is owing to the fact that in many 

 forms the spicular constituents of the fibres have been completely obliterated by 

 fossilization, and notably is this the case with most of the specimens from Triassic 

 strata ; so that an attempt to establish the genera on similarity of minute structure 

 can be, at the best, only partially successful. Where, however, the spicular characters 

 of the type species of a genus can be ascertained, it is desirable to include in it only 

 those sponges with a similar minute structure, and thus constitute the genus on a 

 natural basis. Fossil calcareous sponges, like siliceous forms, very frequently are 

 similar in form and in the canal-structure, whilst the minute components of the 

 fibres, when examined under the microscope, are found to vary very considerably. 

 Dunikowski has proposed to subdivide the family according to the various combina- 

 tions of uniaxial, three-, and four-rayed spicules, in the same manner as Haeckel has 

 divided the recent Leucones ; but I have found it extremely difficult in practice ta 

 determine in thin sections whether many spicules are uniaxial or three-rayed forms, 

 and it is equally difficult to decide whether a spicule in which three rays are 

 s)iown in the section may not have possessed a fourth ray, so that this arrangement 

 appears to me not to be available for fossil sponges. 



There are, however, in the Pharetrones which I have examined, some clearly 

 marked types of spicular structure which may be briefly mentioned. In the first of 

 these, which I propose to call the " Corynella " type, the fibre is nearly entirely com- 

 posed of subequal, elongate, filiform, three-rayed spicules, arranged parallel to each 

 other in the direction of the axis of the fibre. In these spicules the basal ray is 

 so slightly developed that in transparent sections they may easily be mistaken for 

 uniaxial forms. Relatively large three- and four-rayed spicules are rarely present in 

 the fibre, though they form, in some species at least, an open dermal layer on the 

 surface of the sponge. This type of spicular structure is well developed in the 

 Cretaceous species of Corynella and in Tremacystia. In a second type, which may 

 be designated the "Sestrostomella" type, the skeletal fibres are built up of relatively 

 large three- and four-rayed spicules disposed in the central portions of the fibre and 

 surrounded by smaller and very irregular forms, which frequently appear as sinuous 

 lines forming the exterior border of the fibre. This structure is characteristic of 



