408 M. K. A. Zittel on Fossil HexactinelUda. 



edges of a regular octahedron. From the size of this space, 

 enclosed by solid siliceous rods, the strength of the octahedral 

 edges, and the more or less regular or distorted form of the 

 lumina taking the place of the lateral faces of the open 

 octahedron, valuable systematic characters are to be obtained. 

 When the state of preservation is good, it may be ascertained 

 with perfect certainty that the axial canals pass uninterruptedly 

 through the cavity, and form a very elegant axial cross within 

 it. But as their siliceous envelopes are very thin they are 

 easily destroyed ; and the octahedral nodal points then acquire 

 the hgure described by O. Schmidt in ^cyphia striata^. 



Among living Hexactinellida we are acquainted with lat- 

 ticed frameworks with such octahedral nodal points (lantern- 

 spicules) only in Myliusia Grayi^ Bow., and Myliusia ZitteJu^ 

 Marsh. MS. ; in the fossil forms, however, they are widely 

 diffused [Coeloiitychium^ Ventriculites, Becksia, Plocoscyphiay 

 Pachyteichisma, &c.). 



A peculiar phenomenon is presented in the genera Farrea, 

 DactylocalyXj and Aplirocallistes by the exceedingly minute 

 sexradiate stellules, which are firmly attached by one ray to a 

 siliceous fibre of the latticework, but otherwise resemble the 

 other sexradiates of the framework in every respect, and also 

 possess fine axial canals, whilst elsewhere canals are never to 

 be observed in the root-like processes or spines of the siliceous 

 fibres. It is possible that these stellules, with which I am ac- 

 quainted also in some fossil Hexactinellida, may be young, still 

 undeveloped spicules ; but possibly they fulfil the function of 

 the flesh-spicules, although they do not lie free in the sarcode. 

 Bowerbank figures such small spicules in several species of 

 Farrea ; Carter regards them as after-formations. 



In general, the differences presented by the latticed frame- 

 work of the fossil Hexactinellida are so considerable that in 

 many cases we can determine generically even a small isolated 

 fragment or a preparation. The size of the meshes, the octa- 

 hedral or solid constitution of the crossing-nodes, the orna- 

 mentation of the siliceous fibres with prickles, spines, or root-like 

 processes, as well as the arrangement of the coalescent sex- 

 radiates themselves, furnish important systematic data, which 

 vary but little in the same species or even genus. 



Nevertheless the one-sided consideration of the microstruc- 

 ture of the framework would lead to no satisfactory result. 

 Many forms differing greatly in other characters possess an 

 almost exactly accordant skeletal structure. Thus W. Mar- 

 shall has already shown that the framework of Eurete agrees 



* Atlant. SpoDg. Taf. ii. fig. 16, 



