W. J. Sollas — A New Cretaceous Sponge. 399 



and consequently the meshes of the net are each about xo^o- inch 

 long on each side. 



Mr. Carter^ has shown that the siliceous "fibres" of the Vitreo- 

 TiexactinellidcB are produced by the cementation of a number of 

 hexactinellid spicules together by a coating of siliceous matter: 

 and in fact that just as acerate spicules are imbedded in keratose 

 fibre to form the skeleton of such sponges as the Ghalinidm, so 

 hexactinellid spicules are imbedded in siliceous fibre to produce the 

 rigid skeletons of such forms as Farrea, Aphrocallistes, Dacty- 

 locahjx, Euplectella, and their allies ; and it naturally follows from 

 this, that so long as the sexradiate spicules retain their rectangular 

 tri-axial forms, the fibres built upon them will present, wherever a 

 spicule occurs, a similar regular sexradiate knot ; and what appear to 

 be points of intersection of three crossing fibres are really points where 

 the six arms of a spicule regularly radiate. Thus, in the sponge under 

 consideration, each rectangular intersection of the linear depressions 

 represents the place where siliceous fibre once formed upon the four 

 arms of a sexradiate spicule at right angles to one another ; the 

 other two arms are invisible, one because it passes directly into the 

 " coprolite," the other because it, in common with the outer half of 

 the cast of the fibres, has been worn away while the fossil lay 

 exposed on the sea-floor. 



The network arising from the union of the spicules might be 

 very irregular, if the latter were dispersed without any arrangement, 

 as, for example, in Dactylocalyx (Stutchbury) ; if, however, they be 

 grouped according to a general plan, a very regular network may 

 result, as in Farrea occa, and the dermal reticulation of the sponge 

 under consideration. In the latter the fibres exposed on the surface 

 can be arranged in two groups, as will now be described. 



At the base of the sponge are four nearly equidistant centres, from 

 which the fibres regularly radiate. Starting with one of these centres, 

 we may trace a set of fibres passing from it, and diverging from one 

 another as they go ; the more medial fibres of the group ascend 

 the side, cross directly over the summit, and descend on the oppo- 

 site side; those which lie nearer the sides of the group simply 

 curve round the sides of the sponge ; but all agree in finally con- 

 verging to the centre facing their point of origin. Thus all these 

 fibres are simply loops of different lengths and obliquity connecting 

 two opposite centres. 



The fibres from the other pair of centres are disposed in exactly 

 the same manner, and the two sets of loops thus come to be placed 

 at right angles to each other, and so to produce a regularly rectangular 

 meshwork. This is well shown in the accompanying woodcut, 

 which is a map of one side of the Museum specimen ; it is not 

 drawn to scale, and only professes to be accurate so far as it shows, 

 in the area represented, every fibre and the intersection of every 

 fibre in their true relations. The fibres from the centre ' A ' may 

 be observed on the left passing in curves towards the centre ' 0,' 

 which is here only obscurely indicated, and as they do so, the fibres 

 which curve in a similar way from ' B ' to ' D ' may be seen to 

 ^ Ann. and Mag. Nat. Hist., vol. xii. pp. 46-49. 



