Merlia norraani, Kirkp. 239 



wliicli siliceous spicules are present, growing in a calcareous 

 honeycomb-like framework constructed of vertical tube.-? with 

 perforate horizontal tabulae. The upper part of Merlia re- 

 sembles an ordinary siliceous sponge. The lower part in 

 the "crypts" * or honeycomb spaces is composed mostly of 

 very large granular cells, more or less separated by connective- 

 tissue cells, the whole being in continuity, through the 

 central orifice usually present in each tabula, with the tissues 

 of the upper part of the sponge. 



Further, the masses of crypt-cells are surrounded by a kind 

 of epithelium. The large crypt-cells have nut grown down 

 into the crypts, but have developed in the position in which 

 they are found : for I have seen, by a process of grinding 

 down, crypts at the base of the sponge full of crypt-cells, and, 

 after having destroyed the latter with Eau de Javelle, have 

 oiten seen the orifice in the roof of the crypt, i. e. in the 

 tabula, narrowed to the smallest slit sometimes no more than 

 1 /x in width. As some of the crypt-cells are about 200 /u, in 

 area end on, it would be ridiculous to assume that these huge 

 masses of huge cells have grown down through five or six 

 stories and squeezed their way through slits sometimes only 

 1 /x in diameter. Moreover, the crypt-cells are not loose and 

 isolated, but enmeshed in a network of connective tissue and 

 surrounded by a kind of epithelium. 



In some ground-down sections of the skeleton I have seen 

 tabulae absolutely imperforate, but I had not observed whether 

 the crypts thus roofed over had contained cells or not. 



Accordingly Merlia is not a parasite growing over and 

 into a calcareous organism, and it is not a Foraminiferan. 

 Further, it is certainly not a Coelenteiate or a Polyzoon. In 

 fact, it is now beyond a doubt seen to be a sponge and no- 

 thing but a sponge. 



Merlia is of unique interest, not only in itself, but also 

 on account of the resemblance, in some respects, of its calca- 

 reous skeleton to certain of the Monticuliporas. 



Not only is there, on surface view, a polygonal reticulum 

 with tubercles at the nodes of the rneshes, and not only are 

 there tabula? with a central hole, but there are radial sutural 

 markings precisely as in Rhaplddopora (Chcetetes) stromato- 

 poroides, lioemer (Nicholson and Foord, "On a new Genus 

 of Devonian Corals/' Annals & Mag. N. 11. (5) xvii. 1886, 

 p. 393, pi. xvi. tig. 5). Of course these resemblances may be 

 merely homceomorphous. 



* The useful terms " crypt " and " crypt-cell "were suggested by Prof. 

 E. A. Minchin. The cells, which look like large granular gemmule 

 cells, appear to me to be probably " calcoblasts." 



