208 PEOr. H. A. KICHOLSON AND DE. J. MURIE ON THE 



or radiating canals or grooves which are approximately horizontal, 

 and which are arranged in minor systems round numerous secon- 

 dary centres. These canals or grooves, however, as we shall here- 

 after show, have no distinct walls, are of comparatively large and 

 not microscopic dimensions, and perforate the horizontal laminae 

 obliquely, a large portion of their course being thus in the inter- 

 laminar spaces. In fact they appear to be often confined to a single 

 interlaminar space, and thus to be more of the nature of branched 

 grooves on the surfaces of the horizontal laminae than actual canals. 

 In no case, certainly, do these canals run in the horizontal laminae, 

 BO that they do not constitute a canal-system comparable to mi- 

 croscopic tubuli of certain of the Foraminifera. Principal Dawson 

 (' Dawn of Life,' p. 159) describes a specimen of Coenostroma in 

 which " the plates are traversed by tubes or groups of tubes, which 

 in each successive floor gave out radiating and branching canals 

 exactly like those of Eosoon, though more regularly arranged." 

 We have not, however, seen any traces of these in anj of the 

 specimens which we have examined ; and it is possible that Dr. 

 Dawson is merely alluding to the radiating canals wliicli are 

 found in Coenostroma, Caunopora, and other forms, and which, as 

 M^e have remarked, do not properly form a canal-system in the 

 substance of the skeleton of the fossil itself. 



4. The Barcode- cliamhers. — The general form and arrangement 

 of the sarcode-chambers of the typical Stromatoporoids can be 

 best studied in naturally or artificially decalcified specimens in 

 which the skeleton has been silicified, or, on the other hand, in 

 decalcified specimens in which the skeleton has remained calca- 

 reous whilst the cavities have been infiltrated with silica. In the 

 first class of specimens we find the horizontal laminae to be sepa- 

 rated by spaces of very various width in different species, these 

 interspaces being in turn broken up at shorter or longer inter- 

 vals by the radial pillars. In the second class of specimens we 

 get casts in silica of the sarcode-chambers, and we can thus study 

 their general form and arrangement more minutely. In these 

 specimens we find that each interlaminar space, though on the 

 whole continuous, is really composed of a number of separate 

 rounded chambers, connected by wide openings and marked ofii" 

 by the radial pillars. Hence the siliceous cast of a single inter- 

 laminar space presents us with a layer of rounded oval or circular 

 siliceous masses, connected with one another on all sides by stolons 

 of silica of considerable size, and separated by small perforations 



