44 BRITISH STROMATOPOROIDS. 



As regards the " radial pillars," the following are the chief variations to be noted : 

 — In the genus ClatJirodictyon, Wich. and Mur., the radial pillars are incomplete, 

 or even almost obsolete as distinct structures. In some forms of the group, such as 

 G. regidare, Rosen, the radial pillars are confined to their respective interlaminar 

 spaces, running from their lamina of origin to the lamina next above, but not being 

 continued through more than one interlaminar space (Plate V, fig. 1). In G. stria- 

 tellwm, D'Orb., a nearly allied Silurian species, the pillars are not only confined to 

 their respective interlaminar spaces, but many of them are imperfect, and fall short of 

 the lamina next above that from which they arise (Plate V, fig. 3). In the more 

 typical species of Clathrodidijon, such as G. cellulosum, Nich. and Murie, 

 G. vesiculoswm, Nich. and Mur., G. variolare, Rosen, and G. fastigiatum, n. sp., 

 the concentric lamina? are crumpled into numerous minute undulations, which 

 become continuous with the radial pillars. In these cases, therefore, the radial 

 pillars become largely confounded with the concentric laminae, the appearance 

 exhibited by vertical sections (Fig. 2, a, and Fig. 3, B, and Plate V, figs. 5 and 6) 

 being that of vesicular tissue composed of larger or smaller cells arranged in rows. 

 The radial pillars in these cases have, however, nevertheless, a real existence, as 

 shown by the fact that their cut ends can generally be recognised clearly in 

 tangential sections (Fig. 2, B, and Fig. 3, a). 



In the genus Labechia, B. and H., the radial pillars reach their maximum 

 of development, being exceedingly stout, pointed at their free ends, and, as a rule, 

 continuous from the epitheca to the upper surface (Fig. 5). In this genus, also, as 

 to a less extent in some species of Clathrodidijon, adjoining pillars may become 

 closely united by their sides, thus giving rise to short flexuous rows, or sometimes 

 (as in L. alveolaris, n. sp., from the Wenlock Limestone) to a reticulated tissue not 

 very unlike that of such " Tabulate Corals " as Alveolites. 



As regards the genus Actinostroma in particular, and, indeed, as regards the 

 Stromatoporoids generally, much question has arisen among different observers as 

 to whether the radial pillars are hollow or solid. The earlier observers generally 

 believed them to be the former ; later observers, working mostly with thin sections, 

 have generally maintained the latter view. For my own part, I have previously 

 regarded the radial pillars as being solid; but more extended observations have 

 shown me that this is certainly not invariably the case. In some forms (e. g. in 

 certain species of Actinostroma and Glathrodictyon) no traces of any central aper- 

 ture can be detected in cross-sections of the radial pillars. In other cases there is 

 clear evidence of the existence of an axial tube in the pillars. At the same time, 

 there is no ground for supposing, as was thought by many of the earlier observers, 

 that the radial pillars were inhabited by zooids, or that they are in any way com- 

 parable with the zouidal tubes of Millepora. On the contrary, it is still uncertain 

 if they were ever really opeu at their free extremities, even where, as in Labechia, 



