54 BRITISH STROMATOPOROIDS. 



As the astrorhizas are mere grooves on the surface of the last-formed layer of 

 the skeleton, in their typical condition at any rate, it follows that they are not only 

 present on the free surface of the colony, but also on the surface of successive con- 

 centric lamina? ; since each lamina in turn constituted for a time the actual surface. 

 As, however, each successive lamina is produced, the astrorhizal grooves on the 

 surface of the lamina below necessarily become roofed over by the new layer, and 

 are thus converted, in all the parts of the skeleton below the surface-lamina, from 

 open grooves into canals. Hence, in vertical sections of such Stromatoporoids as 

 possess astrorhizas, the cut ends of the astrorhizal canals appear in the section at 

 various points as larger or smaller round apertures (Plate V, fig. 6, and Plate XI, 

 figs. 12 and 14). They are, however, necessarily without any proper walls, their 

 lower margin being formed by the lamina to which they belong, their upper margin 

 by the lamina next above, and their sides by the irregular radial pillars which 

 connect these two lamina?. 



There is one Stromatoporoid, viz. Stromatopora dlscoidea, Lonsd. (=8. elegans, 

 Rosen), from the Wenlock Limestone of Sweden, Esthonia, and Britain, in which 

 the astrorhizal canals seem to depart in important respects from their ordinary 

 form. The elucidation of the true structure of this singular type is attended with 

 unusual difficulties, as, for some reason difficult to explain, most specimens have 

 undergone a more or less complete secondary crystallisation of their skeleton, even 

 when the surface- characters are retained in admirable preservation. Superficially 

 regarded, 8. dlscoidea, Lonsd., is remarkable for the generally large size of the 

 astrorhizas, for the minute subdivision of the main channels, and for the extremely 

 perfect inosculation established between the entire system of astrorhizas (Plate III, 

 fig. 3). Thin sections show, however, that the astrorhizal canals are not, as usual, 

 mere shallow grooves on the surfaces of the successive laminae, but that the laminas are 

 obsolete, and the astrorhizal canals are converted into comparatively deep channels, 

 with perpendicular sides which extend downwards through the thickness of each 

 successive " latilamina." In other words, beginning as open grooves on the surface 

 of the primitive crust, the sides of the astrorhizal canals grew upwards to form so 

 many deep narrow channels with vertical walls, these channels extending through 

 the whole of the first " latilamina." When the second " latilamina " is formed, new 

 astrorhizal grooves are produced, which pass through the same process of develop- 

 ment ; and so on through the entire system of " latilaminas " of which the skeleton is 

 made up. Moreover, the tabulate zooidal tubes open into the sides of these deep 

 channels, and are, in fact, confluent with them. Perhaps, therefore, the most correct 

 way of regarding the astrorhizal grooves of 8. dlscoidea, Lonsd., would be to consider 

 them as really formed by the serial junction of the zooidal tubes in sinuous lines, 

 much as we see in the serially-united polypes of Diploria and other types of Corals. 

 Be this as it may, the result of the peculiar constitution of the astrorhizal system 



