RADIAL PILLARS AND CONCENTRIC LAMINiE. 45 



E.and H., there is clear evidence that they were hollow internally. Even where the sur- 

 face carries perforated tubercles (as in Stromatoporella laminata, Barg., PI. X, fig. 4), 

 it remains to be shown that these tubercles are the upper ends of the radial pillars. 



In various types of Actinostroma, such as A. clathratum (Plate I, figs. 10 and 13), 

 tangential sections show that the exterior of the pillars is of a denser structure 

 than the interior. The cut ends of the radial pillars thus show a dark external 

 ring and an internal lighter space, or, in some cases, a dark outer ring and a minute 

 central clear spot surrounded by a dark ring. This appearance, which is very 

 distinct in some specimens, though not recognisable in others, would seem to show 

 clearly that the radial pillars were primitively furnished with a minute central canal, 

 which probably became largely or entirely filled up in the process of growth. 

 There is no reason to think, however, that this axial canal opened on the surface, 

 as the pillars in Actinostroma clathratum and its allies appear to end superficially in 

 blunt imperforate tubercles (Plate II, fig. 11). 



In Labechia, E. and H., similar appearances have been long since recognised as 

 existing in a still more marked form (Steinmann " Ueber fossile Hydrozoen," 

 ' Palasontographica,' 1878, Plate XII, figs. 10 and 11 ; and Nicholson, " Pal. Tabu- 

 late Corals," Plate XIV, fig. 4). Thus in tangential sections of Labechia conferta, 

 Lonsd. (Fig. 5) one can almost always detect in the cut ends of the radial pillars a 

 minute central dark or light spot, surrounded by a well-marked concentrically- 

 laminated ring ; and there is no reason to doubt that this central spot marks the 

 position of a small central canal. That the same phenomenon is much less fre- 

 quently recognisable in vertical sections, is easily explained by the fact that it is 

 necessarily only an occasional thing for the section to cut a radial pillar precisely 

 in the median plane. There is, however, evidence, as will be subsequently shown, 

 that the radial pillars of Labechia conferta have really a cribriform structure. The 

 central canals of the pillars are, in any case, of small size, and it is doubtful if they 

 are continued to the summits of the pillars. The pillars, in fact, terminate super- 

 ficially in blunt tubercles, which as a rule show no evident signs of a perforation at 

 their summits (Plate III, fig. 12). In other specimens, however, there does appear 

 to be an opening at the summits of some of the pillars (Plate III, fig. 14), though 

 whether this appearance may not be the result of weathering is difficult to decide. 



In a beautiful species of Labechia which I have found in the Devonian Limestones 

 of South Devon, and which I shall name L. serotina, a much larger and more con- 

 spicuous axial canal is developed in the radial pillars (Fig. 4). Tangential sections 

 of this species show that each of the radial pillars is traversed by a large central 

 tube, which is seen in long sections to be crossed by numerous thick, curved, 

 transverse partitions, to a large extent obliterating its cavity. I do not know the 

 upper surface of this form, and cannot positively assert that the axial canals of 

 the pillars may not sometimes be open above ; but in those pillars which terminate 



7 



