520 Dr. H. A. Nicholson and Mr. A. H. Foord on a 



ened, while a more or less conspicuous crystalline structure of 

 the wall is almost always developed. This crystalline struc- 

 ture is shown in two ways. In the first place, the walls of the 

 corallites, as seen in either horizontal or vertical sections, 

 exhibit dark and light patches, often angular, and sometimes 

 very regularly disposed (PI. XV. fig. 7 a, and PI. XVII. 

 fig. 3), the cause of this being clearly the different orientation 

 of the calcite crystals traversed by the plane of the section. In 

 the second place, contiguous visceral cavities are seen in trans- 

 verse sections to be united by dark lines or bars which run 

 transversely across the walls and are sometimes very regular 

 in their arrangement (PI. XVII. fig. 3). These bars are the 

 beginning of the radiate crystallization which ultimately gives 

 rise to the " steUimicans " structure. 



In other specimens, again, or in particular parts of a speci- 

 men, we find the radiating crystalline structure further deve- 

 loped, a zone of radiating crystalline fibres now lining each 

 corallite, but the walls of the corallites still remaining visible 

 (woodcut, fig. B). This lining may be so thick as to leave 

 open only a small central space in each corallite (PI. XVI. 

 fig. 5), or it may extend quite to the centre of the corallites, 

 in which case each tube is filled with a more or less marked 

 crystalline stellate mass (PI. XVI. fig. 3, and woodcut, fig. E). 



The next stage is the more or less complete obliteration of 

 the walls of the corallites. This sometimes takes place 

 while the central portion of the visceral chamber is yet unin- 

 filtrated (PI. XVI. fig. 2 a) . More commonly the obliteration 

 of the walls is accompanied by the complete infiltration of the 

 corallites, in which case there is developed the typical " steUi- 

 micans " structure previously described (PI. XVl. fig. 2, and 

 woodcut, fig. D). 



Vertical sections show much the same differences in the 

 extent of the infiltration and subsequent crystallization by 

 which they have been affected ; but the walls are usually less 

 easily recognizable in these than in transverse sections 

 (PI. XVI. fig. 1 c, and part of figs. 4 and 6 ; PI. XVII. fig. 4). 

 It is usually the case also that the tabulse have been completely 

 obliterated ; but in most sections we may find here and there 

 smaller or larger portions of the visceral cavity — always occu- 

 pying the centre of the cavity — to which the infiltrating 

 material, owing to some local cause, has not penetrated, and 

 in which we meet with transparent calcite intersected by the 

 remnants of the tabula (PI. XVL figs. 1 c and 7, PL XVII. 

 fig. 4, and woodcut, fig. G) . This fact, among others, would 

 show that the stellate crystallization has struck inwards from 

 the walls towards the centre of the tubes, and not vice versa. 



