NATURAL HISTORY SOCIETY OF GLASGOW. 259 



4, c, is as regularly tuberculated as that shown in his figures of H. 

 mirahilis. Except in very mucli worn specimens, the tubercles are 

 always present upon both large and small diameters of the corallum, 

 and as they are sometimes seen to vary slightly in number and 

 regularity, even upon parts of the same stem, mere irregularity of 

 occurrence is not, therefore, to be considered of any specific value. 



The costse and intercostal spaces are also characters that vary 

 considerably in this coral, and cannot, I think, be depended upon 

 as points of specific distinction, as some specimens of the largest 

 diameter have costse less developed than those seen upon more 

 moderately sized stems. The intercostal spaces are wide, shallow, 

 or deep, according to the diameter of the specimen, and the 

 prominence of the costge. 



In H. Lyelli, the horizontal section of the corallum is stated to be 

 hexagonal, in H. mirahilis, mostly circular; this is a point that 

 appears to me of no value in the diagnosis of the species, as the 

 section varies with the prominence of the costae and the convexity 

 or concavity of the intercostal spaces, and this variation may some- 

 times be seen upon the same specimen. In general, the greatest 

 number of specimens of all the various diameters met with, approach 

 the hexagonal form; one specimen, however, in my collection shows, 

 in a cross section, a stellate or six-sided angular form, and another 

 well marked fragment is quadrangular in section, having four small 

 tuberculated costag. Internally this specimen has only four septa 

 corresponding with the costae, instead of six, the normal number. 

 Other specimens occur which show a nearly circular section, and 

 on some of these, instead of costse, we have only longitudinal 

 grooves to the floor of which the booklets were fixed. These 

 difi'erences in the external form of the stem of the corallum I 

 consider to be mere exceptional variations of one species. 



The endotheca and dissepiments are also characters that vary. 

 In many specimens the endotheca has either never been well 

 developed or has been destroyed by the crystallization of the 

 internal structure of the corallum. The dissepiments are stated 

 to be wide apart, but one longitudinal section shows 30 in the 

 length of one inch, being less than one half line apart. 



It will be seen from the above remarks that this coral varies, to 

 a certain extent, both in its external form and internal structure, 

 many of its parts depending upon the state of preservation in 



