144- 



■early growth lines ; the diameter is 2 mm., and the septa are 

 practically the same as in the previous stage (fig. 11). A 

 third is \'^ mm. high with a very marked looping outwards 

 of the epitheca between the septa, of which the primaries 

 extend further into the corallite. 



In fig. 12 is represented a stage where the primary septa 

 are beginning to give off trabeculae ; the specimen is 175 

 mm. high and 25 mm in diameter, and has well-marked 

 growth-lines on its epitheca. In fig. 13, a specimen \'s mm. 

 high by 2 mm, in diameter, the primary septa have all fused 

 with one another by trabeculae, but as yet there are no signs 

 of any tertiary septa. The latter have appeared in the next 

 stage (fig. 14a) 4 mm. high by 3*5 mm. in diameter. The 

 •calicle has begun to elongate the terminal septa of its long 

 axis belonging to the first cycle. The six primaries 

 are still distinctly the larger, but six secondaries have 

 fused with them and with one another. The corallite in side 

 view (fig. 14b) shows marked accretion lines and the 

 •characteristic wavy structure of the epitheca. The former 

 represent slight additions to the epitheca, and do not corre- 

 spond to the lines of growth in the older specimens, which 

 are markedly larger and consist of many such. 



The further changes lie in the gradual fusion of the 

 tertiary septa to those of cycles I. and II. by trabeculae. 

 This does not take place generally until there has been a 

 very marked and considerable increase in the size of the 

 corallite, nor usually until after it has become free. Fig. 

 15 represents a free specimen 6 mm. high, 9*5 mm. long 

 axis of calicle, and 4*5 mm. short axis of same. Only 

 •one septum of cycle III. has as yet become fused with 

 those of I. and II. This is situated in an end space between 

 septa of cycles I. and II., and it will be noticed that the 

 tertiary septa in these four spaces are larger than in the side 

 •ones, being indeed the first to fuse. Those in the next four 

 spaces are larger than those in the central ones, which are 

 naturally from the method of growth of the coralkim the last 

 to join up, only indeed fusing in the largest corallites. In 

 the figure it may be observed that in two of the end spaces 

 of the calicle between septa of cycles I. and III. quinary 

 septa have appeared on either side of the quaternary, which 

 are everywhere complete. 



The corallum, where it breaks off from its psdicle, varies 

 •considerable, but is usually in the stage with 12 septa fusing 

 with only traces of the tertiary septa (fig. 16). A series of 

 perforations appear right round the corallite in one of the 

 accretion bands near the base, but what causes these I have 

 b3en quite unable to determine. In no case is there any 



