THE ANATOMY OF THE MADREPORARIA. 3 



tops of both ray specimens having been broken off, I have not 

 been able to prove this ; nor again to investigate the method of 

 budding; but in M. as per a is such another central cavity with 

 six septa, which is continuous with that of the apical polyp. 

 All other polyp cavities converge towards, and, by means of 

 canals, eventually open into, this central cavity, but no more 

 definite connection is traceable. Tissues not unlike mesenteries 

 are sometimes visible in it, but the alcohol in which the speci- 

 mens were killed did not penetrate sufficiently rapidly to 

 preserve the central parts in good histological condition. In 

 some sections the six septa are not recognisable, and the axis 

 of the branch is occupied by a wide-meshed network of coral ; 

 this is probably due to reabsorption of part of the skeleton. 



In transverse section are also seen concentric series of lon- 

 gitudinal canals (c".) permeating the corallum ; their 

 arrangement appears to indicate that the radial growth of the 

 branch is effected in the following manner. Directly beneath 

 the external body wall of the colony a series of longitudinal 

 canals runs between the costse (fig. 4, c^) ; and it is probable 

 that, for increase in the diameter of the branch, the costse grow 

 outwards, and then, bulging laterally, fuse over these canals, so 

 as to enclose them entirely in corallum (c/. fig. 10, x.). Thus 

 there results a series of internal longitudinal canals, concen- 

 trically arranged, with radii of coral between them which 

 represent former costse. Not only does the appearance of such 

 a transverse section as fig. 3 suggest that this is the mode of 

 growth, but also " dark lines of growth " (fig. 5) run radially 

 from each costa towards the centre, so continuously as to indi- 

 cate that what was a costa when the diameter of the branch 

 was very small, has continued to grow as such, and to be still 

 such, when the diameter is very much larger. New costae, 

 when required owing to the increased circumference of the 

 branch, appear to take their origin from the point of fusion of 

 previous costge. 



More minutely, growth is effected, presumably by the 

 activity of calycoblast cells, through the addition to and for- 

 mation of crystalline ellipsoids, similar to those described 



