22 MADREPORARIA. 



in which concentration of function is advantageous, and the whole stock may acquire a high 

 d^ree of organic unity (Siphonophora). While lastly the well known strobilation of Aurelia 

 may certainly be regarded as some very specialised variation of the same process. Now, since 

 the Stony Corals belong to this same group, it might have been expected that this same principle 

 would assert itself in spite of the secondary secretion of hard calcareous skeletons. These 

 skeletons might lead to variations of the process, but not prevent it altogether. 



It is, of course, impossible here to attempt a review of the group : a few examples must 

 suffice. Let us then in the first place consider the large fossil solitary corals, whose stocks at 

 the very first glance suggest the strobila of Aurelia, although it is also obvious that they are 

 in reality variations only of the same process. In the cases we refer to, the polyps, when 

 they reached their full size, appear to have budded, and the bud to have repeated the 

 form of the parent and to have secreted a skeleton like that of the parent just where it was, 

 on the top of the one from which it sprang. 



In this way these fossil stocks show series of skeletons, like rows of saucers or cups, one in 

 the other — rows which are sometimes of great length. Such cases are examples of the budding 

 of single adult organisms and the consequent formations of stocks, the very last individuals of 

 which were alone alive. 



Considerable interest attaches to the way in which these skeletons rest upon one another, 

 although we can assert little about it — it needs investigation. We may say, that however 

 irregidarly it began, it is certain that little by little the skeletons of the buds gradually 

 acquired the power of attaching themselves to the parent skeleton in such a way as to continue 

 its form and symmetry. We may, perhaps, correlate this with the rise and development of 

 the septa, over which the new ectoderm of the bud would naturally bend, and in the folds thus 

 produced commence to secrete its own septa. The coincidence, or rather continuity, of the 

 peripheral walls would probably be brought about more easily in the deep conical forms than 

 in the shallower saucer-like forms. And if memory does not deceive, the fossils which show 

 very irregular piles of calicles sitting in one another are most frequently found to have shallow 

 saucer-shaped calicles.* When this process is complete, that is, when the skeleton of each 

 generation appears simply to carry on the skeleton of its parent symmetrically, the only trace 

 of the process in the fossil would be in the series of tabulae. 



If tliis is a true account of the phenomenon, that the earlier cruder form of metamerism 

 with its piles of obviously discrete calices, passes gradually into a continuous skeleton, the 

 original segments of which are now so disguised as to be seen only in the succession of the 

 tabulffi, we shall have, in the formation of every such coral, to distinguish two methods of growth. 

 In the earliest stages of its individual development the growth will be normal growth de novo 

 from the larva, with the gradual withdrawal of the expanding polyp from the cup which is 

 progressively too small for it. This results in the deposition of one series of tabulse. When the 



• I am indebted to Mr. W. D. Lang, of the Geological Department, for calling my attention to 

 the interesting case of Parasmtlia of the chalk in which the successive calicles fail quite to fit into 

 one another symmetrically. 



