26 MADREPORARIA. 



2. The columnar form may be easily deduced from this by very rapid central growth and 

 formation of tabular floors, so that in section it would look like a pile of caps of nearly the 

 same size fitted into one another. The lateral edges may descend only a short way, or very 

 far and be very thin, cL fig. 2, diagram E. For a short edge see PI. XIII. fig. 5 ; for a long 

 edge see PI. XII. fig. 13. 



3. TJie Ptdvinate growth-form (fig. 2, C). — This may not be normal, but simply an adapta- 

 tion for the regeneration of injured surfaces of a stock. Assuming it to be normal, the caUcles 

 of the original stock all grow rapidly in length, even those round the edges, but without 

 much lateral expansion. The first cushion-shaped mass is thus formed. 



The uppermost calicles of this first stock still grow the fastest and tend to shoot above the 

 rest of the surface and, as explained above, to throw out a new edge, see diagram A. A second 

 cushion is thus formed. 



There are no very young stocks showing this formation from the beginning, but such 

 cushion-shaped masses are found, e.g. PI. XI. fig. 5. Again, when the tops of great convex stocks 

 have been killed down, one finds a group of calicles which has survived and continued to 

 grow, forming a cushion-like mass upon the dead stock. 



4. The cxpancliiu/ Sheaf Formation. — This is the most remarkable of all the growth-forms 

 in the genus, and took a considerable time to unravel. The central calicles lengthen so rapidly 

 that the primitive laminate septa seem to proliferate and to carry up the central region so fast 

 that the skeletons of the calicles are a mere sheaf of long, twisted, nearly vertical laminae 

 sometimes so confused that the calicles opening among them lose nearly all traces of radial 

 symmetry.* This shooting up of the septa as a sheaf of laminse may, if moderately rapid, 

 form a nearly rounded mass, see PI. XI. fig. 6. The laminae in this coral are shown in PI. I. 

 fig. 6. But when the growth is very rapid it results in a stock with straight sides sloping 

 inwards towards the base : compare the diagram fig. 2 D, which represents my interpretation 

 of the process ; cf. further PI. XIII. fig. 12, with its top calicles shown in PL VIII. fig. 1, and 

 PL XIII. fig. 8, with its caUcles on PL VII. fig. 4. 



A further specialisation of.this sheaf formation is described in the text, see G. Bed Sea 4, 

 and PI. VII. fig. 9. 



5. Branching growth-forms. — These forms will arise if the growth of the central region is 

 irregularly distributed. The rising stock will then fork and produce branching clusters, either 

 with long edges, fig. 2 E, also PL XII. figs. 12 and 13, or with short edges, PL XI. fig. 1. 



In concluding this sketch of the chief morphological details discoverable from the known 

 specimens it remains to be stated that by far the larger number of the specimens are shallow- 

 water forms. The few which have been brought up from deeper parts of reefs justify us in 



* In the large specimen of Astrmopora figured in Vol. II. PI. XXVI. the skeletons of the 

 calicles are here and there composed of confused clusters of twisted laminre running out into flame- 

 like arrangements and obliterating the radial symmetry of the calicles. These are clearly analogous 

 phenomena. 



