80 MADREPOEARIA. 



The colonies range in form from flat expanses to globular masses. In the former, new 

 buds are formed, chiefly round the growing edge, wliile the existing polyp-cavities lengthen 

 but slowly ; in massive specimens, the polyp-cavities lengthen rapidly, and, as the intervals 

 between these cavities increase, buds appear between them. These massive growths, 

 increasing in tWckness more rapidly than in breadth, reach a stage when the basal regions 

 die, so that, whUe the top is continually expanding, the base has ceased to grow. These 

 become top-heavy and roll over. The rolled dead portion is then grown over by fresh 

 growths. In this way, globular forms occur which may show no point of attachment. 



The whole skeletal mass uniting the polyps, and rising above the basal epitheca, is thus 

 septal or ccenenchymatous. The intracalicular portions of the septa are but very feebly 

 developed, while the extracaUcular (costal, ccenenchymatous) portions are well and often 

 strikingly developed. 



We may, perhaps, describe the coenenchyma as being typically built up of two elements, 

 a vertical and a horizontal, in the following way. The costaj stand out round the calicle as 

 tapering echinulse, sloping upwards all round the aperture of the calicle. These echinulfe, 

 which constitute the vertical or upright elements of the coenenchyma, may or may not show 

 the primitive connection with the septa. The horizontal element is a series of more or less 

 flat perforated plates, which bind all the uprights together, so that the whole coenenchyma 

 appears to be composed of tiers of horizontal synapticular floors, supported by columns 

 passing through them. 



This typical structure of the coenenchyma shows great variations. In flat, horizontal 

 coraUa, the vertical element may be reduced to thin, irregular pillars, while the horizontal 

 element is very conspicuous, making the whole surface of the corallum look smooth and flat. 

 On the other hand, in globular specimens, or on portions of specimens growing rapidly in 

 thickness, the upright, vertical elements stand high, giving the whole a strikingly echinulate 

 appearance, while the horizontal floors are reduced to smaU connecting trabeculse. 



Again, both elements may lose their distinctive character, becoming less definitely 

 upright or horizontal, the whole coenenchyma then forming a reticulum, barely recognisable as 

 belonging to the type. Lastly, by the thickening of the elements into solid, u-regular 

 granules, the whole may look like a rough agglomeration of cemented grains with interstices. 



The coenenchyma may assume more than one character on one and the same specimen, 

 according to the position of the part. For instance, in a globular specimen, the uppermost 

 surface, where the corallum is increasing more rapidly in thickness, is a light open reticulum. 

 This gradually changes at the sides and under surface into an almost smooth, continuous 

 synapticular floor, broken only by the calicles and by pores. 



The discarded epitheca plays here and there an important part. If the corallum is a 

 flat expanse, the epitheca grows with the growing edge of the corallum, either adherent to 

 a substratum, or freely rising from it. Wherever the epitheca does not keep pace -witli the 

 outward growth of the ccenenchymatous skeleton, the latter may droop over its edge, or send 

 down long calicular tubes or rootlets, which latter are slowly grown over by the epitheca. 



