18 MADREPORARIA. 



may occur in cases where the cocnenchyma is spongy, and the reticuhim abuts against the 

 septa, here closing the interseptal loculus, there leaving it open. The continuation of the 

 interseptal loculi into tlie channels of the coenenchyma is an index of the amount of inter- 

 communication which exists between the polyp-cavities and the intercalicular canal system. 



The Columella. — This varies greatly both in size and character. Taken in conjunction 

 with characters other than the form of the fossa, on which it largely depends, it becomes of 

 value for classification. 



The Coeneiichyma. — This is sometimes very abundant, at other times sparse ; or in 

 other words, tlie calicles are at times wide apart, at others closely crowded. There is great 

 variation in this matter. Nevertheless, when either of these characters is sufficiently pro- 

 nounced and uniform in its occurrence over the whole corallum, it is of undoubted value for 

 classification ; but the variations are too great to admit of minute differences being taken into 

 account. 



The character of the surface of the coenenchyma is subject to great alterations in one 

 and the same corallum, being dependent largely upon the .slope. On steep slopes, down which 

 nutrient fluids would tend to stream, a pronounced ridge-and-furrow system as a rule will 

 be found. This may be nearly straight or else gyrate, making a kind of arabesque pattern. 



The ridges may be continuous, or broken up into lengths ; sometimes practically dis- 

 appearing, and leaving nothing but a granulation. They may be toothed in one or two or 

 more rows, the teeth being arranged longitudinally like the teeth of a saw, or transversely ; 

 or they may stand up like stiff bristles, giving the whole surface a velvety appearance. Or 

 again, the ridges may arch irregularly over the furrows, which then look like the burrowings 

 under the surface of minute worms, which here and there break through. 



The furrows, again, may not be parallel, but reticulate over the surface, so that the 

 ridges break up into irregular islands within the meshes formed by the furrows. The pores 

 leading into the canal system open in the bottom of the furrows, and show traces of 

 variation which do not appear to be accidental. 



The ridge-and-furrow system may be altogether displaced by a spongy texture either of 

 lamellaj or of fibres. A coarse spongy character may occur at any time, in patches, as an 

 index of very rapid growth ; it is, for instance, the invariable character of the growing edges. 

 It is only of value when it is found over the whole surface of the corallum, there being no 

 indication of abnormally rapid growth. 



There is some ground for believing that the aspect of the coenenchyma is due to 

 local influences. A group in the collection from Formosa have a remarkable superficial 

 resemblance. The same is the case with the specimens from Tongatabu, and, again, from 

 Shark's Bay. These resemblances are entirely superficial, and due in each case to some local 

 peculiarity of the coenenchyma. It appears as if, in each case, the special envu-onment fixed 

 the character of a coenenchyma. Whether this environment is climatic or faunistic it is 

 impossible to say. The number of larva; of parasitic organisms found attached to corals seem 

 to forbid us to exclude the latter suggestion. 



