IS GENEEAL EEMAEKS. 



which need not be considered now. To take another instance : Dana included M. echinata, 

 M. rosaria, anAM.florida in his group D. These three species appear to me to differ in so many 

 points from each other that they have been referred to three different suljgenera. Even for 

 specific purposes it is doubtful how far habit forms a reliable character. Dana points out 

 that the habit of a specimen is to some extent dependent on its situation. He states (p. 434) : — 

 " The cespitose species appear to vary in shape somewhat according to the depth at which 

 they grow. In those near the surface the branches spread more nearly horizontally and are 

 consequently more crowded and more completely coalescent ; whilst those at greater depths 

 have a more ascending mode of growth and the less coalescent branches appear longer and 

 more subdivided." Klunzinger noted in the Red Sea the extreme variability of M. corymbosa 

 and other species. Pourtales also, taking note of the difference in the relative exposure of 

 M. palmata, M. cervicornis, and M. prolifera, on the West Indian Reefs, hinted that the 

 three species of Lamarck may prove to be variations of one species, dependent on environ- 

 ment for their precise habit. Such is the opinion I have arrived at after comparison of a 

 very large number of specimens, although it must be admitted I have no precise information 

 as to environment. 



2. Form of the Radial Corallites. — The form of the radial corallites is subject to considerable 

 variation, in many cases even in different parts of the same colony ; but, as Klunzinger has 

 already pointed out, in every species the variation takes place on a fixed plan, so that each 

 species has its own particular type of corallite, or combination of such types. The character 

 of the corallites appears, however, to enable one to distinguish species rather than groups, 

 and if applied to divisions based on habit it undoubtedly breaks down. Dana's division A 

 may be taken as an illustration. M. palmata, Lamk., and M. cyclopea, Dana, are placed in 

 one section, characterized by a tubular form of corallite, whilst M. conigera, Dana, occupies 

 another section, characterized by the presence of labellate corallites, but also by the occurrence 

 of numerous incipient branches scattered over the surface of the corallum. In M. palmata and 

 M. cyclopea labellate corallites are of frequent occurrence, and in certain forms of the former 

 constitute by far the greater portion of prominent corallites. Thus in this case it is not the 

 form of the corallites but the occurrence of conical incipient branches on the surface which 

 forms the important distinction, and even then M. cyclopea occupies an intermediate position 

 between M. palmata and M. conigera. In another case Milne-Edwards and Haime associate 

 together such diverse species as M. echinata, M. tubulosa, and M. cuspidata in a group 

 characterized by the presence of very long tubular corallites. 



3. Condition of the Surface of the Corallum. — In the genus Madrepora one may distinguish 

 two principal groups of species — the one characterized by the presence of marked costulse on 

 the corallite-wall and of an open network between the radial corallites, the other by an 

 extremely compact or almost solid surface clothed with delicate echinulations ; but innumerable 

 intermediate conditions occur. Doubtless the condition of the surface does reflect to some 

 extent the physiological activity of the living colony and as such must be regarded as a 

 character of value, but only when associated with other characters. In practice it is found 

 extremely difficult to make important use of such a character, partly because the surface is 



