IN THE COCOS'KEELING ISLANDS. 45 



stated loc. cit) the mode of gemmation is centrifugal in Iso])om, as in 

 Madrepora, s. sir. Other points distingnishing Anacropora from most 

 species of Madrepora are the formation of the axis of tlie hraiiches by a 

 spongy coenenchyma, whereas in many (if not all) MadreporsB this, in 

 accordance with the centrifugal habit of budding, is occupied to a greater 

 or less distance from the ends of the branches by the downward prolon- 

 gations of the septa and the intersei^tal spaces of the apical calicle. The 

 rudimentary condition of the external part of the calicle distinguishes 

 Anacropora; for although it is commonly found (I refer to the sunk 

 calicles occurring in so many species between the prolonged tubular or 

 nariform ones) in some, it is never, so far as my knowledge extends, found 

 in all the calicles in any Madrepora. 



. Although in its general appearance it differs remarkably from even 

 the branched species of Montipora, yet the structural differences which 

 separate Anacropora from this genus are T^yy far less distinctive than 

 those w^hich separate it from Madrepora. In the first place, in spite of 

 its external resemblance to Madrepora, it has the same system of calicular 

 budding (viz. centripetal, from the distal coenencliyma) whicli we find 

 well developed in the ramose Montipora ; the trabecular structure and the 

 two-cycled arrangement of the septa is the same in both genera. On the 

 other hand, whereas in Anacropora there is always an undifferentiated 

 coenenchymal apes, devoid of calicle^, to the branches, in Montipora this 

 apex appears always to bear at least one calicle on its surface. In Anacro- 

 pora the calicles are always rather distant and tend to form lines, and are 

 slightly raised above the surface, forming low hill-like eminences, whereas 

 in the ramose Moiitipora (e.g, digitata, Dana, di^-aricata and superficialip, 

 Bruggeraami),vA\i(t\\on the wiiole most closely approach Anacropora, the 

 calicles open flush with the surface, are crowded indiscriminately, and no 

 linear arrangement is apparent. In Montipora foliosa, it is true, the 

 calicles, especially on the posterior aspect of the corallum, are elevated in 

 a similar manner; but the foliate growth and the monticular hder-cdM- 

 cular eminences of the ui^per surface seem to remove this species far from 

 the ramose Montiporse. Jt seems to me not improbable that, for the 

 reasons I have indicated, these ramose forms may have to be separated 

 from the foliate and massive species of Montipora. 



The relations of Anacropora may be thus shortly stated :— Anacropora 

 has the general growth of Madrepora, but the manner of budding of 



Montipora. 



The following is a description of the single species referable to this 

 genus which I am able to describe; owing to the interest attaching to the 

 type, I have allowed myself to give its characters at f liII length : 



AxACRorouA FonnEST, Bidley. 



Corallum branching frequently, clicTiotomously, occasionally subtri- 

 chotomously ; brandies given off in saccession in a subspirai manner, 

 the planes of successive bifurcations varying from about 30'' to 100" with 

 regard to each other ; angle between branches composmg bifurcation 80 

 to 100^ Stem and branches slightly curved, the apical branches more 

 strongly so, cylindrical, except the terminal branches, which tciid to curve 

 outwards and taper gradually io points ; diameter, mam axes G-7 milhm , 

 intermediate and teminal branches about 4 milhm., greatest length 

 between bifurcations of main branches about 30 milhm., terminal twigs 

 25 millim. long. Cah"cles arranged more or less definitely, for tlio inost 

 part in series wrhich follow approximately the longitudinal axis of the 

 stem and branches, the calicles of one series alternating witc those of the 



V 



5 



