CORALS FROM MURRAY, COCOS-KEELING, AND FANNING ISLANDS. 157 



any of the preceding dead growth in the axis of the large detached excrescence (b); this fact 

 shows how richly the corallum sends up its irregular masses. 



"a. Murray Island, Torres Straits. 85. 6. 30. 3. (Type). 



"b. (Fragment of same specimen)." 



Dr. F. Wood Jones has sent to the U. S. National Museum two pieces, evidently 

 representing one species of Montipora, which grows on the barrier of Cocos-Keeling. 

 A description of these pieces follows : 



One piece represents the expanding lamina of the base (plate 64, figs. 4, ^a, ^c). It is 

 82 mm. long, from 41 to 53 mm. wide; up to 9 mm. thick back from the edge; edge about 

 3 mm. thick. Upper surface undulate, with one low hillock; lower surface more even, 

 epitheca ranges from 8 to 38 mm. from the edge. Living tissue somewhat reflected under 

 the edge. 



Calices on upper surface about 0.75 mm. in diameter, margins sunken, boundaries 

 usually distinct and rather thick; on lower surface distinct, about o.t; mm. in diameter, 

 margins flush with the surface or slightly protuberant, surrounded by distinct, thickish, 

 ring-like walls. 



Septa on the upper surface, two well-developed cycles; of the primaries, 4 are about as 

 long as half a radius, i or 2 directives longer; the 6 primaries may meet in the bottom of the 

 calice to form a weak columella. Lower surface, two cycles present, but smaller than on the 

 upper surface. 



Coenenchyma, streaming layer nearer the lower than the upper surface. On lower 

 surface much compacted within the epitheca. Outside the epitheca it forms a flat, open 

 reticulum between the calices. On the upper surface trabeculre grow upward and form 

 erect, densely frosted tubercles which are intimately connected with the septa on the outer 

 edges of the calicular walls. Frequently they form palisades surrounding the calicular 

 apertures, and sometimes are fused one to another. In places where the calices are suffi- 

 ciently distant from one another there are tubercles in the interspaces. The reticulum is 

 loose, so that openings may be seen between the tubercles. 



The other specimen (plate 64, fig. 3) represents a lobate protuberant above the 

 explanate base. It is 67 mm. tall, and is most irregular in its subdivisions, which may be 

 pointed or obtuse with ends about 7 mm. in diameter and without tubercles on the terminals. 

 The calicular characters and the tubercles are similar to those of the upper surface of the 

 basal part of the colony. The surface is spiny, the tubercles are up to 1 .5 mm. tall, and may 

 be loosely fused in rings around the calices or in places there are indefinite indications of loose 

 radial fusion. 



Habitat and color, Cocos-Keeling Islands. — Dr. Wood Jones remarks that this 

 species "lives on the barrier; exposed parts of zooids is fine purple." 



This species greatly resembles Montipora hispida Dana, from which it differs 

 by having the tubercles closely associated with the caHcular margins, whereas in 

 the latter they stand back from the caHcular margins. It is also close to M. stcllata 

 Bernard; in fact, the latter is probably based on the explanate basal part of a 

 corallum. 



Bernard describes 15 species oi Montipora which have separate tubercles that 

 tend to form rings or palisades around the calices, and three others may have the 

 tubercles associated with the calicular margins. Of these, 6 are reported from the 

 Great Barrier Reef or Torres Strait: M. variabilis Bernard, M. stcllata Bernard, 

 M. lichen Dana, M. granulata Bernard, 71/. aqiiituherculata Bernard, and M. informis 

 Bernard. On page 158 is a tabular summary of the characters of these species, 

 except M. lichen, which is not closely related. 



The coral on which the next description is based belongs in this group 

 and is very close to M. informis and M. stcllata, but it does not precisely accord 

 with either. As a sufficiently large suite of specimens for ascertaining its range of 

 variation is not available, it can not be determined whether it is a distinct species 

 or only a variant. Under these circumstances it seems best merely to indicate its 

 relationship with the nearest recognized species. 



