88 MADEEPOEARIA. 



is a thicker layer of reticulum here than round the margin, and the projections are conical. 

 Long cylindrical calicles hang down from the under surface. 



In young, flat specimens, the calicles are fairly evenly distributed, with well-marked 

 interstices, but in older, humped specimens they are often closely crowded, especially on the 

 slopes, so closely that the costal echinulations have no room to develop; the walls are a 

 thin latticework, the exsert septa running out as long tapering spines united by ' delicate 

 mural trabeculse. The interstices in this case are deep narrow grooves with vertical walls 

 (PI. XXVI.). 



The intracalicular portions of the septa are weakly developed. Two cycles, with traces of 

 a third, appear in the protuberant calicular walls, while far down in the calicles the septa of 

 the first cycle alone become prominent, but do not reach the centre. 



Cffinenchyma in the young and flat stage shows very definite synapticular floors in well- 

 developed tiers, but as the coraUum gets old and humpy, growing irregularly in thickness, the 

 horizontal layers are less and less developed, while the vertical elements become more and 

 more leaf-like and predominant. The interstices are highly reticular and porous. 



There are five specimens of this coral which I have identified with Lamarck's original 

 Astrea myriojyhthalma. Lamarck's description agrees in every detail. The specimens are 

 large encrusting plates, " very rough," with " humpy " surfaces. The growth is distinctly 

 explanate. This identification prevents me from accepting that of Dr. Klunzinger. As far as 

 can be judged from the fragmentary specimen of Klunzinger's Astrceopora myriophthalma in 

 the National Collection it shows unmistakably the pulvinate type of growth, and cannot there- 

 fore be identified with a species forming large plates with humpy surfaces. 



The five specimens, four of which are from Mauritius and the fifth probably from the 

 same place, form an interesting series showing the flat, plate-like growth of the younger 

 specimens with the gradual formation of humps, until the largest specimen is an irregular 

 mass diff'ering so markedly from the younger forms, that but for the transition forms of the series 

 it would have been certainly classed as a new type. The calicles in the largest specimen 

 are crowded, with lattice-like walls, the septa running out into fine tapering spines (these are 

 here and there bent over the aperture in a way not easy to imderstand). The coenenchyma in 

 this same specimen consists largely of thin leaves standing out thickly crowded. At the 

 edges, the new calicles appear among these leaves as irregularly bounded spaces without, or 

 \\ith very slight traces of septa. 



The great mounds which this originally flat coral ultimately forms are due to successive 

 layers, each one humping irregularly in the middle. These layers appear one on the other, 

 each new layer creeping over its dead or dying predecessor. The growing edge of the ccenen- 

 chymatous skeleton seems generally to grow out faster than does the epitheca, but this does 

 not seem in all cases to cause the edges to droop. Long pendant cylindrical calicles, however, 

 appear on the imder surface, sometimes bent in all directions. These are slowly grown over 

 and ultimately closed by the advancing epitheca. I have selected as the type specimen the 

 one which shows best the transition from the flat young to the mound-like adult. 



a. Mauritius. (Type.) 



b. Mauritius (old specimen). 



c, d. Mauritius (younger flat specimens). 



e. Habitat imknown. [PiCgister No. 93. 7. 1. 17.] 



