150 MADREPORAEIA. 



The septa are thin and lamellate, and in the shallow calicles show the typical formula. 

 In the deep calicles, they are close rows of long delicate spines, which unite deep down to form 

 a light columellar tangle without symmetry. As the calicles become shallower, the septal 

 teeth become shorter and more typical, and the columellar tangle more compact, till in the 

 shallowest it is a flat, nearly solid floor covered with coarse granules which shows clear traces 

 of the typical rosette arrangement of jagged and straggling granules. 



In the longitudinal section, the walls are stout and solid looking, the septa are irregularly 

 perforated with very large and very minute pores irregularly scattered. 



Tliis specimen, with its method of growth, which may have been purely accidental, is 

 interesting because it shows some of the same variations of the calicles as are seen in G. North- 

 West Australia 6, and perhaps throws some light on them. It certainly suggests that the 

 deep calicles with regularly reticular walls, which are only so far known in these two repre- 

 sentatives of the genus,* may be more favourably situated than those with thin lattice 

 walls. 



In this one specimen, then, we have a very wide range of calicle variation, and obviously 

 all of tliem due to the accidents of position ; see observation under the last heading. 



In general colour and appearance, it recalls 0. Great Barrier Beef 2 from Palm Islands. 

 But its skeletal elements are more delicate, and the septa do not form such conspicuous wall 

 striae, and there is no median furrow on the walls. 



On the dead parts of the upper surface, what seems to be a young colony of four calicles 

 appears in a deep epithecal saucer. This, however, is hardly a fresh colony in the sense that 

 it started de novo from a larva. It is rather a minute portion of the old colony, which has 

 survived and protected itself with an epithecal wall from the surrounding decay. This may be 

 gathered from the fact that the fossse run down into the old colony. 



a. Zool. Dept. 1903. 4. 3. 3. 



There is, further, a worn massive fragment which may be the same kind as this. Its 

 skeletal elements are thickened by post-mortem aqueous corrosion. The specimen G. Ellice 

 Islands 1 encrusts a part of it. 



The calicles are from 3 to 4 mm. across, were evidently somewhat deep, the walls were 

 reticular, and the septal formula complete. It is clear tliat with the range of calicle variation 

 shown on specimen a, this dead fragment may, so far as size of calicles is concerned, safely be 

 placed here. 



h. Zool. Dept. 1903. 4. 3. 2. 



157. Goniopora Ellice Islands (4)4. {G. Ellidana quarta.) (PI. VIII. fig. 3; 



PI. XVII. fig. 17.) 



[Funafuti, coll. Sollas ; British Museum.] 



Description. — The corallum forms tufts of short, thin, bent stems, which fuse together, 

 swell into rounded and angular knobs at their tips, and frequently divide. The stems 

 are from 7 to 10 mm. thick. 



* We find a similar variation in Porites Fiji Islands 18, see Vol. V. p. 56. 



