SUPPLEMENTARY GONIOPOR^. 149 



support. We can trace stages in the proliferation of the reticulum, recalling those sliown in 

 the Paris basin group which would thus appear t« have grown in an unfavouraJile environment. 

 All traces of calicles may be obliterated in a flaky reticulum, as seen in Vol. IV. PI. X. fig. 6, 

 with intermediate stages not unlike fig. 5 of the same plate, wliich seem to lead us gradually 

 to the calicles now figured. 



This coral, which the formal systematist might be inclined to throw aside as abnonnal, is 

 of great interest. For we can learn from such accidents almost as much as we could from 

 any experiments in skeletal variation that we could devise. It is only Ijy a study of variation, 

 however caused, that we shall ever arrive at a natural system of the corals. For an instructive 

 case see Pontes Bahamas 1. 



There are two large double calicles of a shape and character wliich suggest that they are 

 due to abortive fission. 



a. Zool. Dept. 1903. 4. 3. 1. 



156. Goniopora Ellice Islands (4)3. {G. Ellinnna Inilii.) (PI. VIII. fig. 2; 

 PI. XVII. fig. 16.) 



[Funafuti, coll. Sollas ; British Museum.] 



Description. — The corallum, having been apparently attached to the side of some steep 

 portion of the substratum, grew out laterally and upwards, hoof-shaped, faintly and irregularly 

 cloven, and with the sole uppermost. The living colony covers the sharp edge of the hoof 

 extending backwards like a thick cushion over the previous cushion-shaped soles, and 

 downwards as a tliiu creeping layer with conspicuous, wrinkled, epithecal pellicle. The 

 condition of the fracture suggests that the whole stock has Ijeen built of at least two such 

 hoof-shaped colonies. The specimen came away completely from the rock, exposing the 

 epithecal base of the original colony. 



The calicles are subcii'cular, with immense numbers of intracalicular buds ; they are 

 as large as 4 mm. round the growing edge of the hoof, but on each side, upper and lower, 

 of tliis edge they decrease in size to 2 mm., but decreasing more rapidly on the upturned 

 sole than on the lower side. They are deepest (3 to 4 mm.) near the extreme edge, but get 

 shallower till on the under surface, on nearing the epithecal film, they are flush with the 

 smface. The walls are thick and regularly reticular over the upper surface and show the 

 septal elements as thin radial lamellse, which seldom run right across the wall, but either end 

 up against a zigzag median line or fork and are lost in the reticulum. Along the gi'owing 

 edge of the hoof, there is a tendency for the whole skeleton to tiu-n into a lamellate stroma 

 such as that described as typical of the expanding sheaf method of growth and figured Vol. IV. 

 PI. VII. fig. 4. This lamellate character is seen in all the calicles of the upper smface, but less 

 pronounced and more regular than right at the edge, forming a reticulum recalling that seen 

 in patches on G. North- West Anstr-alia 6, see Vol. IV. PI. IV. fig. 6. Just bek)w tlie sharp 

 edge of the sole, the walls suddenly thin away so that many are simple latticeworks, but 

 gradually thicken again in the tyjiical manner, remaining simjile ami not reticulai' as they near 

 the epitheca. 



