SUPPLEMENTARY GONIOPOR^. 155 



directive, usually called the ventral septum. Wliere the septa meet, the granules may fuse 

 to form paU or remain separate, but they do not rise at all to break the curve of the concavity. 



A list of the known tliin, explanate Gonioporce will be found on p. 174 of Vol. IV. A 

 comparison of their calicles, as given in the magnified figures, is very instructive. The 

 thinnest forms tend to have very granular surfaces, representing the tips of the short trabeculae, 

 cf. PI. IV. fig. 1, and PI. V. fig. 4 of Vol. IV. with the figure of this coral. An interesting 

 variation on these is shown on PI. V. fig. 5 of Vol. IV., in this the trabecular elements are 

 feebly developed as compared with the horizontal ; the surface is consequently composed of 

 flakes. These specimens are also instructive from the fact that it is impossible to group tliem 

 under a single heading, because though showing the same general characters, they all differ in 

 many ways, and are further distributed so widely. In the meantime, therefore, in view of the 

 extreme plasticity of coral-stocks, we can do nothing but regard their special characteristics as 

 due to their localities. Their genetic relationships can be only puzzled out in time. The fact 

 that the calicles seem to vary with the form of the corallum no matter how accidentally that 

 form is acquired, should ptit a stop to all attempts to base specific affinities upon calicle 

 resemblances alone. For another striking instance of this total nonplussing of the prevailing 

 attempt to guess at specific relationships, see the next coral G. North- West Australia 7. 



a, h. In spirit, with a bleached fragment. Zool. Dept. 82. 2. 23. 120-1. 



163. Goniopora North-West Australia (7)7.* {G. Australice ocddentalis septima.) 



(PI. VIII. fig. 8.) 



[Adolphus Island, coll. Saville-Kent ; British Museum.] 



Description. — The corallum is massive, but owing to the destructive action of a sponge 

 close under the surface, it is apparently tliin (4 mm.) and encrusting, with edges supported by 

 very wrinkled epitheca, which follows the irregularities of the substratum. 



The calicles are subcircular, 2 mm. in diameter, shallow. The walls are low, round- 

 topped, not of even thickness all round, finely reticular, and frosted. The septa do not appear 

 as wall strioe, but begin to project slightly below the margin, their thin edges, very frosted, 

 descend vertically for a short distance and then curve out concavely to join the large 

 columellar tangle. The full septal formula is very apparent. In sections the septa are very 

 perforated, the round pores frequently arranged in series which slope steeply upwards from 

 the wall towards the centre. The columellar tangle is very large and flat, and appears to 

 consist of numbers of frosted granules ; the palic formula is confused, owing to the fact that 

 frosted granules are not limited to the principal points of septal fusion. Near the tliin edges 

 of the corallum, the usual rosette is visible to the naked eye. 



Wliat remains of the natural colour of tlie coral extends some 3 ■ 5 mm. below the surface, 

 and some 2 to 2*5 mm. below the level of the surface of the columellar tangle. 



The growth form of this coral is very puzzling. It appears, as stated, to be thin 



* For the earlier forms, see Vol. IV. p. 63. 



X 2 



