FIJI ISLANDS. 307 



The water on the reef edge was usually not much more than 

 ankle-deep, but the breakers sent from time to time so strong a 

 current inwards across the barrier, that it was difficult to keep 

 one's footing. On the reef were resting irregularly shaped 

 masses of solid stony corals, portions of various Astrceidce, Pori- 

 tidce, or of reef rock, thrown up upon the marginal platform of 

 the reef by the surf, and reminding one, as they rested in all sorts 

 of positions, of the scattered rock fragments on a glacier. Some- 

 times they even rest on a narrowed support like " table-stones," 

 having become first cemented to the platform, and subsequently 

 gradually undercut by the waves. Dana has figured such table- 

 stones. It is these thrown-up fragments which are, as has been 

 described, the only portions of the actual reef visible from a 

 distance. 



The chief differences between the fauna of the Fijian reefs 

 and those of Bermuda, are the absence at Fiji of any large 

 quantities of coral formed by Milleporidce and large branching 

 Oculinidce, and the absence of the large flexible Gorgonidce, 

 which form so striking a feature at Bermuda. The great abun- 

 dance of Madrepores forms the characteristic feature in the 

 Fijian reefs. I saw, however at Fiji, no Madreporas so large 

 and fine in growth as those of St. Thomas. 



On the reef-margin, by turning over the cast-up rock frag- 



ACEOCLADIA MAMILLATA. 



ments, I found a few cowries, some huge Troclii, also specimens 

 of Turbo operculum, and other shells. Various Holothurians and 

 a large bright ultramarine-coloured Starfish (Ophidiaster), were 



x 2 



