CORAL i^OEMATIONS. 33 



so the reef continually subsiding at the base and rising at the summit, grew to a 

 far greater thickness than 120 feet. 



Thus was explained the formation of barrier reefs at great distances from the 

 shore. At one time they fringed the coast, which slowly sank with the general 

 movement of subsidence, while the reefs continued awash, thanks to the incessant 

 labour of their coralline inhabitants. The mainland, which foi-merly served to 

 support the superstructure, gradually sank deeper and deeper, thus continually 

 retiring from the outer barrier of the steadily rising coral reefs. The passage 

 also became gradually enlarged, and bj^ the disappearance of the central nucleus 

 itself the inner waters were at last transformed to a lagoon. Certain archi- 

 pelagoes, such as the Low Islands, are compared by Dana to a vast cemetery, 

 where every atoll marks the site of an engulfed land. 



According to this theory it would therefore be easy to determine the character 

 of the oscillating movements to which the oceanic islands are subjected. The 

 reefs raised to great heights above the sea would thus indicate an area of upheaval, 

 the fringing coralline rocks would imply a state of comparative stability on the 

 seaboards, while the barriers and the atolls might be likened to floats placed on the 

 sites of submerged lands. Most of the Pacific islands — that is to say, all those that 

 follow from Pitcairn in the Low Archipelago to the Philippines along a line 

 passing north of Tahiti and Samoa — would thus belong to a zone of depression, and 

 these scattered groups might be regarded as fragments of a vanished continent, 

 stretching across the south side of the North Pacific Ocean. 



Such is Darwin's theory, which, however, can scarcely be applied with any 

 probability to all the oceanic lands girdled by coral reefs. Wherever the rocky 

 pedestals supporting the superstructures of living polypi themselves consist of 

 calcareous secretions to any great depth, there can be no doubt that subsidence has 

 really taken place. But verifications have hitherto been made only at a limited 

 number of points, and in the absence of direct observations it would be rash to do 

 more than regard subsidence as very probable wherever the outer walls of the 

 coralline islands plunge rapidly — as, however, they rarely do — into abysmal depths. 

 Thus near Enderbury, in the Phœnix Archipelago, the soundings reveal 1,800 

 fathoms within 3 miles of the shore, 900 fathoms at 1,400 yards from Danger 

 Island, near Vanikoro, while one of the reefs at Tahiti indicates a seaward slope of 

 72 degrees. 



On the other hand, observations made in the vicinity of certain coralline islands 

 show that at the foot of an escarpment less than 200 feet high, there stretch vast 

 platforms where fragments of volcanic origin have been found scattered amongst 

 crumbling blocks of coral. In this case it is quite possible that eruptive cones 

 eroded by the waves to a slight depth below the marine surface may have served 

 as foundations for the coral-builders, or else that their structures have been raised 

 on rocks entirely formed by other organisms working at considerable depths. But 

 many protracted observations must still be made before the diverse coralline islands 

 can be classified according to their origin and history. Several groups, such as 

 the Low Archipelago, Fiji, the Pelew, Solomon, and Tonga Islands, supposed by 

 3—0 



