270 THE INFLtTEHCE OF INANIMATE SUKRODNDINGS. 



were formed during a period of elevation, if we duly realise in 

 oiu" mind the effects tliat must have ensued from the combined 

 action of upheaval and of more or less constant currents on the 

 form of the growing reef To the eastward the impact of the 

 current on the island is perpendicular to the face of the coast ; 

 it therefore beats upon the slowly rising sea-bottom, and, being 

 tolerably strong in itself, it prevents the perpendicular growth 

 of the reef corals. The result is that the reef itself is driven 

 very close to the coast. Moreover, between it and the foot of 

 the land no deep channel can be formed by currents, for here 

 the very hard black rock of the Andesite cliff prevents any s\ich 

 rapid grinding down as is easily effected in the coralline limestone. 

 On the west coast it is otherwise. Here the currents rarely 

 impinge directly on the face of the reef, but only at an angle. 

 I have ali'eady shown that whei'ever strong currents sweep 

 past a reef at an angle these are forced to grow perpendicularly or 

 nearly perpendicularly upwards; this is everywhere the case on 

 the western shore. The fact that the outer reef here is separated 

 from the islands by a channel above forty fathoms deep and 

 many miles wide, finds an easy and unforced explanation on the 

 assumption of an upheaval. It is certain that the enclosed 

 island of Babel thuap was formerly much broader than it now 

 is, as is proved by the existence of the little island in fi'ont of 

 Roll (see Map), which is now far from the land on the surface 

 of the inner reef. Now, if we suppose — as we must even on the 

 theory of a subsidence — that the island, which here consists 

 almost entirely of tufas, originally extended nearly to the 

 western outer reef, only nairow fringing reefs could have ori- 

 ginally been formed on the western side ; but these must from 

 the first have grown perpendicularly, because they were impinged 

 on by tangential currents ; and since the tufa on this side could 

 offer but a feeble resistance to the action of surface-water and 

 rain, aa well as of the surf beating on it, a small channel might 

 soon be formed between the reef proper and the coast. This 

 channel would presently gi'ow wider, in propoiiion as the enclosed 

 island, consisting of soft stone, was gradually eaten away, and 

 during slow upheaval it would continue to grow deeper, in pro- 

 portion as the old porous portions of the reef and the rock in 



