86 PROFESSOR T. W. EDGEWORTH DAVID AND MR. G. SWEET. 



Minor oscillations of the shore line followed, during which debris from the breccia 

 sheets became overgrown by Lithothamnion and coral, as at Telele, and near 

 Luamanife, on the main island, as well as in the case of the Amatuku conglomerate, 

 and thus an upper crust of rock formed itself above the old breccia sheet. Finally, 

 a downward movement predominated, and portions of the raised breccia sheet at the 

 centres of the islets, where they were beyond the protection of either the inner 

 or outer Hurricane Banks, would be dissolved away under the action of rain-water 

 charged with humous acid, whereas the portion of the breccia sheet under these 

 Banks would be more or less protected from corrosion by the overlying rubble and 

 sand. Hence, towards the centres of the islets, a corrosion hollow would be gradually 

 channelled out of the breccia, down to near the level of low tide, so that at high tide 

 a strip of sea-water would occupy the centres of the larger islets, as is now the case. 

 Such hollows would be bounded by strips of breccia of almost the full original 

 thickness of the sheet. If, therefore, through denudation and erosion, operating 

 from the ocean lagoon wards, the Hurricane Banks were driven in and pushed over the 

 strips of breccia, and across the reef platform, a low wall of breccia would be left 

 facing the ocean, marking the former foundation of the Hurricane Banks. Meanwhile, 

 the detrital material of the outer bank, driven rapidly across the old corrosion 

 hollow in the breccia, would find support against the strip of breccia preserved under 

 the inner Hurricane Bank. This is precisely w^hat has happened at Fualifeke and 

 Pava Islets, w^here the outer wall of breccia marked O.2.D. (see Plate 13) marks the 

 original position of the outer Hurricane Bank (see specially Section 1 of Fualifeke). 

 What is, perhaps, even a more instructive case, is afforded by the islets of Amatuku 

 and Mulitefala (Plate 12). The northern shore of Amatuku still holds the advanced 

 position on the breccia line O.2.D., but Mulitefala Islet has been driven from 

 the outer defence lagoonwards, its rallying point in this case being chiefly some 

 masses of bedded calcareous sandstone, w^hose lagoonward dip shows that they were 

 formed on the lagoon shore of the reef (see Section 2 of Mulitefala, Plate 12). 

 It is, however, at the south-west portion of the atoll that this action is perhaps best 

 illustrated. A long but interrupted line of breccia outliers runs from near Fuagea 

 Islet to Avalau, and thence bends round to Motungie Islet (see Plates 3, 4, and 5, 

 and particularly Section 1 of Plate 4 of Tefala). There has obviously at one time 

 been a long island here from near Fuagea, down to as far south as Avalau, and from 

 here right around to Mateika. 



We return now to the consideration of the further growth of the islets of 

 the atoll. Inside the Hurricane Banks blown sand and fine rubble gradually 

 accumulated, and the islets became clothed with vegetation, and stocked with 

 their present fauna. Meanwhile the deposit of silt at the inner angle of the 

 main island gradually led to the death of the belt of Heliopora caridea, 

 \\'hich up to this time still flourished in the lagoon at this part of the atoll, 

 though it continues to flourish luxuriantly in the lagoon (from a little below the 



