SUBMARINE DRAINAGE. 269 



sea-water can. psrcolate througli the walls of the atoll-reef into 

 these reservoirs, and consequently proves that the' soil must be 

 pierced by larger or smaller channels. The raised hmestone 

 islands of Kokeal are similarly porous. One very small island 

 close to Coroere has quite the structure of a tru.e uplifted atoll. 

 A high wall of metamorphic coralline limestone surrounds a 

 deep lagoon on all sides, and no channel leads from this to the 

 sea at the present day. There is a little ridge, which may for- 

 merly have been a channel, to be crossed in order to reach the 

 lagoon. Nevertheless the same fish are found in the lagoon as 

 outside, the water is equally salt, and its level rises and sinks 

 regularly with the tidal flow and ebb. The following circum- 

 stance is even more conclusive. On one of my excursions round 

 Pelelew, before going into the harbour of Nasiass from the sea, I 

 saw out at sea a wide current of yellowish water, almost fresh, 

 flowing from the land; its edge was pretty sharply defined 

 against the sea water. It was exactly low water. The natives 

 informed me that this current was always there, though some- 

 what more feeble at high tide, and strongest during the rainy 

 season. Now in the whole of Pelelew there is not one stream 

 from which this current could proceed. All the surface water 

 percolates at once through the excessively porous soil, and it is 

 only during heavy storms of rain that drainage streams form 

 in the gullies, and these totally disappear again in a few hours. 

 This water then runs away through the soil down to a certain 

 level. But, instead of reappearing at the same level on the sea- 

 shore, its flow is sitbmarine for a considerable distance from 

 land — a proof that it must have found its way through vei-y deep- 

 lying channels. 



If we duly consider all these conditions, it becomes clear 

 that there is no serious difficulty in the way of the hypothesis 

 that the lagoon extant in Kriangle should have been the result 

 of the action of currents on the porous soil during a period of 

 slow upheaval. 



With regard to the reefof Babelthuap, there are other objec- 

 tions in the way of the subsidence theory ; by far the most grave is 

 the gradual seaward slope of the eastern outer reef. But this 

 and all other difficulties vanish with the supposition that the reefs 



