As this second summit is at about the right depth below the surface, coral grows vigorously upon it forming a surface reef, slightly hollowed out in the middle. The third summit is like the first. _Plate XXXIX_ [Illustration] [Illustration: Figs. 86 and 87. Two views on Rawaya Upper figure in the north of the peninsula, lower near salt works in the south. Both shew one of the canal-like inlets of the sea which cut up the western side of the peninsula. In the lower one the inlet is partially cut off from the sea and the great heat has evaporated its water leaving a lake of salt. The low ground of Rawaya is shewn behind as a mere line on the horizon in the upper, but in the lower photograph the hill of Abu Shagara is included.] In B the thin line represents the final stage of A, further elevation and abrasion, with coral growth, resulting in the levelling down of islands and reefs and partial filling of the deeper lagoons as shewn by the shaded area of the diagram. Summit No. 1 is not only cut away altogether but hollowed out into a shallow lagoon, the deep lagoon has been narrowed considerably while the ring-shaped reef on summit 2 is much as before, but has spread out and encloses a larger lagoon, thus becoming a small atoll. Summit 3 shares the fate of number one. Now compare the outline of the original hill range in A, with the shaded line in B, and the levelling action of the sea, both upwards and downwards, is evident. This explanation of the origin of the reefs between Makawar and Rawaya can obviously be extended to those to the south as far as the Têlat Islands, and to the whole barrier system in fact. The reefs south of Salak are similarly related to a large area of raised coral extending from the point northwards, and though there is no bay here, corresponding to Khor Dongonab, there is a large salt marsh separating this from other raised coral to the west, and formed by the filling in of a bay by blown sand. The diagrammatic map overleaf makes this clearer, and shews that on land we have continuations of both kinds of reef, the barrier being continued as the eastern coral ridge, the fringing reefs of Salak Seghir being one with the limestone on the west side of the swamp. Similarly Ras Benas to the north (lat. 24° N.) and the angle at the entrance to the Gulf of Suez, have reefs and islands in continuation of them southwards, the former being named Makawar, in this as in appearance recalling the island off Ras Rawaya. It is now evident that the origin of both barrier and fringing reefs is identical with that of the whole coast-land, and is not to be looked for in any laws of coral growth, or marine sedimentation and abrasion, these factors having merely affected the summits of submarine hills hundreds of miles long, nearly two thousand feet high, often peculiarly narrow, and always more or less parallel to the axis of the sea-filled Rift Valley.