Hence also the almost perfect level of the coast-line, which, spite of “faulting” by which a few small areas rise to a higher level as hills, and the opening of fissures, preserves the same level, within 20 feet, for several hundred miles. At the same time the elevation has been effected in several stages, as evidenced by the existence of level parallel lines of cliff along the sides of hills, e.g. Jebel Zêt in the Gulf of Suez, Jebel Makawar on the Sudan coast, which were cut out by the sea when the hills were at lower levels, and by successive beds of coral at different levels on hill sides in positions that could not be due to tilting of the hill during elevation. Also, at various sheltered points of the shores of Port Sudan and Suakin harbours, and elsewhere, the latest stage of this elevation can be traced in the form of a low cliff, standing a few yards back from the sea, fronted by a reef flat now dry land, though only a foot or two above the sea level. The cliff is undermined exactly as are those still under the influence of the waves, and even the detailed marking of the rock surface characteristic of this marine erosion remains, not yet obliterated by the flaking away of the surface through the action of the sun’s heat and the cold of the clear nights, or by the filing action of the sand blasts of summer. As explained in Chapter VIII, page 107, much of the fringing reef is really a part of the coral limestone of the coast, and its formation needs no further explanation, but it is interesting to note how greatly it varies in width in correspondence with the height of the land behind it. For instance, about Suakin it is up to 1½ miles wide, at Port Sudan only one-third of this, in correspondence with the fact that the shore about Suakin is raised scarcely two feet above sea level, whereas at Port Sudan it rises six to ten feet. This variation is exactly what we should expect on the theory of reef formation by abrasion, the cutting down of the low-lying land involving the removal of comparatively small masses of rock and so proceeding quickly. Again, on the coast about Ankêfail (see Map, p. 139), where the land is as high as at Port Sudan, the reef is only a few yards wide, but this may be partially attributed to the shelter from waves provided by the large island Makawar. That the differences in breadth are so marked shews that abrasion has had much more to do with the formation of the reef flats than has growth of coral, for we see no reason why this latter factor should not have operated equally well over the whole coast and tended to equalise the reefs’ breadths.