But it was also explained in Chapter VII that ordinary corals cannot grow in very deep water, and as we find depths of even 200 fathoms just over the edges of these reefs we are confronted with a problem. We have explained the origin of only the surface of the ground, what lies beneath and how it came about that there was a foundation ready, within the narrow limits of depth in which corals could build, continuous through so many hundreds of miles, are the real problems. The barrier reefs, right away from land, from which they are separated by deep water, still more conspicuously need an explanation. These are the main questions of this chapter. _Barrier Reefs._ The barrier system is not a single linear reef, or line of reefs, but rather a line of areas of shallow water full of reefs of all sizes, generally more or less crescentic or ring-shaped. The details have not been surveyed, except very partially in some cases, the charts from which the map on page 126 is copied merely giving the outlines of the areas on which the reefs stand. Some of these areas are very broad, the southernmost, Towartit, being eight miles across, their size, intricacy, and their being completely useless to all navigators but a few pearl fishers, preventing their survey within the outer borders, except in the case of that which bounds the passage to Port Sudan on the north, shewn on the map opposite. This area is obviously a continuation of the barrier system, spite of the fewness of its reefs, and the fact that over the greater part of it an average depth of 10 fathoms obtains. It is a young reef, mostly not yet grown to the surface. _Origin of Barrier Reefs._ The origin of these reefs cannot be explained by any of the theories discussed in Chapter VIII. Darwin’s theory is quite inapplicable as the coast has risen continuously throughout recent geological time, and no currents could have carved out such a channel as that separating the reefs from the land, with such irregular great depths as are shewn on the maps, where, within the barrier depths over 150 fathoms are seen in proximity to soundings of only 30 to 40 fathoms, or even close alongside surface reefs. Finally, coral growth alone, as already mentioned, could not give rise to such sheer precipices as those in which these reefs generally end. The examination of two features of the land makes all clear at once. These are the promontories and sandstone hills, of which the Rawaya peninsula is the best example, and worth describing in some detail. [Illustration: Fig. 84. Approaches to Port Sudan, shewing Sanganeb atoll, portions of the barrier reefs and fringing reef. The latter dotted.] The map opposite shews that Rawaya is a large area joined by a very narrow neck to the mainland and enclosing a large bay, Khor Dongonab, about 20 fathoms deep.