Imagine land newly raised from the sea upon which coral growth is only beginning. In section its coastline would be a more or less gradual slope (to take the simplest case) as the line A, B in Diagram 2, sea level being represented by the line C, D. Suppose the scale to be such that the depth C to A is about 50 fathoms. Now it is found that under the best of conditions reef corals do not grow at this depth; if the conditions are less favourable so the maximum depth at which

these corals grow is decreased. (That this fact is the crux of the problems to be discussed later may as well be noted at once.) Coral growth will be most luxuriant in the shallow water, and the first stage of our reef will be a mound of coral of the shape shewn in section by the dotted area. Between E and D this comes to the surface, and the corals, projecting above the water at lowest spring tides, are killed at the top, so that E to D becomes an almost flat surface of dead corals, some of which may, however, be still living where their bases are immersed in clear sea-water. A continuation of this process gives us a reef flat of considerable area, indicated by the line F, D, the slope A, F becoming correspondingly steepened. At the point F the waves have thrown up a long low mound of coral fragments and shells, which, in the way described below, may be consolidated into a ridge of solid rock.

It is easy to see how an extension of coral growth would make A, F a regular precipice, as F approximates to C. What happens after F, the reef edge, grows out to water 50 fathoms deep, where no living foundation can be laid for

the support of the still growing reef above? Diagram 3 explains. Passing seawards from F is the gentle slope formed by the breaking waves, next a precipice, followed by a very steep slope to the sea bottom beyond A formed of the broken and dead corals fallen from the growing zone above, and which forms the foundation on which the shallow water corals extend the reef seawards.

Now at the same time as growth has added to the reef seawards the waves have cut down the land on the other side. Consider this case separately and then combine with that above. As before, A, B in Diagram 4 is the outline in section of recently formed land upon which the sea has as yet had no action, and C, D is the level of lowest tides, C′, D′ that of the highest. Between these two levels, upon the land mass lying between D and D′, is the never ceasing beat of waves and the wear of silt-carrying currents, so that in time the land is eaten away along a line a little above C, D, say X, Y, and a cliff, Y, Z, is formed. We are assuming that the material of the land is sufficiently coherent to form such a flat and cliff, but even so in general X, Y becomes a sloping shore, not a flat. It only remains as a flat if for some reason the seaward surface at X is protected against further detrition by waves, e.g. by the growth of corals and stony seaweeds. If they are present, even if their growth adds nothing to the mass of the rock, it hinders its decay and causes the formation of a reef flat in place of a sloping shore.

Now these two processes, addition by growth and abrasion by wave action, go on simultaneously, and to get at the true method of formation of a reef flat in the Red Sea the two diagrams must be combined, as in Diagram 5, where as before F is the raised reef edge and F, Y the whole extent of the reef flat, and the cliff Z, Y is undermined as shewn. Where, as in some seas, F to D is recent growth and X to Y is rather older coral rock, it is impossible to locate accurately the dividing line between reef formed by recent growth and that cut out of the land, but near shore, where the flat is free from mud and sand, its surface is seen to consist of sections of the constituent shells and corals, cut as cleanly as if done by a stonemason (see Plate XXXIII). Even such hard shells as those of the “giant clam,” Tridacna, are cut across at the same level, thus shewing very clearly the origin of the surface by the planing down of a mass of rock to that level.

The boat channel indicated between F and G remains to be accounted for. As the reef flat widens it results in a great area covered by shallow water at high tide level and partly bare at low. Exposed to a tropic sun life is impossible for any but a few specialised forms, the rock is unprotected from such wave motion as there is, and from boring organisms, which agencies quickly reduce its level. Strong currents also flow over the surface, for the breakers throw water over the raised edge which may have to travel several miles before reaching a gap through which it can return to the sea, and this with tidal currents cause swift rivers of muddy water to flow over the reef flat parallel with the shore. The obvious result is the hollowing out of a boat channel[50], and the accumulation in it of great quantities of mud and sand, which in many places form the greater part of its actual surface, but which eventually are swept out to sea.

The presence of certain marine flowering plants of grass-like form (Cymodocea and other genera) assists, if not wholly responsible for, the formation of such accumulations by binding the mass together by their strong and tangled roots and rhizomes.

When the channel has become broad and deep enough, coral growth may resume sway in it, sometimes to such an extent as almost to block it up again.

I need offer no proof of the formation of a reef flat and precipice by coral growth, the thing is obvious, at least in the case of ordinary fringing reefs. But the hollowing out of the boat channel, and with it that of other lagoons enclosed by coral, is less obvious, and it is natural to assign to a feature so distinctive of coral reefs an origin more directly dependent on the laws of coral growth. The proof comes from a consideration of the simplest case. Do we know of any reefs where solution and abrasion have formed these characteristic features without aid from coral growth? We do.

Plate XXXIII

The east coast of the island of Zanzibar is, like that of the Sudan on the Red Sea, composed entirely of elevated coral. But for some reason this abundant growth ceased shortly after the elevation of the island, and the reef edge bears now nothing but a little stony and filamentous seaweed, and in deeper water forests of sea-grass (Cymodocea). The reef is very wide, up to three miles, the edge regularly raised, and boat channel, as above mentioned, generally well developed. On the raised edge of the reef are numbers of stones, a foot or two in diameter, composed of the same recrystallised coral rock[51] as the shores and cliffs of the island. Now this rock differs very widely from that formed of recent coral in its hardness and weight. Its specific gravity totally forbids the assumption that these stones were torn from the reef by breakers and cast up in their present position above low tide level and indeed, though constantly among them and turning them over to search for specimens of marine life, I never saw one that had recently been moved by the waves, much less broken away from some projection of the submarine precipice.

In fact these are the hardest remnants of the mass of rock which has been removed in the cutting out of the reef, and their presence proves (1) that this was the mode of formation of the reefs, (2) that the addition by growth taking place since the elevation of the old reef has been either nothing or very inconsiderable. Here solution, attrition, and boring organisms alone have carved out from dead rock all the features of a reef which has grown up undisturbed[52].

The present flora of the reef edge may have been preceded by a flora and fauna capable of affording a more efficient protection, as is at present the case in the adjacent and similar island of Pemba, where the reef is narrower and consequently cleaner, and some stunted corals and Tubipora grow on the outer slope of the reef edge, a position where such species are never found in Zanzibar.

The absence of corals from the outer slope of the reef edge is remarkable seeing that they flourish in a few places in the boat channel, so much so in one place as to almost block it up and form a new reef surface. They flourish too round all the many sandbanks and islets of the channel which separates Zanzibar from the mainland of Africa. The mud from the broad reef flat together with the strong currents that impinge upon these coasts are amply sufficient to prevent the settlement of the delicate coral larvae, if not to destroy full-grown colonies.

Another case from the Cape Verde Islands, where reef corals do not exist at all, is shewn on Plate XXXIV. A reef flat, with raised definite edge and miniature boat channel complete, has been cut out of sandstone, the edge of which was protected by a growth of stony seaweed (lithothamnia), and vast numbers of the shelly tubes of that strange animal Vermetus[53]. These two organisms combine to form a continuous crust over the whole surface of the seaward edge of the sandstone, and so greatly delay its removal by the sea, but landwards, this protection being absent, the reef flat is hollowed out into a “boat channel.” This sandstone is a local deposit just to the south-west of the town of St Vincent, but the volcanic rocks of which the island is composed are cut down to a narrow flat in the same way, but less regularly.

Plate XXXIV

A third case, from the Mediterranean near Alexandria, is so striking as to be worth illustrating, though only the embryo of a reef, as it were a ledge a few yards wide, has been formed as yet. The rock is a calcareous sandstone, a consolidated dune, and the protecting organisms are much the same as those found in the Cape Verde Islands, but here forming a less coherent coating to the rock. The regularity of the ledge laid bare by the retreat of a wave is very striking.

Reefs may shew other features, no one arrangement can be taken as typical of all. Instead of the smooth slope and rounded ridge which compose the reef edge on this coast and that of Zanzibar it is usual, in many oceanic reefs, for the growing edge to be cut into by deep and narrow fissures, up which the great breakers send violent torrents of water.

The land, or reef islands, may be either portions of the reef elevated above sea level, containing fossil corals in the positions in which they grow, or it may be partly formed of a mass of corals thrown up by storms backed generally by an accumulation of sand. The coral rock thus elevated may be, as in the Red Sea, but little different from the original material of which it was formed, but more generally it is much altered. The continual wetting by spray or rain and drying under the tropical sun has a very marked effect in hardening and consolidating elevated coral, or coral sand. The upper parts are dissolved, and as the water sinks into the porous corals and becomes supersaturated with lime, the latter is crystallised out, thus filling up all cavities with crystalline limestone. Thus in the end the highly porous heterogeneous limestone becomes a rock of exceeding hardness, crystalline and homogeneous. All the more delicate organisms are dissolved, only the largest remaining recognisable. At the same time as sea-water contains magnesium carbonate as well as limestone, and the former is less soluble than the latter, it tends to be deposited more quickly, so that it comes to replace the original limestone to some extent[54]. The alteration in the external appearance of the rock is very marked. Instead of the yellow, rather shapeless, cliffs of the Red Sea coast, in most other parts of the world, where tides supply spray and there is a considerable rainfall, we have coal-black rock with a very peculiar surface, all covered with sharp points and knife edges separating depressions left by the solution of the stone by water, hence the name “coral rag” applied to such rock. Where it forms the shore of a sheltered bay its homogeneity causes the undermining by the sea to go on to an astonishing extent before the unsupported piece falls away from the cliff to which it is attached. Such projections of the rocks which may be much longer than those shewn on Plate XXXV, also illustrate the hardness of this recrystallised material, for on striking one with a hammer a loud clear bell-like note is produced. Given the right conditions and we have the same peculiar result in the Red Sea and even in the Mediterranean. For instance, a considerable swell breaks at times on the narrow reef fringing the east side of the Tella Tella Kebir Islands, thus keeping the cliff behind it drenched with spray. In consequence the rock has become like that of Zanzibar and British East Africa. And generally, wherever the coral rock is exposed to spray it takes on these characters partially or completely, as is the case at the bases of all the cliffs along a narrow band just about sea level, where the rock is “’twixt wind and water.” Here the outer part is converted into a black, hard, and pitted crust, higher up it is harder than normal but above gradually passes into the slightly altered rock of the normal cliffs. Such a crust also covers the reef flats of the Red Sea, the reef within consisting, as before noted, of loose masses of coral bedded in with shells and sand. A portion of this crust is photographed on Plate XXXIII; the upper surface (Fig. 72) with its sections of contained shells has already been referred to. It is nearly smooth and very hard. The under side of the same fragment is shewn in the next figure and is seen to consist of an irregular mass of shells and coral branches lightly cemented to the crust, from between which the sand, which has not been consolidated, has fallen away. The formation of beach sandstone is practically the same process of cementation, by alternate solution and deposition of lime, taking place in a mass of shell and coral sand instead of larger fragments, the rock following exactly the curve of the sandbank, of which it is obviously a part which has been consolidated in situ.

Plate XXXV

Coral reefs are classified into three sets according to their relation with other land[55].

I. Fringing reefs, which, as the name implies, border the land, are continuous with it, and the seaward edge of which can be reached by wading.

II. Barrier reefs, which run parallel to the coast but separated from it by deep water navigable for coasting vessels larger than canoes.

III. Atolls, ring- or crescent-shaped reefs having no obvious relation to any land and typically found far out in the ocean, from the great depths of which they rise with steep slopes to, at most, a few feet above high tide level.

Fringing reefs we have already dealt with; the two agents described—growth and abrasion of coral—will account for all of them. Barriers and atolls are more puzzling. Why should the barrier form its line parallel to the coast, though at a distance from it, and the very existence of atolls is one of the most striking phenomena of Nature.

The problem is complicated by the fact that ordinary reef corals die out at a depth of 50 fathoms or so. Now 50 fathoms is a mere nothing compared to the depths from which the Pacific atolls rise, and is only a quarter the depth often found within a few hundred yards of the Red Sea reefs. How then to account for the building of reefs in deep water?

One suggestion was that atoll rings were formed by the growth of a mere cap of coral round the edge of the craters of huge submarine volcanoes. But that postulates far too large a number of such immense volcanoes[56], and the early stages of these formations have not been found. Darwin’s hypothesis was hailed with joy as the obvious solution, and held the field against all rivals for many years. Briefly it is that corals formed a reef by direct growth in shallow water on the coast of an island, forming a fringe thereto in the way explained above. Now is postulated one of those great, slow earth movements such as have very often occurred in the past and are occurring at the present day. In this case the island is to sink slowly, at such a rate that the reef grows upwards as fast as it is submerged. The result is obviously a mass of corals of a thickness equal to the total sinking movement of our island, though every individual coral grew while in water under 50 fathoms deep.

When our island is half submerged the fringing reef has become a barrier, when wholly gone the reef ring remains enclosing an empty lagoon, and is the only mark of the grave of a drowned island. Thus Darwin’s theory has the further merit of referring the two forms of reef, barrier and atoll, to one common cause, the sinking of the land. But we have no idea of how the original islands were formed in such numbers, and many believe that no such vast sinking of the ocean basins has occurred since they were formed. Also, if solution be ignored, it is difficult to see why, as the island sank, coral growth did not close in over the submerged land, and so form a vast reef flat instead of leaving a lagoon up to 50 fathoms deep.

To settle the matter an expedition was sent to a typical atoll, Funafuti, and a boring 1200 feet deep was made to find out what the interior of the reef is made of. The material brought out of the bore hole has been carefully examined by experts, and reported to consist of the remains of exactly similar corals to those found near the surface, and this result was taken by one or two geologists as complete vindication of Darwin’s theory. But apart from the extreme difficulty of the identification of all coral species, especially those which have been subject to partial crystallisation and so on, one remembers that a considerable part of the foundations in deep water are formed of corals which have fallen down the steep slope from the growing reef above, so that their presence buried a thousand fathoms deep proves nothing, while the boring at Funafuti only went to about 200 fathoms.

After all it is easier to imagine that the atoll grew up from the bottom of the deep sea. The only postulate is a chance elevation on the sea bottom. On such elevations it is found that the remains of marine organisms, including deep sea corals (as distinct from reef builders), tend to accumulate much more rapidly than on the floor of the surrounding depths. The elevation is consequently slowly but surely raised, and the higher it grows the more rapid the accumulation, until at last reef corals obtain a footing forming a cap or pinnacle reaching to the surface. From this masses of coral, sand, stones or large boulders, are always falling on to the foundation slopes, forming successive sloping layers indicated by the dotted lines, upon which fresh growth takes its rise[57]. When the coral reef has become of some breadth (and atoll rings may be 30 miles or more across) a boring at the edge might descend for a thousand fathoms and never meet with the original foundations, but would pass only through recent corals fallen from the shallow zone.

We should expect to find a continuous surface of coral at sea level. As a matter of fact there is a broad lagoon, generally of considerable depth, one or two gaps through the encircling reef giving communication with the open ocean. This is the natural result of the causes described when dealing with the boat channel of a fringing reef; it is the same thing on a much larger scale. Seeing that the rate of growth of coral masses is always only the excess of growth over destruction and solution, the presence of growing corals is no evidence against the fact that the lagoon shores may be undergoing destruction, and that such coral growth as is present may add nothing to the inner sides of the reef in the end. No more does the accumulation of great quantities of mud prove that the lagoon will in time be quite filled in. Mud and sand[58] are but stages in the destruction of coral rock, and its presence where that process is going on is to be expected. An abnormal tide, a shift of the currents, and vast quantities are swept out through the gaps in the reefs. My home on the Red Sea is beside a large landlocked lagoon in which coral gardens of great luxuriance, whence collections of many species can be procured, are frequent. Spite of this, the evidence is as clear as possible that its shores and islands are undergoing rapid denudation, and its reefs are being cut down by currents to banks below water level. As in the Red Sea the level rarely alters by more than a foot once in the twenty-four hours, and often the rise or fall is much less, the action of tidal currents is at a minimum, yet even so they produce well-marked effects.

Barrier reefs may be formed from fringing reefs by the enlargement of the boat channel, while the reef is extending seawards.

The island of Zanzibar, 60 miles long by 20 wide, and 20 miles from the mainland of Africa, seems to be a part of the East African barrier system, and it certainly was separated from the mainland by the destruction of the intervening land; the shallow dividing channel being full of shoals and sandbanks formed by cutting down of islands. The fauna of Zanzibar, including leopards, serval cats, &c., can be accounted for in no other way. The Great Barrier of Australia, a thousand miles long, is the same thing on a vastly greater scale. But, as described in the next chapter, the Barrier system of the Red Sea is quite another thing, and its mode of formation may possibly be unique in the world.

CHAPTER IX
GEOGRAPHY OF RED SEA AND FOUNDATIONS OF THE REEF SYSTEMS

There are two seasons when rain may be hoped for, viz. the “kharîf” which centres round August, and which is referred to in the frontispiece, and the winter months, but if rain fell for an hour or two on three days it would be considered a liberal supply for the whole year in most places. At Dongonab there has been no rain (above a millimetre or two) since December, 1907, though one or two showers have fallen on Rawaya and Makawar[59]. There is of course much more rain on the hills than on the plains, but even so grass grows only in scattered areas to which the people migrate.

Tides. The Red Sea undergoes considerable variations of level at its extremities, up to seven feet at Suez, but in the middle the variations are small, only a few centimetres at Port Sudan. At Dongonab the difference between highest and lowest levels recorded is 80 cm., but the maximum change in any 24 hours is rarely over 30 cm. Records shew a distinct tide, but this may be interfered with by changes of level due to wind and changes of atmospheric pressure, and in any case one of the usual two tides of the 24 hours is practically suppressed, the water remaining near high tide level until it falls for next day’s tide. In the summer the average level is lower than in winter and the tidal effects are partially masked by the results of the peculiar climatic conditions. The water may remain low for days, so that all the coral which has grown above that level since the last occasion of extreme low water, which may have been one or even two years ago, dies off.

I suppose that every school-boy looking at an atlas, is struck by the peculiar shape of the Red Sea, and is led to ponder on the usefulness of this peculiar canal, the sole value of which is that it gives communication between Europe and the East, a value which needed but the trifling addition possible to human effort to make it the great highway of the world. Its own shores are desolate wastes, in itself it has no attraction for traffic, and even its shape seems to indicate that it is but a passage to other seas. (See map inside the cover.) For so narrow a sea, only a little over a hundred miles wide, the depth is great, two hundred to five hundred fathoms at the side and a thousand in the middle. These peculiarities are also well marked in the deep Gulf of Akaba which bounds Sinai on the east—the Gulf of Suez, on the west, being a shallower branch valley. Both these gulfs, like the Red Sea, are bounded on either side by high mountains, and those of the southern part of Sinai are particularly grand in the savage barrenness of their jagged peaks and vast precipices.

The Gulf of Akaba is directly in line with the Jordan Valley, a similar depression on a smaller scale, only partially occupied by water, the Dead Sea, while southwards we find another dry valley running through British East Africa and adjoining territories, a great trough bounded by plateaux, several thousand feet above its bottom. We can thus trace

this trough-like valley from Palestine to some degrees south of the Equator as a stupendous crack in the earth’s surface, well named “The Great Rift Valley[60].” The Red Sea is its greatest section, its total depth here being, say, 5000 feet from the summit of the mountains[61] to sea level and 6000 feet to the sea bottom, 11,000 feet in all.

The formation of such a valley, by the dropping down of a series of strips of country below the level of the remainder, is illustrated by Diagram 7. To study the simplest possible case we draw a section through the ground and imagine it formed of three kinds of rock, of which two form horizontal sheets, AA and BB, over the third CC. These were originally unbroken, and in the positions shewn by the lines of dashes, but were broken by the dropping down of the central part to form the valley shewn here in section. The floor of the valley has the same structure as the original surface of the ground, the same three beds, A, B and C, occurring in the same positions, but at a lower level. They are found again in each of the steps on the valley’s sides, their regular reappearance in this way being conclusive proof of the earth movements postulated.

The vertical lines FFFF, between each step and the next drop, along which the continuity of the beds is broken, are termed “faults,” a geological term which should be remembered.

Rift valleys are found elsewhere in the world, but are exceptional, ordinary valleys, with their winding courses and rounded outlines, having been formed by the action of streams, which slowly wash away the ground and hollow out their courses to the sea.

The actual structure of the middle portion of the Red Sea Valley is shewn diagrammatically by the section on page 145. Five steps are shewn, Nos. 2 and 3 being further separated by a minor fault valley. The details are described later.

The southern part of the sea, below Masawa, has recently been subjected to volcanic[62] action; many of the islands there are quite well-preserved volcanic cones, but as regards the rest of the sea, though earth movements have been frequent and considerable, there are now no traces of volcanic action, and the movements that have occurred have not necessarily involved cataclysms greater than severe earthquakes.

There are however in the north two islands the existence of which is most readily explained by volcanic action. I refer to the coral formations known as “The Brothers” and “Daedalus Shoal,” the former a pair of low islets, the latter a flat reef, rising out of the centre of the sea and surrounded by water hundreds of fathoms deep. They are extremely steep-sided cones, and what could form and support such structures far out from land is puzzling. A certain view of another section of the Rift Valley, that once seen can never be forgotten, seems to offer an explanation. After passing through the forests of the Kikuyu Plateau by the Uganda Railway one comes out into the open on the brink of the great escarpment of the Rift Valley and looks across a trough 3000 feet deep to the similar forest-clad heights of the Mau on the other side. The continuity of the valley is rudely broken by two volcanic cones rising abruptly in the middle of the flat bottom of the trough. On consideration the strangeness of their appearance in the middle of a valley passes away, one sees that the bottom of such a rift must be a zone of weakness of the earth’s crust where volcanoes might naturally be expected to arise.

If the water were removed from the Red Sea Valley would not the appearance of The Brothers and Daedalus be very much like that of the two volcanoes of British East Africa, allowing for the steeper angle at which their materials would lie under water? Given such cones of loose volcanic ash, &c., wave action would quickly level down their summits until coral growth afforded protection and formed a cap of rock, part of which is now raised again above sea level as the islands on one of which the lighthouse is built.

The ring-shaped reef of Sanganeb[63], opposite Port Sudan, which is outside the Barrier system and separated from it by water 400 fathoms deep, may be built on a similar foundation. Like the two coral reefs above it rises with extremely steep slopes from this deep water, and is the summit of a submarine pinnacle rather than hill.

On the other hand, the foundations of these strangely isolated reefs may be like a certain island which, rising high above sea level, shews its structure, a centre of olivine rock fringed with coral. This island is variously known as Zeberjed, St Johns, and Emerald Island, the latter name due to its possession of mines for peridots, which are worked by the Khedive of Egypt. Its position is 23° 30′ N., distant about 60 miles from the African coast, a formation quite independent of the sides of the Rift Valley. It is an example of the “Block Mountains” described by Professor Gregory, portions of the original earth surface which have remained standing when the surrounding country dropped down to form the trough of the Rift Valley, not a mass of land thrust upwards and subsequently carved into peaks and valleys by running water, which is the way ordinary mountain ranges are formed.

One gets a good idea of the structure of the Red Sea coasts on leaving the Gulf of Suez for the voyage south, before the ship’s course passes far from land. On the western horizon is a range of wild mountains, a grey plain ending in a yellow shore-line separating them from the sea, and the off-lying islands are of the same colour. The plain is formed of gravel from the high hills, its yellow border seawards being coral limestone, and the islands also. In the sea are numerous reefs, here of very intricate plan, lines of white breakers separating the deep blue black water from large areas of green and brown shoals in waveless lagoons. There are deep channels between these reefs and the shore, which is itself fringed by a shallow reef with its edge at low water level but bearing perhaps one or two fathoms of water on its surface within.

This being the simple structure of both sides of the whole Red Sea trough I may proceed to describe in detail one section of the coast, that bounding the territory of the Anglo-Egyptian Sudan, between 18° and 22° N. This section includes two (Ras Rawaya and Ras Salak) of the three promontories which break the straight line of the west coast north of Masawa, the third being Ras Benas, further north. The map opposite shews clearly the fringing reef which lies along the whole coastline, the numerous harbours, of which Port Sudan, Suakin, and Trinkitat are of commercial importance[64], the deep channel separating the fringing from the barrier reefs, and the atoll of Sanganeb on which the lighthouse is built.

On land the bases of the high mountains are indicated, and certain lower hills, of sandstone, which rise in the midst of the maritime plain. A striking fact is visible on first inspection of this map, viz. that not only is the Red Sea a nearly parallel-sided trough but that the constituents of the sides are themselves placed in lines parallel to the coast. The Archean hills[65], the lesser sandstone ranges, the coral bounding the maritime plain, and the barrier reefs, are all four roughly parallel to the main axis of the sea.

We will consider each feature in more detail. For the Archean hills consult the extremely interesting memoirs of the Egyptian Geological Survey[66]; for our purposes it is enough to note that they are all of ancient igneous and metamorphic rocks, that they rise to heights of from four to eight thousand feet, and the valley bottoms are generally flat and filled in with gravel.

The maritime plain is from five to ten miles wide, sloping up regularly from the sea towards the bases of the hills, where it may attain a level of several hundred feet. Except at its seaward edge, it is composed of black gravel, the product of the decay of the hills carried down by the torrents resulting from the rare but furious rain-storms, and spread out to form the plain. Sand-hills occur, but not very commonly, though the gravel is mingled with sand throughout, and in sections of the plain exposed by wells, layers of gravel alternate with sand, fine or coarse, as far as the deepest borings have been carried[67].

The pebbles, though black predominate, are of a most remarkable variety of kinds and colours. Bright green and red, yellow and clear white are abundant, and any square yard would yield a rich collection in Petrology. As the torrents open out into the level plain they lose themselves, continually taking to fresh channels, so that the débris from series of hills quite distant from one another are mingled; in a given spot gravel from one valley is laid down this year, from another and totally distinct one another. One would expect gravel which had been carried by torrents a distance of many miles to be rounded down by friction into smooth boulders or pebbles, like those of our home streams. As a matter of fact it is nearly always angular, the rounded surfaces we should expect being rarely met with on the surface. The pebbles, as we now see them, have been re-formed from larger stones since their transport through the valleys and over the plain. Large stones, lying half buried in smaller material, shew the usual rounded surfaces of water-borne rock, but they are invariably split up by fissures, which may be half an inch broad, so that the stone is as it were built up of angular fragments fitted together after the style of a puzzle picture. During the hundreds of years they have lain there, apparently secure from all interference, they have been exposed to innumerable fierce heats and cold nights, which, causing successive minute expansions and contractions, have at last split the stones into small pieces. This is the origin of the irregularly shaped gravel; first indeed it was rounded by the grinding and pounding of the torrents of hundreds of successive winters, then it was split up again by the silent invisible stresses of heat and cold.

Plate XXXVI