The Origin of Continents and Oceans/Chapter 12

by Alfred Wegener

CHAPTER XII

THE CONTINENTAL MARGIN

On the margin of a continental block there lies below the floor of the ocean an approximately vertical parting plane between the sial and sima which does not correspond to the natural layer-like arrangement of light and heavy materials, but which only exists because of the solidity of the block of sial. On this account special forces are exerted here which strive to bring about the natural arrangement of the masses, and are therefore in opposition to the molecular forces of the block. In this connection there are a whole series of phenomena, which will be dealt with in the following pages.

As Schiötz first recognized from the gravity measurements taken on the Fram which was drifting over the margin of the continental shelf of the Polar seas, and Helmert^[1] deduced in detail later, the pendulum observations on the margins of the continental blocks show a characteristic disturbance of gravity, which is reproduced from Helmert in Fig. 37. As the coast is approached from the land the gravity increases to a maximum on the shore-line itself, then rapidly falls and reaches a minimum at the place below which the floor of the deep sea begins, after which the normal value appears again at greater distance from the coast-line. The occurrence of this disturbance of gravity can be represented in somewhat the following way. The observer on the land, who inland has obtained normal values, obtains a maximum at the coast, because he approaches the heavy sima of the ocean-floor, which lies obliquely beneath him. To be sure, this excess of gravity should be counterbalanced by the fact that the uppermost 4 km. are replaced by the less dense sea-water.
Fig. 37.—Disturbance of gravity on the margin of a continent, after Helmert. But these layers lie by the side of and not beneath the observer, and thus instead of reducing gravity again to its normal value, cause a deviation of the plumb-line, in the nature of an attraction of the plumb by the continental platform. The observer on the sea, on approaching the coast, experiences the reverse. The amount of gravitational force is affected by the diminution of mass beneath him; the increase of mass on the land side of him cannot, however, influence the amount, but only the direction of the force of gravity, so that a minimum of gravity results.

An island or group of islands which forms the summit of an isolated fragment of sial swimming in the sima must manifestly be surrounded by this gravity disturbance in a ring-like manner. On account of this, the gravity is greater than normal on the island itself, and especially so on the shore, whilst outside it there is a circular area on the sea where it is below normal. The observation that was early made that the pendulum measurements on islands yield gravities greater than normal is explained in this manner. The view of many authors that the Pacific islands are purely volcanic cones which are simply mounted on the floor of the ocean, and are supported by it, cannot be established by gravity measurements. These favour the ideas advocated, for example, by Gagel for the Canary Islands and by Haug for many Pacific isles, that all these islands are pieces of the sialsphere, but that they are in many cases so completely covered with lava that their sial cores are nowhere exposed.

These circumstances may be considered in yet another light, and one which is adapted to illustrate directly their effects. In a continental block, the pressure must manifestly increase with the depth, according to a law different from that which prevails in oceanic areas. If we compare the pressures^[2] at similar depths, we find that all over the continental block—with the exception of its upper surface and its under surface—the pressure is greater than in oceanic areas. If we take as a basis the numerical proportion assumed in Fig. 5 (p. 30), we obtain as the values for this excess of pressure in the continental platform:—

• At 000,100 m. elevation excess of pressure is 000 atmospheres.
• At„ 000,000 m.„ elevation„ excess„ of„ pressure„ is„ 028 atmospheres.„
• At„ 004,700 m.„ elevationdepth excess„ of„ pressure„ is„ 860 atmospheres.„
• At„ 100,000 m.„ elevationdepth„ excess„ of„ pressure„ is„ 000 atmospheres.„

The excess of pressure thus increases very rapidly in the uppermost portions because there rock is in contrast to air; in the next section only about two-thirds as quickly, since here water is present in the oceanic area. The maximum excess of pressure is reached at the depth of the ocean-floor. At still greater depths it will become smaller again, since now the heavy sima lies in the oceanic area, and causes there an accelerated increase of pressure; and on the under-surface of the continental blocks the pressures must naturally have been equalised.
Fig. 38.—Effect of the pressures on the margin of the continent (diagrammatic). This difference of pressure produces a region of stress on the vertical continental margin, which is struggling to press out the material of the continental platform into the oceanic space, and most of all in the layer that forms the deep-sea floor of the ocean.^[3] If the sial were mobile, it would spread out in this layer. Now that is not the case. But the sial is, however, plastic enough to yield noticeably to these enormous forces, as is well shown in the step-like faults, which as a rule accompany the margin of a continent (Fig. 38). This lateral forward flow of the deeper plastic layers is also the reason for the fact that the margins of blocks split asunder and widely separated, as South America and Africa, have the parallelism far better preserved in their coast-lines than in the border-line between the continental slope and the deep-sea floor.

It is not inconceivable that the vulcanicity occurs so frequently near the coasts, because the inclusions of sima of the block are pressed out by the field of stress described above. This explanation is more especially possible in the case of oceanic islands which are surrounded in a circular manner by this field of stress.

A special kind of force must occur on the margin of the plastic continental blocks when these are loaded by a covering of land-ice. If one loads a plastic cake, it will, in its effort to reduce its thickness and extend horizontally, become affected by marginal cracks. This is the explanation of the formation of fiords, which exist in astonishing uniformity on all formerly glaciated coasts (Scandinavia, Greenland, Labrador, the Pacific coast of North America north of 48° and of South America south of 42°, as well as the South Island of New Zealand), and have been already traced back by Gregory, in an extensive yet much too little valued examination of them, to the formation of faults.^[4] On the basis of my own observations in Greenland and Norway, I believe that the explanation of fiords as erosion valleys is incorrect, though it is still much advocated at the present day.

By a great number of soundings on the Atlantic continental margins, attention is drawn to a peculiar phenomenon which appears to indicate the submarine continuation of river valleys. Thus the valley of the St. Lawrence is prolonged into the shelf in front down to the deep sea; similarly that of the Hudson may be followed to a depth of 1450 m. The same is the case on the European side, before the mouth of the Tagus, and specially in the “Fosse de Cap Breton,” 17 km. north of the mouth of the Adour. But the most beautiful phenomenon of this kind is certainly the Congo Channel in the South Atlantic,^[5] which may be traced to a depth of 2000 m. According to the customary explanations, these channels are drowned erosion valleys, which had their origin above the water. This, however, appears to me to be extremely improbable, first on account of the great amount of submergence, secondly on account of their general distribution (by a sufficiently large number of soundings they will probably be found on all continental margins), and thirdly because only a certain selection of river mouths show the phenomenon, whilst those lying between them do not. I believe it to be much more probable that we are here also dealing with rifts in the continental margin which have been used by the rivers. In the case of the St. Lawrence, this fissure-nature of its bed is, as a matter of fact, proved geologically. In the case of the “Fosse de Cap Breton,” which forms the innermost end of the deep-sea rift of the Bay of Biscay, with its book-like opening, its very position renders this explanation plausible.

But island festoons form the most interesting phenomenon of the continental margin, and are especially well developed on the east coast of Asia (Fig. 39). If we consider their distribution in the Pacific Ocean, we see a system developed on a large scale. This is especially the case if we conceive of New Zealand as a former festoon of Australia; then the whole west coast of the Pacific Ocean will be covered with island festoons whilst the east coast is free from them. In North America still undeveloped beginnings of the formation of festoons can perhaps be detected in the separation of islands between latitudes 50° and 55°, the bulging of the coast near San Francisco, and the separation of the Californian coastal ranges. In the south West Antarctica might possibly be claimed as a festoon (in this case probably a double festoon).
Fig. 39.—Festoons of North-east Asia.
(Depth contours, 200 and 2000 m.; ocean deeps, dotted). As a whole, the phenomenon of the festoons points to a displacement of the continental masses west of the Pacific Ocean, directed towards the west-north-west, therefore about due west for the position of the poles in Pleistocene time. This direction also coincides with the long axis of the Pacific (South America-Japan) and with the dominant direction of the ancient Pacific island rows (Hawaian Islands, Marshall Islands, Society Islands, etc.). The ocean deeps, inclusive of the Tonga deep, are arranged as rifts perpendicular to this displacement direction, and thus parallel to the festoons. There can certainly be no doubt that all these phenomena are in causal connection with one another. If we take a circular sheet of rubber and draw it out in one direction, we obtain a similar picture. One diameter increases, the other diminishes. By the stretching of the rubber, all groups of points (groups of islands) are drawn into chains parallel to the length, and rifts are torn open perpendicular to the direction of tension. The island festoons of East Asia are thus in close relation to the structure of the whole Pacific Ocean.

Absolutely similar festoons are present in the West Indies; and the arc of the South Antilles between Tierra del Fuego and Graham Land can also be claimed as consisting of a free festoon, although with a somewhat different significance.

The uniform arrangement of the festoons in echelon is very striking. The Aleutian Islands form a chain which farther east in Alaska is no longer a coastal range, but comes out from the interior. They end near Kamtschatka, and thenceforward the Kamtschatka chain, hitherto an interior range, and the Kuriles, form, as the outermost chain, the festoon. This ends in turn near Japan to give place to the Sakhalin-Japan chain, which up to that point is an inland range. This arrangement can be still further followed south of Japan until the relations become confused near the Sunda Islands. The Antilles also show exactly the same arrangement. It is obvious that this echeloned formation of the festoons is a direct consequence of the arrangement in echelon of the former coastal mountain chains of the continents, and thus goes back to the general law of the echeloned succession of folds described above. The strikingly similar length of the arcs (Aleutian 2900, Kamtschatka-Kuriles 2600, Sakhalin-Japan 3000, Korea-Riu-Kiu 2500, Formosa-Borneo 2500, New Guinea-New Zealand formerly 2700 km.)^[6] could perhaps have been predetermined in this fashion tectonically by the former structure of the coastal mountain systems.

The remarkable similarity of the festoons in their geological structure has been already mentioned. The concave side always bears a row of volcanoes, obviously a result of the pressure, which results from the bending and forces out the inclusions of sima. On the other hand, the convex sides bear Tertiary sediments, whilst these are usually absent on the corresponding shore of the mainland. This signifies that the separation first took place in very late geological time, and that the festoons still formed the margin of the continents at the time of the deposition of these sediments. These Tertiary deposits everywhere show great disturbances of the strata as a result of the tension that occurs here on account of the bending, and leads to fissuring and vertical faulting. Nippon (Hondo or Honshiu), Japan, has been broken up, in the “Fossa Magna,” by the excessively strong bending. That this outer margin of the festoon appears to be elevated, in spite of the depression connected everywhere else with extension, indicates a tilting movement of the festoon, which one can believe to be caused by the fact that it is dragged along at its ends by the westerly drift of the continental blocks, while deep down it is held back by the sima. The ocean deep that usually accompanies the outer margin of the festoons is apparently connected with the same process. Attention has already been drawn to the fact that these ocean deeps never form on the freshly exposed surface of the sima between the continent and festoon, but always only on the outer margin of the latter, and therefore on the margin of the old ocean-floor. It appears as a rift, one side of which is formed of the greatly cooled ancient deep-sea floor, which is already solidified to great depths, and the other side of the sial material of the festoon. The formation of such a marginal fissure between the sial and sima would be very intelligible, and quite in keeping with the tilting of the festoon already referred to.

The bulging outline of the continental margin behind the festoons is strikingly shown in Fig. 39. It is seen, especially if we consider the 200 m. depth contour in addition to the coast-line itself, that the continental margin is always the reflected image of an S-shape, whilst the festoon lying in front forms a simple convex curve.
Fig. 40.—Manner of origin of island festoons.
A, Section. B, Plan. (The greatly cooled portion of the sima is denoted by dashes.) This relation is diagrammatically illustrated in Fig. 40, B. The phenomenon is developed in the same manner by all three of the festoons contained in Fig. 39, and is also shown for example by the continental border of East Australia and its former festoon, formed by the south-eastern prolongation of New Guinea and New Zealand. These curved shore-lines denote a compression parallel to the coast, and thus also to the direction of strike of the coastal mountains. They are to be considered as great horizontal folds. We are here dealing with a phenomenon forming part of the powerful compression which the whole of eastern Asia has undergone in a north-east and south-west direction. If the attempt is made to smooth out this sinuous line, the distance between Further India and the Bering Straits, which at present amounts to 9,100 km., would be increased to 11,100 km.

Thus, according to our idea, the island festoons, especially those of Eastern Asia, are to be considered as marginal ranges, which, by the westerly drift of the continental masses, become separated from them, because those festoons remain attached to the deeply-solidified ancient floor of the sea. The younger and more mobile ocean floor crops out in a window-like fashion between them and the continental margin.

This is a different theory from that which F. v. Richthofen^[7] has put forward, starting, of course, from quite other assumptions. He thought that the festoons originated through a tensile force in the crust of the earth coming from the Pacific. The festoons of islands, together with a broad zone of the neighbouring mainland, also characterized by the curved course of the coast and of the elevations, would form a great system of fractures. The area between the chain of islands and the coast of the mainland would be the first continental step which would be submerged below the surface of the sea on the west, as a consequence of a tilting movement, whilst the eastern margin would project as an island-festoon. Richthofen believed he could recognize two further similar steps on the mainland, the depression of which, however, was less. The regular arc-like form of these fractures certainly raised a difficulty; nevertheless, it was thought that by referring to curved cracks in asphalt and other substances this objection could be invalidated.

Though it must be freely acknowledged that this theory possesses the merit historically of being the first to break consciously with the dogma of a universally effective arching pressure, and to call in tensional forces for the explanation of earth structures, yet little time is needed to show that it does not fit in with present-day knowledge. The ocean charts, incomplete as they are owing to the absence of soundings, show decisively that the connection is quite broken between the festoon and the main block.

If the movement of the continental block did not, as in Eastern Asia, occur at right angles to its margin, but parallel to it, the coastal chains could be stripped off by lateral displacements without the appearance of a window of sima between them and the main block. The basic principles are the same as for the similar phenomena which were illustrated for the interior of the continental blocks in Fig. 33 (p. 166), but mentally transferred to the continental margin. If the block moves towards the sima, then marginal folding occurs, and may be accompanied either by overfolds or thrusts or by echeloned folds, according to the direction or the movement. If it moves from the deep-sea floor, then the coastal chains will split off. But if the movement is a shearing one, then we have a fault with lateral displacement, and the marginal chain slides longitudinally. In this case also, the chain adheres to the solidified deep-sea floor. This process is especially well seen in the depth chart of the Drake Straits in Fig. 14, on page 71, at the north end of Graham Land. Similarly the most southerly chain of the Sunda Islands (Sumba-Timor-Ceram-Buru), which has probably at one time formed the south-easterly continuation of the islands lying in front of Sumatra, has slipped past Java, until it was arrested by the advancing block of Australia-New Guinea.

California is another example. The Californian peninsula shows dragging phenomena on its lateral projections which appear to demonstrate a progressive impulse of the land-mass towards the south-east. The point of the peninsula is already thickened to an anvil-shaped form by the frontal resistance of the sima, and the peninsula seems as a whole to be much shortened, as is shown by a comparison with the outlines of the Gulf of California. According to Böse and Wittich,^[8] the most northerly portion has only recently been elevated out of the sea and that to an amount of as much as 2000 m. a clear sign of strong compression.
Fig. 41.—California and the earthquake fault of San Francisco. From the contours it can scarcely be doubted that the southern point actually lay in former times in the notch of the Mexican coast which is in front of it. The geological maps show that at both places post-Cambrian intrusive rocks exist, the identity of which has however not yet been demonstrated.

But besides the shortening of the peninsula itself, there appears to be a sliding towards the north^[9] in which the coastal chains immediately to the north also take part. The great bulge of the coast-line near San Francisco is explained in this manner by compression. This idea is confirmed in a striking way by the well-known fault connected with the earthquake of San Francisco on April 18th, 1906, shown in Fig. 41 (taken from Rudzki).^[10] The eastern portion was thrown southwards, the western northwards by it. The survey measurements, as was to be expected, showed that the amount of this sudden displacement gradually became less with increasing distance from the fracture, and at a great distance was no longer recognizable.
Fig. 42—Movement of one of the surface elements intersecting the fracture, after Lawson. The crust of the earth had naturally been in slow continuous motion before the displacement at the fault. Andrew C. Lawson has compared this movement between the years 1891 and 1906, with the direction of the fault movement.^[11] He arrives at the conclusion based on the Point Arena group of observations, and shown in Fig. 42, that an object on the surface above the latter fracture has moved in the fifteen years about 0.7 m. from A towards B, then was divided by the formation of the fracture whereby the western half was thrown about 2.43 m. towards C and the eastern about 2.23 m. towards D. The continuous movement between A and B, which must be regarded as relative to the main mass of the North American continent, shows that the western margin of the continent is kept back continually to the north by attachment to the sima of the Pacific. The fracture only signifies a sudden adjustment of the stress, and does not move the continental block as a whole.

In this connection we might refer to another also very interesting portion of the earth’s crust, which has certainly been but slightly investigated, namely, the continental margin of Further India.
Fig. 43.—Depth chart of Further India.
(Depth contours, 200 and 2000 m.; outline of ocean deep, dotted). The deep sea-basin to the north of Sumatra is of special interest in this connection. The kink in the Malacca peninsula corresponds to the abrupt northern termination of Sumatra; but it is not possible again to cover up the window-like uncovering of the sima-sphere recognizable north of this island, by again straightening out the Malacca peninsula. That is at once shown by the island-chain of the Andamans, which lies before the window. We must obviously here assume that the great compression of the Himalayas has exerted a pull on the ranges of Further India in the direction of their length, that under the influence of this pull the Sumatra range has been torn off at the northern end of the island, and that the northern portion of the chain (Arakan) has been drawn northwards like an end of a rope into the great compression, and is still being so drawn. Planes of differential movement must have been formed by this process on both sides of this enormous horizontal displacement. It is interesting that the outermost marginal range, the Andaman and Nicobar Islands, remained fast in the sima, and it was only the second chain that experienced this remarkable displacement.

Finally, the well-known difference between the Pacific and Atlantic types of coast may be considered. The Atlantic coasts display fractures of a plateau-land, whilst those of the Pacific are distinguished by marginal mountain chains and the presence of ocean deeps in front of them. The coasts with the Atlantic structure also include those of East Africa with Madagascar, India, West and South Australia, as well as East Antarctica, whilst with the Pacific type there are those of the west coast of Further India and of the Sunda Archipelago, the east coast of Australia with New Guinea and New Zealand, and West Antarctica. The West Indies, including the Antilles, have a Pacific structure. A different behaviour of the force of gravity also corresponds to the tectonic distinctions between these types.^[12] The Atlantic coasts are, apart from the disturbance of the continental margin described above, isostatically compensated; that is, the floating blocks are in equilibrium. On the other hand, deviations from isostasy prevail on the Pacific coasts. Further, it is known that the Atlantic coasts are relatively free from earthquakes and vulcanicity also, whilst the Pacific are rich in both. Where a volcano does occur on a coast of Atlantic type, then the lava, as pointed out by Becke, has systematic mineralogical differences compared with the Pacific lavas, and is notably heavier and more ferruginous, thus appearing to originate from greater depths.^[13]

According to our ideas, the Atlantic coasts are in all cases those that have been formed through the splitting of the blocks, for the first time in the Mesozoic and in places considerably later. The floor of the sea in front of them thus exhibits a relatively freshly exposed surface of sima, and must on that account be considered as relatively fluid. On these grounds it is not surprising that these coasts are isostatically compensated. Further, on account of this greater mobility of the sima, the continental margins experience only a slight resistance to displacement, and therefore become neither folded nor squeezed, so that neither marginal mountain chains nor volcanoes occur. Earthquakes are also not to be expected here, since the sima, by its power to flow, is mobile enough to render all necessary movements possible, without discontinuity. Expressed in an exaggerated manner, the continents act in this case as solid blocks of ice in fluid water.

1. F. R. Helmert, “Die Tiefe der Ausgleichfläche bei des Prattschen Hypothese für das Gleichgewicht der Erdkruste und der Verlauf der Schwerestörung vom Innern der Kontinente und Ozeane nach den Küsten,” Sitzber. d. Kgl. Preusz. Akad. d. Wiss., 18, pp. 1192–1198, 1909.
2. Strictly speaking, the vertical pressures. According to Rudzki there are altogether six pressures acting on a cubical mass of a solid body, namely, three normal to its walls (compressive), and three tangential (shearing). Expansion (dilatation) can be conceived as being a negative compression, so that the pressure can be positive or negative. In our case shearing forces will be assumed to be non-existent.
3. The relations are exactly the reverse of those set forth by Willis, when he assumes an advance of the heavy oceanic rocks against the deeper layers of the continental blocks (Research in China, 1, p. 115, etc. Washington, 1907).
4. J. W. Gregory, The Nature and Origin of Fiords, pp. 1–542. London, 1913.
5. See the map given in Schott, Geographie des Atlantischen Ozeans, p. 102. Hamburg, 1912.
6. The West Indian arcs show, however, a gradation: Lesser Antilles-South Haiti-Jamaica-Mosquito Bank 2600, Haiti-South Cuba-Misteriosa Bank 1900, Cuba 1100 km.
7. F. v. Richthofen, “Über Gebirgskettunge in Ostasien. Geomorphologische Studien aus Ostasien,” iv, Sitzb. d. Kgl. Preuss. Akad. d. Wiss., Berlin, Phys.-Math. Kl., 40, pp. 867–891, 1903.
8. Short communication from E. Böse. The paper is to be found in the Parergones del Instituto Geologico de Mexico.
9. Or a lag of the peninsula in respect of a southerly movement of the mainland relatively to the sima.
10. Rudzki, Physik der Erde, p. 176. Leipzig, 1911. Compare also Tams, “Die Entstehung des kalifornischen Erdbebens vom 18. April, 1906,” Peterm. Mitt., 64, p. 77, 1918.
11. Andrew C. Lawson, “The Mobility of the Coast Ranges of California,” Univ. of California, Publ., Geology, Vol. 12, No. 7, pp. 431–473, 1921.
12. Otto Meissner, “Isostasie und Küstentypus,” Peterm. Mitt., 64, p. 221, 1918.
13. Walther Penck distinguishes a third still heavier kind of magma, which he calls the Arctic magma, and the place of origin of which he places at a still greater depth (“Die Entstehung der Gebirge der Erde,” Deutsche Revue, 1921).