The Origin of Continents and Oceans/Chapter 3

by Alfred Wegener

II. DEMONSTRATION

CHAPTER III

GEOPHYSICAL ARGUMENTS

The statistics of the contours of the earth’s crust lead to the remarkable result that there are two heights which occur most frequently, whilst the stages between them are quite rare. The higher stage corresponds to the continental platforms, the lower to the oceanic floor. If the whole surface of the earth be divided into square kilometres and these are arranged in a series according to their height above sea-level, the well-known diagram of the so-called hypsometric curve of the earth’s surface is obtained, which clearly shows these two steps. According to the most recent calculations of H. Wagner,^[1] the frequency of the different levels is shown numerically in the following way:—

Depth below sea.									Elevation above sea.
Below	6	5–6	4–5	3–4	2–3	1–2	0–1		0–1	1–2	2–3	over 3 km.
	1.0	16.5	23.3	13.9	4.7	2.9	8.5		21.3	4.7	2.0	1.2 per cent.

This series is best illustrated by another method based on the somewhat older, but only in unessentials different, figures of Trabert,^[2] given in Fig. 5. It is applied to stages of 100 m., so that the percentages are naturally only about one-tenth of those given above. According to this, the maxima lie at a depth of about 4,700 m. and at an elevation of about 100 m.

These figures draw attention to the fact that the increase of the number of soundings always shows an increase in the steepness of the declivity from the margin of the continent or shelf. This is emphasized
Fig. 4.—Hypsometric curve of the earth's surface, after Krümmel. by any comparison of older ocean charts with the most recent one taken from Groll.^[3] For example, whilst Trabert gives, even in 1911, 4.0 per cent. for the stage 1 to 2 km., and for the 2 to 3 km. 6.5 per cent., we find that Wagner, whose figures are based ultimately on Groll’s chart, gives for the same distances only 2.9 and 4.7 per cent. respectively. Thus it is certainly to be expected that in the future both the frequency-maxima will be shown to be somewhat more sharply separated than the present observations indicate.

In the whole of geophysics there is scarcely another law of such clearness and certainty as this one, which states that there are two favoured levels on the earth, which occur alternately side by side and which are represented by the continents and the floors of the oceans. Therefore it is very remarkable that for this law, which has been well known for at least fifty years, no explanation has ever been sought. Only Sörgel,^[4] in his polemic against the displacement theory, attempts to trace it to elevation and subsidence. Even this attempt rests on an erroneous consideration.
Fig. 5.—The two frequency maxima of elevation.If, as he believes, only a single equilibrium level existed, disturbances thereof, such as elevations and subsidences, could then only give rise to two differing frequency-maxima, if physical causes existed for a preference of just these elevations. Since this is not the case, the frequency should simply be controlled by the Law of Errors of Gauss, the approximate course of which is drawn as an interrupted line in Fig. 5, because the deviations from the level of equilibrium must naturally be fewer as they become greater. Thus there should exist a single frequency-maximum somewhere in the region of the mean sphere level (−2,450 m.). Instead of this we see two maxima, both of which have a curve with a course similar to that of the law of errors. From this it must be concluded that there are already two undisturbed original levels, and from this the step seems inevitable, that in the continents and the floors of the oceans we have two different layers of the body of the earth, which—expressed in a somewhat exaggerated form—act as water does between great sheets of ice. This step seems so easy and obvious that the next generation will certainly wonder that we should have hesitated such a long time over taking it.
Fig. 6.—Diagrammatic section through the margin of a continent. In Fig. 6 a diagrammatic vertical section is given across the margin of a continent according to this new conception.

It is at once necessary to exercise caution against any exaggeration of this new conception of the nature of the oceanic floors. In our comparison with the tabular icebergs we must certainly also consider the fact that the upper surface of the sea between them can be again covered with newer ice, and further also, that smaller fragments of the iceberg which have become detached from its upper margin, or have risen from its foot, deeply submerged as it is under-water, could cover the surface of the water. In a similar manner this will occur at many places on the ocean floors. Islands are always larger pieces of continent, which, with their substructure, reach 50 to 70 km. under the floor of the ocean, as is shown by gravity measurements. They are comparable to the non-tabular icebergs.

Although this argument of the double frequency-maximum is quite sufficient to prove the correctness of the idea illustrated in Fig. 6, it nevertheless might be asked whether the other results of geophysics are compatible with it.

It is obvious that the gravity measurements in the oceans will be satisfied by our assumption in at least as good a manner as by the earlier one, according to which the outermost skin of rock is thinner here, but is, however, not quite absent. For they only signify that the rock beneath the oceans is heavier than that under the continents. We need not go any further into this question.

From investigation in earth magnetism, to which A. Nippoldt drew my attention, the view is generally held that the floors of the oceans consist of more strongly magnetic, therefore probably more ferruginous, material than the continental blocks. This is especially emphasized in the discussion over Henry Wilde’s^[5] magnetic model of the earth. Wilde covered the oceanic areas with sheet-iron in order to obtain a distribution of magnetic force corresponding to the earth’s magnetism. A. W. Rücker^[6] described this attempt in the words: “Mr. Wilde has produced a good magnetic model of the globe by means of an arrangement which consisted in effect of a primary field due to a uniformly magnetized sphere, and a secondary field due to iron, placed near the surface and magnetized by induction. The principal part of the iron is placed under the oceans. … Mr. Wilde attaches the greatest importance to the covering of the oceans with iron.” Raclot^[7] has also recently confirmed the opinion that this attempt of Wilde represents very well in broad outlines the picture of the distribution of the magnetism of the earth. Nevertheless, up to the present there has been no success in deriving, by calculation from the observation of the earth’s magnetism, the difference between continents and oceans, apparently because it is overlain by another, much greater, field of disturbance of yet unknown origin. The latter shows no relation to the distribution of the continents, and also cannot do so, as is evident from its great alterations expressed in the secular variation. But nevertheless, the results of the study of the earth’s magnetism do not at all vitiate, according to the view of such specialists as A. Schmidt, who will not yet acknowledge without reservations the convincing character of Wilde’s experiment, the assumption that the floor of the oceans consists of more ferruginous rock. Since it is generally assumed that the iron content increases with depth in the silicate mantle of the earth, and further, that the interior of the earth is mainly composed of iron, this signifies that we are dealing with a deeper layer. Now, as a rule, magnetism disappears at red heat, a temperature which would be reached, on the basis of the usual geothermal gradient,^[8] at a depth of about 15 to 20 km. The strong magnetism of the floor of the ocean must therefore occur in the uppermost layers, which appears to agree quite well with our assumption that there the weaker magnetic substances are quite absent.

Our assumption is also upheld by the study of earthquakes. E. Tams^[9] has compared the velocity of propagation of the surface seismic waves through continents with those through oceanic areas, and found the following values:—

1.—OCEANS.
					Number.
Californian	Earthquake,	April 18, 1906	v. =	3.847±0.045	km./sec.	09
Columbian	Earthquake,„	January 31, 1906		3.806±0.046	km./sec.„	18
Honduras	Earthquake,„	July 1, 1907		3.941±0.022	km./sec.„	20
Nicaraguan	Earthquake,„	December 30, 1907		3.916±0.029	km./sec.„	22
2.—CONTINENTS.
Californian	Earthquake,	April 18, 1906	v. =	3.770±0.104	km./sec.	05
Philippine	Islands, I „	April 18, 1907		3.765±0.045	km./sec.„	30
Philippine„	Islands„II „	April 18, 1907		3.768±0.054	km./sec.„	27
Buchara	Islands, I „	October 21, 1907		3.837±0.065	km./sec.„	19
Buchara„	Islands, I „	October 27, 1907		3.760±0.069	km./sec.„	11

Even if individual results do sometimes overlap, a considerable difference can on the average be recognized to the effect that the velocity of propagation through the oceans is about 0.1 km. per second greater than through the continents. This agrees with the expected theoretical values obtained from the physical properties of volcanic igneous rocks.

On the other hand, Tams also tried to combine the observations of as many earthquakes as possible, and thus obtained from the velocity values of 38 Pacific earthquakes an average value of 3.897 ± 0.028 km./sec., and from 45 Eurasian or American earthquakes 3.801 ± 0.029 km./sec., that is, practically the same values as above.

Angenheister^[10] also has recently examined the seismical difference between ocean basins and continental blocks by means of a series of Pacific earthquakes in which at the same time he attempted to differentiate from each other two kinds of surface waves not separated by Tams. He found, on scanty material however, considerably greater differences: “The velocity of the principal waves is about 21 to 26 per cent. greater under the Pacific than under the Asiatic continent. … The times of transit for P and S^[11] under the Pacific with 6° focal distance are about 13 seconds and 25 seconds respectively less than under the European continent. This corresponds to a greater velocity of about 18 per cent. for S under the ocean. … The damping of the principal waves is greater under the Pacific than under Asia. … The period in the concluding phase is greater under the Pacific than under Asia. …” All these differences point to the accuracy of our assumption that the ocean floor is composed of another, and that a heavier material. It is important to note that we are dealing here essentially with surface waves, so that these data become positive proofs of the complete absence in the ocean floor of the lighter outermost crust of rock.

It seems very natural to ask whether it would not be possible to obtain any specimens of this rock directly from the oceanic floor. It will be impossible for a long time yet to bring to the surface specimens of the “country” rock by the drag-net or other means. Nevertheless, the fact that, according to Krümmel,^[12] the greater portion of the specimens brought up by dredging is volcanic, deserves attention; “ predominant pumice … then fragments of sanidine, plagioclase, hornblende, magnetite, volcanic glass and its decomposition product palagonite, also pieces of lava, basalt, augite-andesite, etc., are found.” Volcanic rocks are in fact distinguished by greater specific gravity and greater iron content, and are generally considered as derived from greater depths. Suess called all this basic group of rocks, the chief member of which is basalt, “Sima,” from the initial letters of the principal constituents, silicon and magnesium, in contrast to the other more siliceous group of “Sal” (silicon and aluminium), the chief representatives of which, gneiss and granite, form the substructure of our continents.^[13] Following a short communication from Pfeffer, I should like to write “Sial” instead of “Sal,” in order that there may be no confusion with the Latin word for salt. From the preceding the reader will have already drawn for himself the conclusion that the rocks of the sima group, which we only know as eruptive rocks in the sialic continental blocks where they appear as foreign bodies, have their own place beneath these blocks, and form at the same time the floor of the ocean. Basalt has all the necessary properties for the material of the ocean floors. In particular its specific gravity is in harmony with the thickness of the continental blocks, calculated by other means.

It is not without profit to obtain a few more exact figures on the matter. The thickness of the continental blocks has been calculated by Hayford and Helmert by different methods. Hayford derived the so-called “depth of compensation level” (that of uniform pressure) at 114 km., which is identical with the underside of the continental blocks, from the deviations of the plumb-line at more than a hundred stations in the United States. Practically the same figure, namely 120 km., was found by Helmert from gravity measurements with the pendulum at fifty-one coastal stations. The close agreement of these figures, obtained as they are in different ways, naturally gives to them a heightened degree of accuracy, but it should not be thought that this thickness can be universally assigned to the continental blocks.^[14] This would not be consistent with isostasy.
Fig. 7.—Diagrammatic section through the margin of a continent. The thickness must be estimated as much less in the case of continental shelves, and as much greater^[15] in elevated regions such as Tibet, so that about 50 to 300 km. can be assumed as limits.

We can now easily calculate what the specific gravity of the sial and sima must be in order that a block of sial of about 100 km. thickness (M) may be elevated 4.8 km. above the ocean floor, thus being submerged to a total depth of 95.2 km. (compare Fig. 7). Equilibrium of pressure exists on the lower margin of the continental block; that is, a column of unit cross section extending from here to the surface must always weigh the same, whether it be taken in the continental area or in the oceanic. If we call the specific gravities of sial x and sima y, and consider that that of sea-water, which must also be taken into account, is 1.03, the equation will therefore become

100 x	=	95.2y + 4.7 × 1.03,
orx	=	00.952y + 0.048

Now, since sima rocks like basalt, diabase, melaphyre, gabbro, peridotite, andesite, porphyrite, diorite, among others, have mostly a specific gravity of about 3.0 (only rarely up to 3.3), we can put y = 3.0 and then obtain x = 2.9. Whitman Cross and Gilbert actually obtained the specific gravity of 2.615 for gneiss as an average from twelve specimens. Other observations gave values between 2.5 and 2.47. This small difference, however, can easily be explained, in that the specific gravity in the sial zone as well as in the sima increases with depth, and that basalt is derived from great depths, whilst the specimens of gneiss came from close to the surface. To be sure, this is not mathematically proved, since we do not know the extent of the increase of specific gravity with depth. We only know that, according to research on earthquakes, the average value of 3.4 is yielded for the whole of the approximately 1500 km. thick silicate mantle of the earth. These figures agree, in any case, qualitatively with our assumptions.^[16]

Finally, in this connection the smoothness of the ocean floor must be mentioned, because it also confirms the accuracy of our ideas. It has been long known that the ocean floor often shows astonishingly slight differences in level over wide areas, a circumstance not without practical importance for the laying of cables. For example, out of the 100 soundings carried out for the cable between Midway Island and Guam along a length of 1540 km., the extreme values (5510 and 6277) only differed by 767 m. On a portion 10 geographical miles long, where an average of 14 soundings gave 5938 m., the greatest deviations were + 36 and − 38 m.^[17]

It is true that the principle of the smoothness of the ocean floor has recently become somewhat qualified, for it is seen that the network of soundings is too widely meshed to permit of such conclusions, and that an erroneous impression of great levelness can be obtained on land by similarly dispersed separate measurements of altitude. But, with Krümmel, most workers have returned from the temporary exaggerated scepticism to the view that, leaving aside the oceanic troughs, a fundamental difference exists between land and deep sea, although on account of the loss of gravity under water the slopes could be much steeper there than in the air. In this greater smoothness is manifested a greater plasticity, a higher degree of mobility of the floors of the oceans.

Another deduction from this levelness is the absence of folded chains at the bottom of the sea. Whilst the continental blocks are wrinkled in all directions by ancient and recent folded mountain chains, we do not know, in spite of all the soundings that have been made, a single feature from the enormous area of deep sea which we could claim with any certainty as a chain of mountains. Some, of course, would conceive the elevations of the floor of the Middle Atlantic, and also the ridge between the two troughs lying in front of Java, as corresponding to folded mountains; nevertheless, this view counts so few adherents that we can content ourselves here with a reference to Andrée’s criticism.^[18] How can this absence be explained, since compression must be also assumed in the sima? The answer is obvious if we consider isostasy in relation to mountain building. Mountain building is folding subject to the preservation of isostasy. Since by far the greatest portion of the continental block, a thickness of 100 km., is submerged in the sima, the greatest portion of the thickening of the block by folding must be in a downward direction. Only a very small portion of the compression will be visible as elevation, a subject to which we will return later in Chapter XI. But while the greatest part of the compression of the continental blocks is downward, a compression in the sima cannot in any case lead to an elevation. The material in this case is squeezed out below or to the side, just as water between two approaching icebergs. The objection of A. Penck,^[19] that “the absence of sima folding on the anterior side of the drifting continents appears to be a decisive proof against his conception of the constitution of the crust and the mobility of the continents,” is not valid. On the contrary, we see a confirmation of our assumption regarding the nature of the sima in the absence of such uprising folded ranges in the ocean floor. These masses are certainly folded. If they had consisted of sial, the folding would, at least partly, have been directed upwards and come into view. These circumstances will become still clearer by the considerations mentioned in Chapter IX.

The proofs given in this chapter of the nature of the floor of the ocean speak in a very clear and impressive language. Hence this aspect of our theory has up to the present encountered least opposition, and these ideas have been accepted by most geophysicists.

1. H. Wagner, Lehrb. d. Geographie, vol. i., Allgemeine Erdkunde, Part 2: Physikal, p. 271. Hanover, 1922. The new survey of the oceans by Kossina (Die Tiefe des Weltmeeres, Veröff. d. Inst. f. Meereskunde, N.F.A., Part 9. Berlin, 1921) is taken into consideration in his figures. Our diagrams are taken, however, from the older but little different values of Krümmel and Trabert.
2. Trabert, Lehrb. d. kosmischen Physik, p. 277. Leipzig and Berlin, 1911.
3. Groll, Tiefenkarien der Ozeane, Veröff. d. Inst. f. Meereskunde, Part 2, Berlin, 1912.
4. W. Sörgel, “Die Atlantische ‘Spalte,’ Kritische Bemerkungen zu A. Wegener’s Theorie von der Kontinentalverschiebung.” Monatsber. d. deutsch. Geol. Ges., 68, pp. 200–239, 1916.
5. Proc. Roy. Soc., June 19, 1890, and January 22, 1891.
6. A. W. Rücker, “The Secondary Magnetic Field of the Earth,” Terrestrial Magnetism and Atmospheric Electricity, 4, pp. 113–129, March–December, 1899.
7. C.R., 164, p. 150, 1917.
8. According to J. Friedlaender, Beitr. z. Geophys., 11, Kl. Mitt., pp. 85–94, 1912, the thermal conductivity of the deep igneous rocks which we are considering is even smaller (geothermal gradient of lava 17 m.), so that the thickness of the magnetic layer would only amount from 8 to 9 km.
9. E. Tams, “Über die Fortpflanzungsgeschwindigkeit der seismischen Oberflächenwellen längs kontinentaler und ozeanischer Wege,” Centralbl. f. Min. Geol. u. Paläont., 1921, pp. 44–52 and 75–83, 1921.
10. G. Angenheister, “Beobachtungen an pazifischen Beben,” Nachr. d. Kg. Ges. d. Wiss. z. Göttingen, Math.-Phys. Klasse, 1921, pp. 113–146. The work of Omori mentioned in the previous edition, who found such greater differences, rests on a misconception of that author with regard to the nature of the waves and is to be deleted.
11. P (undae primae) and S (undae secundae) are the designations for the first and second preliminary tremors of the seismogram, which can be traced respectively to longitudinal and transverse elastic vibrations propagated through the interior of the earth.
12. O. Krümmel, Handb. d. Ozeanographie, 1, pp. 193 and 197. Stuttgart, 1907.
13. This division goes back as far as Robert Bunsen, who separated the non-sedimentary rocks into “normal trachytic” (siliceous) and “normal pyroxenic” (basic). Suess, however, devised the more convenient names.
14. These calculations are founded on Pratt’s hypothesis. Schweydar, in a preliminary communication states, on the basis of the hypothesis of Airy, that the thickness of the block is found to be 200 km., which corresponds to a difference of specific gravity of sial and sima of only 0.034 (“Bemerkungen zu Wegeners Hypothese der Verschiebung der Kontinente,” Zeitschr. d. Ges. f. Erdk. zu Berlin, 1921, p. 121).
15. Hayden calculates the depth of compensation of the Himalayas at 330 km., of the lowland at 114 km.; the calculation is, however, not quite free from objection.
16. The following small table may still further explain the dependence of the submergence of the sial on its specific gravity. It gives the thickness of the blocks for a sima of specific gravity of 3.0. If this be reduced simultaneously with that of the sial by 0.1, then the figures of the table are only decreased about 5 per cent.

    Thickness of block of sial in sima of specific gravity 3.0.

    Specific Gravity of sial. 2.6 2.7 2.8 2.9 2.95 Height of surface above sea-level 0100 m. 24 32 048 096 192 km. 4000 m. 53 71 106 213 430 km.„
17. O. Krümmel, Handb. d. Ozeanographie, 1, p. 91. Stuttgart, 1907.
18. K. Andrée, Über die Bedingungen der Gebirgsbildung, p. 86, etc. Berlin, 1914.
19. A. Penck, “Wegeners Hypothese der kontinentalen Verschiebungen,” Zeitschr. d. Ges. f. Erdk. zu Berlin, pp. 110–120, 1921.