THE AGE OF THE EARTH.1

By J. JOLY, F.R.S.

The recent contributions to the data bearing on the subject of the age of the earth have strengthened the evidence derived by two very different methods of computation; that based on the study of solvent denudation and that based on the accumulation of radioactive waste products in minerals. While the indications of both lines of inquiry seem individually rendered more definite by these advances, the divergence in their final results have, if anything, become intensified. I propose in the following pages to review the opposing methods, as briefly as the many details permit, and to discuss the possibility of reconciliation.

THE AGE OF THE OCEAN DERIVED FROM SOLVENT DENUDATION.

Three recent contributions to this subject have appeared: Prof. Sollas's Presidential Address to the Geological Society of London, 1909; a paper on "A preliminary study of chemical denudation," by F. W. Clarke (Smithsonian Miscellaneous Collections, vol. 56, June, 1910); and a paper by G. F. Becker on "The age of the earth" (Smith sonian Miscellaneous Collections, vol. 56, June, 1910).

These recent discussions chiefly center round the ascertainment of the true present rate of supply of sodium to the ocean. The limitations of the method are also discussed.

My own original estimate of the age of the ocean was based on the only data then available the estimates made by Sir John Murray of the average chemical composition of river water and the probable · total annual discharge of the rivers into the ocean. Calculating from its estimated volume and mean chemical composition the mass of sodium now in the ocean, and dividing this by the calculated amount of sodium entering annually from the rivers, the uncorrected age of 99.4 million years was obtained. To this I applied certain corrections, to some of which I shall refer later. The final result of these corrections left the age as from 80 to 90 million years.

1 Reprinted by permission, after revision by the author, from the Philosophical Magazine, London, 8. 6, vol. 22, No. 122, September, 1911, pp. 358-370.

2 Trans. Roy. Dublin Soc., vol. 7, 1899.

Prof. Sollas approaches the question by a recalculation of the average amount of sodium discharged by the rivers annually. He finds that the added results available, as derived from the rivers of North and South America and Europe, give the uncorrected age as 78 million years. After a careful and detailed discussion of the corrections, Sollas concludes that the age lies between 80 and 150 million years; the latter figure being based on extreme assumptions.

Clarke bases his discussion of the question upon what he terms the denudation factor, i. e., the number of metric tons annually removed in solution from a square mile of drainage area. This is estimated for a number of important rivers of the world, accounting in this way for a drainage area of 28 millions of square miles out of the total of about 40 millions which drain to the ocean. The mean value found for the denudation factor is 68.4 tons. Assuming that this denudation factor is a fair average for the whole, the entire matter in solution discharged into the ocean in a year is 2,735 millions of tons. From the chemical analyses of this saline matter for the several rivers, an average composition for each continent is found. When this is weighted for the quantity of water contributed by each continent, a final weighted mean composition is obtained which may be applied to determining the integral of the sodium passing annually from rivers to ocean. this way it is found that 175,040,000 metric tons of sodium are annually discharged into the sea. Clarke next finds the total amount of sodium in the ocean to be 14,130 X 1012 tons. My own results were based on a slightly higher value-15,611×1012 tons. From his figures, Clarke now gets the uncorrected age as 80,726,000 years.

In

Although the numerous analyses which go to build up this result are not of equal value, there are certain satisfactory features in the computation.

It is explained by Clarke that in the wonderfully detailed analyses of the Mississippi by Dole and Stabler, taken along with their work on other great rivers of North America and with the observations of Forbes and Skinner for Colorado, data have been obtained for the United States which are not likely to be much altered by any future analyses. Twenty-two river basins enter into the mean for the United States, giving a mean denudation factor of 79 tons. For the rest of North America an estimate only is possible; but, for reasons given, Clarke concludes that "if we assume that 6 millions of square miles of North America lose 79 metric tons in solution per square mile per annum, and that the composition of the saline matter so transported is that found for the United States alone, we shall not be very far from the truth." Possessing thus a standard based on the drainage of a great continent, we feel confidence in our criticism of other data. The quantity of water thus dealt with is rather more

than one-fourth of that supplied by the entire drainage areas of the earth.

It will be seen from the tables given by Clarke that the mean denudation factor of 68.4 tons is in good agreement with the standard result from North America, nor is it very largely departed from by the factors derived from other continents.

There can, I think, be little doubt that the results arrived at by Clarke and Sollas are not likely to be seriously disturbed in the future. It is most improbable that they require amendment to the extent of 50 per cent. This being so, we conclude that the uncorrected estimates of the age of the ocean as based on solvent denudation is of the order of 100 million years. It remains now to consider the legitimate corrections to be applied to this figure.

At the present moment the most important aspect of this method of evaluating the age of the ocean is involved in its degree of reliability as affording a maximum value of the time elapsed since solvent denudation began. This point I shall therefore specially consider.

The errors affecting the crude result found by dividing the sodium of the ocean by the annual river supply, and tending to make this estimate too small, are:

(a) Underestimation of the sodium now in the ocean.

(b) Neglect of sodium which at some period in the past may have been in the ocean, but is now removed from it.

(c) Overestimation of the legitimate river supply of sodium.

(d) Decreased river supply of sodium in the past.

Of these possible sources of crror (a) may be at once dismissed. The average depth and area of the ocean and its average chemical composition are sufficiently well known to preclude the possibility of any serious error.

In considering (b) it is necessary to bear in mind the magnitude of the quantities involved. The saline matter in the ocean would represent a volume of over 4,800,000 cubic miles on Clarke's estimation. I have formerly pointed out that the rock salt alone would suffice to cover the land area of the globe to a depth of 122 meters. In comparison with quantities so vast all the salt deposits known sink into insignificance; nor is it likely that deposits adequate to enter into consideration exist.

The errors referred to in (c) must be of the nature of cyclic sodiumthat is, sodium which circulates from the sea to the land and back through the rivers to the ocean. Cyclic sodium exists in the form of wind-borne spray, which, descending on the land with the rainfall, augments that which is truly derived by denudation. In arid regions it may settle as dust, to be, under special circumstances, washed ultimately into the sea. Again, the sodium which the rivers

38734°-SM 1911-18

may derive from the ancient salt deposits which have been impounded from the sea is cyclic.

The influence of wind-borne sodium has been fully discussed by Sollas, Clarke, and Becker. There can be no doubt that it is relatively unimportant. My own original correction was 10 per cent of the river supply. Becker, by examining typical cross sections of the isochlors, determined for the rainfall of western North America by the United States Geological Survey, finds that an allowance of 6 per cent is sufficient. Sollas shows that these isochlors indicate that but a small fraction of the sodium chloride of the American rivers can be referred to this source. Clarke, by a somewhat different line of attack, concludes that a correction of 7 per cent on the sodium conveyed by the rivers of the United States is a maximum allowance. Clarke further considers that a correction for sodium chloride carried as dry dust is unnecessary.

In a paper contributed by me to the Geological Magazine (May, 1900) I considered the possibility of oceanic sodium existing disseminated in the sedimentary rocks. Such sodium would be of course cyclic. It was easy to show that, even on excessive estimates of the occluded sodium chloride in such rocks, taken in conjunction with their rate of removal by denudation, this source of supply to the rivers is less than 1 per cent. Clarke reconsiders the question and finds the allowance would not be more than 1 per cent. Three per cent is regarded by Clarke as a maximum deduction for sodium artificially supplied in modern times to the rivers.

Oceanic salt deposits are not very abundant over the surface of the earth, being generally confined to particular formations. That they seriously affect the river analyses of all the great rivers of the world is in the highest degree improbable. In any case if we deduct all the chlorinated sodium from the river supply we must include also all sea-derived sodium. If we effect this calculation, we obtain an age of about 150 million years. I do not think it will be disputed that this figure is in its nature excessive.

There remains the possibility (d) that the assumed uniformity of past and present conditions is illusory; in other words, that special conditions now exist tending to bring about an abnormally great river supply of sodium.

The present is admittedly a period of large land exposure. This, however, involves a fact which must be held in mind. At the present time the land area actually draining into the ocean is about 39.7 millions of square miles. The total land area is, however, rather over 55 millions of square miles. It follows that about 30 per cent of the land area contributes nothing to the ocean. Or, again, the areas which are classed as "rainless"—that is, which have less than an annual rainfall of 10 inches and have no run-off-are estimated as

one-fifth of the whole. Under such circumstances transgression of the ocean upon the land simply results in the diminution or disappearance of the great continental desert regions. It has been shown by Murray that it would require a vertical depression relatively to the ocean of 600 feet in order to reduce the existing land area by 26.7 per cent. Penck, on the same data, concludes that a submergence of 200 meters would reduce the area 29 per cent. A submergence of nearly 1,500 feet is required to diminish the land area 50 per cent. It is for geologists to judge whether world-wide transgressions of these magnitudes obtained for any long periods in the past. So far as I know, paleography would not support such transgressions. A recent study of the Paleography of North America by C. Schuchert 1 leads to the conclusion that the mean area of that continent throughout the past has been about eight-tenths of its present area. In his Traitê de Géologie, De Lapparent, in a series of well-known restorations of ancient geography, shows how far, as judged by the sediments, there was transgression of the sea upon the land at various epochs. It does not appear that we can infer, even at the climax of the great Cenomanian transgression, that the existing land was at any time covered to one-half its extent. And mindful of the fact that the area of denudation is in most cases much greater than that of deposition, when the latter is greatest the necessity of accounting for the former involves the assumption that tracts of land now submerged were then exposed. Without assuming the former existence of lost continents in the central oceanic basins, there seems very strong evidence for the disappearance of former land. The evidence is found in our own islands, in North America, in India, South Africa, and Australia and elsewhere. We have to recognize continual fluctuations, but the evidence for a prevailing reduction of continental areas by as much as 50 per cent, or even 25 per cent, in the past is, so far as I know, not forthcoming. We might go further and state that so great a diminution of existing land area as 50 per cent certainly did not prevail in the past. Such a reduction involves about 25 per cent of the present rate of solvent denudation and increases the age accordingly.

Meteorological conditions, unless occasioned by a prevailing change in the amount of solar heat, can not be supposed to have steadily affected in one direction the rate of denudation. It is worthy of note that the testimony derived from the solvent denudation of the continents shows that climatal conditions do not, within the limits, seriously affect the rate of solvent denudation. This finds explanation in the extremely complex nature of the factors concerned in rock weathering and rock solution. Now, the mere abundance of

1 Bull. Geol. Soc. America, vol. 20, 1910.