unaccounted for. Part of it goes to form pyrite, and part is decomposed by organic agencies and lost, but the proportion of loss is unknown. It is, doubtless, large. The potassium which is taken up by clays or else in glauconite is in either instance represented as silicate, and hence a part of the silica is regarded as in combination. The sesquioxides are calculated as limonite, although a part of them is certainly alumina; but no refinement of a calculation here would change the order of magnitude as given. The several orders of magnitude are probably close to the truth, and we may say with much confidence that the precipitates, including such substances as coral, shell, diatomaceous ooze and what not are formed at a rate of something like 21 X 108 metric tons a year, plus a small but undefined allowance for that part of the sulphur which has been fixed as pyrite.

At the figure given, chemical sediments are now forming in the ocean sufficient to cover 88,000,000 square miles of the sea floor to the depth of 0.0001337 inch annually. The whole area of the ocean is 139,440,000 square miles, but the portion covered by the red clay, where the precipitation is relatively insignificant, must be deducted. If the rate had been uniform throughout geological time, 83,472,000 years these sediments would form a layer about 930 feet deep, but such a calculation is unsound. Large areas of what were once marine sediments are now land, and, moreover, neither the rate nor the distribution of the deposits can have been uniform. The limestones that are forming now are largely derived from the solution of older deposits, Cambrian, Silurian, Devonian, Cretaceous, etc., and their carbonates have been deposited in the ocean, not once only, but possibly several times. In the earliest geologic eras, when sediments began to form, the proportion of carbonates to other salts thrown down must have been much smaller than today. An average thickness of 930 feet over the assumed area is therefore a great exaggeration; and needs to be reduced.

It is probably impossible to determine, with any approach to precision, the actual quantity of marine sediments that have been formed. We can, however, make a plausible estimate, which shall, at least, give us some conception of their order of magnitude. It has

already been shown that the limestones, which are mostly of marine origin, have a volume of 3,916,400 cubic miles. With a specific gravity of 2.7 their mass becomes 42,092 X 1012 metric tons. From the figures given on p. 230 ante, the calcareous and magnesian sediments are now forming at a rate bearing a certain ratio to that of the other deposits, the limonitic and siliceous residues. This ratio, which is roundly 1,650:452, if constant throughout geologic time, would give for the latter class of sediments, proportional to the limestones, a mass of 11,664 X 1012 tons; the sum of both classes of precipitates being 53,756 X 1012 tons. The corresponding average thickness over the sedimentary oceanic area would then be 287 feet, or less than one third of the figure previously given. The actual thickness, however, must be much less; for a large part of the once marine sediments are now elevated into land. According to the best estimates, the land area of the globe is now covered by 23 per cent. of archæan and eruptive rocks, and 77 per cent. of sedimentaries. Adding this sedimentary area to that of the ocean, the total becomes 132,180,000 square miles, and the average thickness of the chemical sediments reduces to 191 feet. At the crude value assigned to geologic time this represents a rate of deposition of only 0.000027 inch annually. If the age of the earth is less than 83,472,000 years, the mean annual rate of deposition will be proportionately increased, but not to anything like the present magnitude.

15

Whether the ratio assumed between the calcareous and siliceous sediments is justifiable or not, is a question admitting of argument. It seems, however, probable, that in the earliest geologic ages, when the land area was occupied principally by igneous rocks, the salinity of the rivers was relatively low, but the proportion of silica to lime in the waters was higher. This suspicion is justified by a study of the river waters of today, especially those issuing from granitoid areas. In such waters silica is often in excess of lime, while in waters from sedimentary areas the reverse is commonly true. The ratio here assumed represents a balancing between waters of both classes, and is therefore as legitimate as any other which might be chosen. Here it must be borne in mind that we are dealing with probabilities only, nothing more.

15 Von Tillo as modified by Becker. See Becker's memoir already cited.

So far, the mechanical sediments, such as silt and sand, have not been considered. From the surface of the United States, according to Dole and Stabler,16 the rivers annually carry to the sea 270,000,000 tons of dissolved substances, and 513,000,000 tons in suspension. If this ratio, which is only approximate, should hold for the whole world, the quantity deposited in the ocean during geologic time would be 102,370 X 1012 tons, and the total sedimentation, chemical and mechanical, becomes 156,126 X 1012 tons. This quantity, distributed over the entire sedimentary area, continental and oceanic, gives an average thickness of about 550 feet, or 0.000079 inch a year.

The total volume of the marine sediments thus computed, is 13,873,000 cubic miles. The volume remaining in the ocean is very nearly two thirds of this figure, 9,239,000 cubic miles. The volume of all the secondary rocks derived from the decomposition of igneous rocks was previously found to be 78,338,000 cubic miles. Hence the portion now on the land area of the globe amounts to 69,099,000 cubic miles of rock, consisting in great part of materials which were never transported very far from their original place of formation.

To the foregoing estimates of the oceanic sediments at least one large but undetermined correction needs to be applied. The ocean receives great quantities of dust, representative of aerial erosion, and also quantities of volcanic ejectamenta. For these nonfluviatile additions no valid estimates can yet be made. The major portion of them, however, must come from disintegrated sedimentary rocks, sands, and soils, and so do not affect to any serious extent our estimates of rock decomposition. The oceanic share of the sediments should be increased, but less than appears at a first glance. The marine sediments now on land must include a part of the contributions made to the ocean by atmospheric transportation. The actual distribution of the sediments is naturally very uneven. They are probably thin near the margin of the red clay, and thick along the continental shelves. Coral rock, for example, has been bored to a depth of 1,100 feet without reaching its limit. The mechanical sediments are of course mainly deposited relatively near to shore.

10 U. S. Geol. Surv. Water Supply Paper, No. 234, p. 83.

In order to prevent misapprehension it may be well to reiterate the statement that these estimates of the marine sediments are necessarily crude, and represent orders of magnitude only. With better evidence, better estimates may at some future time be made, but accuracy in them is unattainable. The other figures given, for the composition of the igneous rocks and of the ocean are probably near the truth but are still subject to revision and improvement.

THERMAL RELATIONS OF SOLUTIONS.

BY WILLIAM FRANCIS MAGIE.

(Read April 20, 1912.)

The thermal relations of solutions afford evidence of a peculiar and valuable kind about the nature of solutions. The electrical conductivity of solutions has been explained by the hypothesis that the molecules of the solute are partly dissociated into ions in the solution. When this hypothesis is further tested on the assumption that the osmotic pressure is proportional to the number of free molecules and ions in the solution, the experimental results of Dieterici, of Kahlenberg, of Jones and others show disaccord with the predictions of the hypothesis. I venture to believe that a study of the heat capacities and the heats of dilution of solutions will confirm the view that the reason for this disaccord lies in the assumed relation of the osmotic pressure to the dissociation, so that while the dissociation hypothesis is confirmed, the relation of the osmotic pressure to the dissociation is shown to be different from that which was originally assumed.

Five years ago I presented to this Society a paper in which I discussed the heat capacities of solutions. I will summarize here the principal results described in that paper, in order to render the present discussion more complete.

The heat capacity of a gram-molecular solution of an electrolyte diminishes with increasing dilution. The change in the heat capacity is directly proportional to the change in the dissociation, determined from the electrical conductivity. In most cases, at ordinary concentrations, the heat capacity of the water is diminished, and we are led to infer an interaction between the water and the molecules and ions of the solute of such a sort as to diminish the freedom of the water. A study of the constants of the formula by which the heat capacity is expressed leads to the conclusion that the heat capacity of the