1902-3.] The Pal^ochemistry of the Ocean. 549 



as they seem to be, if one scrutinizes the proportions that are 

 found in rivers whose waters have been carefully analysed. There are 

 only two rivers, the Amazon and the St. Lawrence, which give nearly 

 the proportion of magnesium called for by Murray's estimates, while the 

 Ottawa, the Mississippi, and the Nile give quantities much below that 

 of the sodium, and the quantities of the calcium are found to vary very 

 much for the different rivers. If we disregard Murray's estimates and 

 base our observations on the analyses of the various rivers, we can 

 safely conclude that, while the quantity of calcium added, except in the 

 case of the Nile, is always, and sometimes very much, greater than the 

 sodium addition, the latter does not probably exceed the amount of the 

 magnesium discharged. In the ocean, however, the sodium, calcium, 

 and magnesium have the proportions of 100, 3.91 and 12.0. 



The comparatively low proportion of magnesium in sea water is 

 explainable. In the first place, as pointed out above, there must have 

 been in the primeval ocean but very little magnesium, owing to the 

 conversion of all the chloride of magnesium into magnesia which is, 

 except in minute quantities, insoluble. The conditions which so affected 

 magnesium chloride left the chlorides of calcium, sodium and potassium 

 unchanged, and in conseqence these went into solution in primeval sea 

 water, and were, therefore, as compared with magnesium, very abundant. 

 Further, the ocean at first must have contained only traces of the latter 

 element and the subsequent addition of it through river discharge 

 would increase the amount in sea water, but not to such an extent as 

 to make it overtake the sodium. 



There is another factor which operated in limiting the amount of 

 the magnesium. This is the tendency shown by the chloride to interact 

 with the carbonate of lime when the latter undergoes deposition to form 

 limestone, and, in consequence, this always contains carbonate of 

 magnesia. When the latter exceeds 10 per cent, the mixture of the 

 carbonates is given the conventional name of dolomite, and in some 

 formations of this kind the magnesia is found greatly to exceed the lime. 

 Dolomites are found in all the periods down to and including the 

 Cambrian and even in the pre-Cambrian, it is associated with the 

 crystalline schists.* An exact estimation of the magnesium so 

 localized is impossible, but on the average it cannot be more than 10 

 per cent, of the quantity of the calcium due to deposition, so that the 

 amount of magnesium removed annually from sea water must fall far 

 behind that of the calcium. It follows from this that whatever were 



' Zirkel, Lehrbuch der Petrographie, 1894, Bd, 4, p. 499. 



