REPORT OF THE CHIEF ASTRONOMER 703 



SESSIONAL PAPER No. 25a 



dolomites among the basic rocks; with argillaceous rocks as the residuum of the 

 decomposition.'* 



The idea that basalt, as a ' comprehensive or synthetic ' rock, might have 

 given the -world's granites through weather-leaching and remelting of the leached- 

 out products can be tested quantitatively with a fair degree of confidence in the 

 result. The ratio of soda to potash, in the average basalt, is about 3-16: 1-55. 

 The ratio in average granite is about 3-31: 4>10; and in the average pre- 

 Cambrian granite, about 3 26: 4 -53. (See Columns 1 and 4 of Table XLIV). 

 Even if all the potash remained among the residual products of the weathering 

 of basalt, it would take nearly three weight units of basalt to make a weight unit 

 of granite, according to Dutton's principle. All the soda, or say three per cent of 

 two weight units of basalt, goes in solution to the sea. Many granite batholiths 

 are known to be at least two miles deep and are probably much deeper. It is safe 

 to postulate that 40,000,000 square miles of the earth's surface is underlain by 

 granite or by the average pre-Cambrian terrane, itself a granite in composition. 

 If we hold that their combined mass is of the minimum average depth of two miles, 

 it follows that at least 200,000,000 cubic miles of basalt, or enough to cover the 

 planet one mile deep, must have been weathered to produce the whole granitic 

 mass. About two per cent, by weight, of the basalt is sodium carried in solution 

 to the ocean. This addition alone would charge the ocean with three times as 

 much sodium as it actually contains. Since nearly all the sodium which has ever 

 entered the ocean is still there in solution, the vast excess calculated shows, 

 without allowing for other sources of oceanic sodium, that the main assumption 

 cannot be true. 



The only remaining interpretation of the acid basement complex of the con- 

 tinental plateaus recognizes in it the material of the earth's primary surface shell. 

 This shell has been denuded, metamorphosed, and largely remelted in the huge 

 batholithic invasions of the Laurentian type. Little or none of the visible pre- 

 Cambrian terrane directly represents the unmodified, original crust, but, in spite 

 of all its vicissitudes, the terrane seems to have retained the average chemical 

 composition of the primary acid shell. 



Our speculation leads, thus, to the conception of an early separation of the 

 earth's outer magmatic layer into two shells ; the underlying one basaltic in com- 

 position, the overlying one granitic in composition. If these are the poles of a 

 gigantic process of magmatic differentiation, the original magma must have 

 been of some mediosilicic type. If it be arbitrarily assumed that the two poles 

 of this differentiation were formed in equal masses, the original magma must 

 have had a composition much like the average auf ite andesite or the average 

 diorite. The following table (XLV) shows the comparison of the mean of the 

 average pre-Cambrian granite and average basalt, with the average diorite and 

 average augite andesite; each average being computed from a large number of 

 analyses, and recalculated as water-free. 



* C. E. Dufcton, Report on the Geology of the High Plateaus of Utah, Washington, 

 ' 1880, pp. 124-125. 



