ORIGIN OF CRYSTALLINE ROCKS. 507 
well known, besides the bicarbonates of soda, lime, and magne- 
. sia, compounds of iron, manganese, and many of the rarer metals 
in solution, and thus the metalliferous character of many of the 
dolomites of the second class is explained. The simultaneous 
occurrence of alkaline silicates in such mineral waters, would » 
give rise, as already pointed out, to the production of insoluble 
silicates of magnesia, and thus the frequent association of such 
silicates with dolomites and magnesian carbonates in the crystal- 
line schists is explained, as marking portions of one continuous 
process. The formation of these mineral waters depends upon 
the decomposition of feldspathic rocks by subterranean or su 
aërial processes, which were doubtless more active in former ages 
than in our own. The subsequent action upon magnesian waters 
of these bicarbonated solutions, whether alkaline or not, is de- 
pendent upon climatic conditions, since, in a region where the rain- 
fall is abundant, such waters would find their way down the river- 
courses to the open sea, where the excess of dissolved sulphate of 
lime would prevent the deposition of magnesian carbonate. It is 
in dry and desert regions, with limited lake-basins, that we must 
seek for the production of magnesian carbonates, and I have ar- 
gued from these considerations that much of northeastern Amer- 
ica, including the present basins of the Upper Mississippi and St. 
Lawrence, must, during long intervals in the paleozoic period, 
have had a climate of excessive dryness, and a surface marked by 
shallow enclosed basins, as is shown by the widely spread magne- 
sian limestones, and the existence of gypsum and rock-salt at 
more than one geological horizon within that area.* The occur- 
rence of. serpentine and diallage at Syracuse, New York, offers a 
curious example of the local development of crystalline magne- 
sian silicates in Upper Silurian dolomitic strata under conditions 
which are imperfectly known, and, in the present state of the 
locality, cannot be studied. f 
Since the uncombined and hydrated magnesian mono-carbonate 
is at once decomposed by sulphate or chlorid of calcium, it fol- 
lows that the whole of these lime-salts in-a sea-basin must be 
converted into carbonates before the production of carbonated 
magnesian sediments can begin. The carbonate of lime, formed 
Tes on eee 
* Geology of Southwestern Ontario, Amer. Jour. Sci II, xlvi, 355. 
t Geology of the 3d district of New York, 108-110, and Hunt on Ophiolites, Amer. Jour. 
Sci., I, xxvi, 236 
