ORIGIN OF CRYSTALLINE ROCKS. 499 
While, however, there is good reason to believe that solutions of 
alkaline silicates or carbonates have been efficient agents in the 
crystallization and molecular re-arrangement of ancient sediments, 
and have also played an important part in that local alteration of 
sedimentary strata which is often observed in the vicinity of intru- 
sive rocks, it is clear to me that the agency of these solutions is 
less universal than once supposed by Daubrée and myself, and will 
not account for the formation of various silicated rocks found 
among crystalline schists, such as serpentine, hornblende, steatite 
and chlorite. When I commenced the: study of these crystalline 
strata I was led, in accordance with the almost universally received 
opinion of geologists, to regard them as resulting from a subse- 
quent alteration of paleozoic sediments, which, according to differ- 
ent authorities, were of Cambrian, Silurian or Devonian age. Thus 
in the Appalachian region, as we have already seen, they have, on 
supposed stratigraphical evidence, been successively placed at the 
base, at the summit, and in the middle of the Lower Silurian or 
Champlain division of the New York system. A careful chemical 
examination among the unaltered paleozoic sediments, which in 
Canada were looked upon as the stratigraphical equivalents of the 
bands of magnesian silicates in these crystalline schists, showed 
me, however, no magnesian rocks except certain silicious and 
ferruginous dolomites. From a consideration of reactions which 
I had observed to take place in such admixtures in presence of 
heated alkaline solutions, and from the composition of the basic 
silicates which I had found to be formed in silicious limestones 
near their contact with eruptive rocks, I was led to suppose that 
similar actions, on a grand scale, might transform these silicious 
dolomites of the unaltered strata into crystalline magnesian sili- 
arther researches, however, convinced me that this view was 
inapplicable to the crystalline schists of the Appalachians, since, 
apart from the geognostical considerations set forth in the previous 
part of this paper, I found that these same crystalline strata hold 
beds of quartzose dolomite and magnesian carbonate, associated in 
such intimate relations with beds of serpentine, diallage and stea- 
tite, as to forbid the notion that these silicates could have been 
generated by any transformations or chemical re-arrangement of 
mixtures like the accompanying beds of quartzose magnesian car- 
bonates. Hence it was that go in 1860, as shown above, I 
AMER. NATURALIST, VOL. V 
