120 Mr. Dana on the Analogies between the 
The vaporization or exudation of magnesia from porphyries, 
as Von Buch supposes, in order to produce magnesian limestones, 
is a less satisfactory hypothesis than to suppose this earth intro- 
duced through heated waters containing magnesia in solution. 
Magnesia is one of the elements of sea-water, and when heated 
the water may have contained or received a much larger than the 
usual supply. But this and other theories are to a great extent 
if not entirely set aside by the discovery that recent coral rock 
in the Pacific often contains a large amount of magnesia. I sus- 
pected this fact when among the islands, both from their hard- 
ness and specific gravity ; and having put some specimens into 
the hands of Mr. B. Silliman, Jr. for analysis, he has obtained the 
very interesting result, that carbonate of magnesia occurs in large 
_ proportions in some of these rocks. 'These analyses will be car- 
ried on with the different varieties of coral, and the conclusions 
which must necessarily be important, will appear in the Expedi- 
tion publications. ‘These facts will account for the occurrence 
of magnesia in limestones not crystalline, which is wholly unex- 
plained by any theory of dolomization heretofore proposed. | 
But the coral rock examined and most compact magnesian 
limestones do not generally contain as large a proportion of mag- 
nesia as dolomite, in which there is about 45 per cent. of the car- 
bonate. We may be compelled therefore to fall back upon a heat- 
ed ocean—the same cause that crystallizes—for the source of the 
added magnesia. 
A strong argument in favor of the metamorphic origin of much 
‘of the primary limestone and its dolomization by the method 
proposed is found in some of its associated minerals, and espe- 
cially in the beds of serpentine or interspersed grains of this 
mineral. 
Serpentine appears to be a deposit from the ancient ocean, con- 
nected with or proceeding from the granitic eruptions, and altered 
through the action of heated magnesian waters. Some evidence 
of this is seen in its position in beds.and not in dykes; in its 
being so often associated with granitic and syenitic rocks, yet 
containing none of the elements of these rocks or but in small 
proportions, which should not be expected if they proceeded from 
simultaneous eruptions in the same regions; in its not altering 
the adjoining rocks like igneous ejections :—and more strongly 
still in its containing so large a proportion of water. It is a fact 
