204 CORALS AND CORAL ISLANDS. 
carbonate of magnesia, magnesia is largely present in some 
specimens of the rock. The rock is hard (H. = 4), and splint. — 
ery in fracture, with a specific gravity 2°690. It affords 
on analysis, 38°07 per cent. carbonate of magnesia, and hence, 
only 61°93 of carbonate of lime. 
Another specimen from the same island, having the spe- 
cific gravity 2°646, afforded 5:29 per cent. of carbonate of mag- 
nesia. 
The former was a compact homogeneous specimen, and the 
latter was partly fragmentary. Recent examinations of coral 
sand and coral mud from the islands, give no different com- 
position, as regards the magnesia, from that for corals, which, 
as the analyses on page 99 show, contain very little or no 
magnesia. The coral sand from the Straits of Balabac, af- 
forded Prof. Silliman carbonate.of lime 98°26, carbonate of 
magnesia 1°38, alumina 0°24, phosphoric acid and silica a trace. 
This introduction of magnesia into the consolidating 
under-water coral sand or mud, has apparently taken place (1) 
in sea waters at the ordinary temperature; and (2) without 
the agency of any mineral waters except the ocean. But the 
sand or mud may have been that of a contracting and evap- 
orating lagoon, in which the magnesian and other salts of the 
ocean were in a concentrated state. It has been already ob- 
served (p. 349), that this was probably the actual condition 
of the elevated portion of the island of Metia, every thing 
about it looking as if it corresponded to the lagoon part of 
the old atoll; and also that the idea of the existence of min- 
eral springs there has no support in known facts. 
X. FORMATION OF CHALK. 
The formation of chalk from coral is known to be ex- 
emplified at only one spot among the reefs of the Pacific 


