IS Ye: 
is 4 
se 
an 
196 Mr. Silliman, Jr. on the Composition of Corals. 
G. The alcoholic filtrate from (F'), containing magnesia and 
lime, is treated by the well known methods of analysis for the 
estimation of those substances. ‘The lime in all cases in these 
researches, was converted into sulphate and precipitated by alco- 
hol. ‘The magnesia was estimated as phosphate. 
H. The crystalline precipitate from (F'), which was collected 
on the filter, contained the alumina and iron, previously in com- 
bination with phosphoric acid or fluorine. This precipitate 1s 
boiled in a capsule with a strong solution of carbonate of soda, 
to decompose the sulphate of lime; it is then filtered, the insol- 
uble residue washed thoroughly and treated with chlorohydric 
acid, the precipitate by ammonia from (F') being added, and the 
whole treated with excess of ammonia. Alumina and iron fall, 
which may be afterward separated in the usual way; but this 
was generally not deemed requisite, the quantity of iron being 
very small in most cases. 
I. The filtrate from (H) is treated for lime by oxalate of am- 
monia, and the oxalate converted into sulphate and weighed: 
this dose of lime had been previously united to phosphoric acid 
or fluorine. 
J. Magnesia is next separated from the filtrate of (1), by am- 
monia and phosphate of soda. 
K. The alkaline liquor from (H) contains another portion of 
nesia, which is separated in like manner as the ammonio- 
phosphate. Much labor is saved if we take care to reserve the 
several portions, from which magnesia has been thrown down, 
and unite them in one filtration and weighing, instead of treating 
them as so many separate portions. 
The minute détermination of all the constituents of the lime- 
water precipitate, was attempted only on those specimens of 
which we had a large quantity at command; for a solution of 
half a pound or more of the coral in nitric or chlorohydric acid, 
was necessary to afford sufficient precipitate for analysis. The 
carbonate of lime, by far the most abundant constituent, was sep- 
arately determined on one gramme, as sulphate of lime, and from 
this the carbonate was calculated, The ratio of phosphates and 
fluorides of the several bases to the entire mass was also deter- 
mined from a distinct portion of coral, two grammes by weight ; 
and from the data thus furnished, we have the means of safely 
_ estimating the organic matter by the loss. 
2 
