f 
J. L. Smith on determining the Alkalies in Minerals. 61 
arises from the fact that the chlorid of calcium has but little de- 
composing effect on the silicates, its action being simply that of a 
menstruum in which the lime can act conveniently on the 
mineral. 
62. The use of lime or its carbonate mixed with chlorid of 
calcium or chlorid of ammonium, for the purpose of effecting the 
decompositions alluded to, would be considered by me of ques- 
tionable utility if the mixture were not so proportioned, and em- 
ployed as to decompose the most difficult silicates if necessary ; 
for unless this be done, we can at no time be certain that the 
decomposition is complete. As a general rule for decomposing 
silicates by lime or soda, it is far better to use a charcoal fire than 
the flame of a lamp, as it is better to heat too high than not to 
heat sufficiently the mineral to be acted on. 
Complete analysis of an insoluble silicate on one portion of the 
Mineral. 
63. The effort to accomplish an analysis of this description 
deserves no encouragement, from the almost invariable inaccuracy 
attending the results. If we havea given quantity of any one of 
the silicates alluded to, requiring analysis, we had better subdivide 
it, however small the entire quantity may be, ascertain one set of 
ingredients by the soda, and the other by the lime fusion ; for 
the results thus obtained, may be relied on as more accurate than 
those furnished by an analysis of the whole quantity, through 
the agency of baryta or hydrofluoric acid. 
64. Should it be desired to undertake the analysis on a single 
portion, I would recommend the silicate to be attacked with car- 
bonate of baryta mixed with the chlorid—three to four parts of 
carbonate and two of chlorid: this mixture can be made te de- 
compose all silicates at a much lower temperature, than when 
the carbonate alone is used, but its action is not near so powerful 
as the carbonate of lime and sal-ammoniac. 
. This terminates an account of my labors in the determin- 
oO 
Many analytical processes mentioned in this article can be ap> — 
plied when operating on soluble silicates. 
Laboratory of the University of Virginia, May 10th, 1853, 
