PotLtoKk— Lines of the Spark-Spectra of the Elements. 185 
is absent, no matter how many other lines appear to agree with 
aluminium lines. 
Analysis of Minerals. 
To prepare minerals for spectrographic analysis they should be 
decomposed in the ordinary way—the solution separated into 
the various analytical groups, and these groups then dissolved 
in hydrochloric acid, or, if necessary, nitric acid, ammonia, or 
caustic soda, and made up to a fixed volume, such as 20 c.c. each 
for every two grams of sample taken, and these solutions then 
sparked. If only a small quantity of an element is present, much 
more of the original material must be taken, and the solution made 
of such a strength as to contain not less than 1 per cent. of the 
element sought for. Iron should be separated from the ammonia 
precipitate by tartaric acid and sulphide of ammonia, as there are 
so many iron lines that it is difficult to detect other elements when 
it is present. A somewhat similar difficulty occurs with cerium, 
but cerium can be completely removed by caustic soda and 
chlorine. 
Silica gives no lines in an acid solution, but very characteristic 
lines when dissolved in caustic soda or with soluble silicates. 
Tungsten gives its lines when dissolved in ammonia; the other 
metals may be sparked in acid solution. The non-metallic elements 
present do not give spectra when treated in this way ; they must 
be sparked in the gaseous condition, at reduced pressure, in 
Geissler tubes. 
Sparking. 
In sparking solutions the most convenient electrodes to use in 
conjunction with this index are gold. A piece of glass-tube, about 
6 inches long, and having a capillary bore, is bent into the form 
of a U with limbs of unequal length, and a small cup, like a 
miniature thistle-funnel, of about 2 ¢.c. capacity, blown at each 
end. The gold wire is then passed through the capillary bore of 
the tube, so as to project a little above the rim of the lower cup, 
and a small piece of thin-walled capillary tube slipped over the 
end of the gold wire, and made of such a length as just to reach 
the top of the wire, and thus draw up the solution and keep the top 
of the wire well wetted when sparking. Owing to the fusibility of 
1 Hartley, Proc. Roy. Soc., vol. xxxv., p. 301; also Brit. Assoc. Report, 1883. 
2B2 
