100 SOME PRINCIPLES AND METHODS OF ROCK ANALYSIS, [bull. 176. 
though not absolutely essential to the attainment of fairly correct results, to decant 
the liquid from the residue, wash the latter a little with anhydrous amyl alcohol, 
dissolve in a few drops of water, and repeat the separation by boiling again in amyl 
alcohol. For washing, amyl alcohol, previously dehydrated by boiling, is to be used, 
and the filtrates are to be measured apart from the washings. In filtering it is best 
to make use of the perforated crucible and asbestos felt, and apply gentle pressure. 
The crucible and residue are ready for the balance after drying for a few minutes 
directly over a flame turned low. The weight of insoluble chlorides actually obtained 
in this manner is to be corrected by the addition of 0.00041 gram for every 10 cm. 3 of 
amyl alcohol in the filtrate, exclusive of washings, if the insoluble salt is entirely 
sodium chloride, 0.00051 gram for every 10 cm. 3 if potassium chloride constitutes the 
residue, and if both sodium and potassium chlorides are present, 0.00092 gram; 
but * * * the entire correction may in any case be kept within very narrow 
limits if due care be given to the reduction of the volume of residual alcohol before 
filtration. The filtrate and washings are evaporated to dryness, treated with sul- 
phuric acid, the excess of the latter driven off, and the residue ignited to fusion 
and weighed. From the weight thus found the subtraction of 0.00050 gram is to be 
be made if sodium chloride constitutes the precipitate, 0.00059 gram if potassium 
chloride alone is present in the residue, and 0.00109 gram if both these chlorides are 
present, for every 10 cm. 3 of filtrate, exclusive of washings. 
Amyl alcohol is not costly, the manipulations of the process are easy, and the only 
objectionable feature — the development of the fumes of amyl alcohol — is one which 
is insignificant when good ventilation is available. 
The process has been used for some months frequently and successfully, by others 
as well as by myself, for the estimation of lithium in waters and minerals. 
SEPARATION OF ALKALIES BY OTHER METHODS. 
When, as may happen in rare instances, it is necessary to estimate 
alkalies in the main portion after elimination of silica, alumina, lime, 
etc., in one of the usual ways, the question of a suitable method for 
separating magnesium becomes important. 
The mercuric oxide method. — The old barium-hydroxide method is 
not to be recommended. The mercuric oxide method of Zimmermann, 
whereby the magnesia is precipitated from solution of the chlorides 
by moist, freshly precipitated, and alkali-free mercuric oxide, can give 
satisfactory results. The oxide is added in excess to the solution in a 
platinum crucible and evaporated to dryness. Then the mercuric 
chloride and most or all of the excess of oxide are expelled by cautious 
heating. On leaching with water the magnesia remains on the filter. 
With more than 1 per cent of magnesia the operation must be repeated 
(Dittrich). 
The ammonium carbonate method. — Lately the once favored method 
of precipitating the magnesium by neutral ammonium carbonate in 
concentrated solution has been again recommended. 1 The magnesium 
solution must be as strongly concentrated as possible, and a great 
excess of ammonium carbonate solution must be used. A voluminous 
1 Wiilflng: Ber. deutsch. chem. Gesell., Vol. XXXII, p. 2214, 1899. The neutral carbonate is pre- 
pared by dissolving 230 grams of ammonium carbonate in 180 cubic centimeters of ammonia of 0.92 
specific gravity and enough water to make 1 liter. 
