Hillbbeand.] QUALITY Of KKAGENTS. 1>5 
uents. It is quite as likely, in fact more than likely, to affect a single 
determination and one which may be of importance in a critical study 
of the rock from the petrographic side. 
VII. QUALITY OF REAGENTS. 
It is due to say that all analyses performed in the Survey laborato- 
ries have been made with the purest reagents obtainable, either by 
purchase in the open market or by special preparation on the part of 
manufacturers or in the laboratory. The best acids made in this coun- 
try are of a high grade and need no redistillation except for .special 
experiments. Ammonia has always been redistilled at short intervals; 
and no sodium carbonate which exceeds 2£ milligrams of total impurity 
(sec p. 50) in 20 grams (0.012 per cent) is used for the main portions, in 
which silica, alumina, etc., are to be estimated. For other portions, as 
phosphoric acid, fluorine, sulphur, a poorer grade is entirely allowable, 
provided it is free from the element to be determined, and from any 
other which might interfere with its estimation. 
Hydrofluoric acid was always freshly distilled with potassium per- 
manganate until the introduction of ceresine bottles afforded an article 
sufficiently pure for all but the most exacting work. Care must be 
exercised even yet, however, that no particles of paraffin or ceresine 
are floating on the acid, and that the latter is free from traces of chlo- 
rine whenever it is to be used for attacking silicates with a view to 
estimating chlorine (p. 103). 
Potassium bisulphate has usually been prepared in the laboratory 
from sulphuric acid and potassium sulphate, since it is not always to 
be bought of satisfactory quality. Even then the normal sulphate 
had flrst to be examined, for it has been found to contain, on different 
occasions, notable amounts of lead, calcium, and silica. 
The phosphorus salt used for precipitating magnesium . has been 
found to contain iron, and calcium is almost always a constituent of 
ammonium oxalate. The latter has therefore to be purified or specially 
prepared, as also oxalic acid, ammonium chloride (in which latter 
manganese has been observed), and occasionally other reagents. Some 
hydrogen peroxide contains fluorine, which renders it unfit for use as 
a chemical reagent. 
A ct C. P.'' label is no guaranty whatever of the purity of a reagent ; 
hence no chemicals should be taken on trust because of bearing such a 
label. Every new purchase should be examined, if it is one in which 
purity is a desideratum. In general all so-called "C. P." chemicals 
should at least stand the tests laid down by Krauch. 1 
Of late years the appearance upon the market of so-called guaran- 
teed reagents promised to meet a long-felt want. But experience has 
1 Die Priifung der ehemisehen Reagentien, 3d cd., Berlin, Julius Springer, 18%. 
