[HARRINGTON | MINERALOGICAL CHEMISTRY 7 
The thanks of chemists and mineralogists are due to Professor Penfield, 
of Yale, for his recent and valuable paper “On Some Methods for the 
Determination of Water,” which have proved useful in his hands. 
There is another question which is important more especially in con- 
nection with the analysis of crystalline rocks, and to which I have seen no 
reference in the text-books. The rocks often contain iron pyrites, and if 
no allowance is made for this in estimating the alkalies by Lawrence 
Smith’s method, the results are entirely erroneous. In the fusion with 
calcium carbonate and ammoniun chloride, the sulphur of the pyrites 
becomes oxydised, and sulphates of the alkalies are produced along with 
the alkaline chlorides and weighed as chlorides. If one considers the dif- 
ference in the molecular weights of the alkaline sulphates and chlorides it 
is obvious that the error introduced in this manner may be very serious. 
The difficulty may be obviated in different ways—in some cases by roast- 
ing the powdered mineral before making the fusion for alkalies, in others 
by removing the pyrites with nitric or nitro-hydrochloric acid ; or again, 
by entirely converting the alkalies into sulphates and weighing as such, or 
by transforming the sulphates into chlorides by one of the well-known 
methods. 

Considerable attention has been devoted of late years to the, examin- 
ation of minerals for elements which in many cases have escaped detection, 
but there is room for a large amount of research in this direction. The 
importance of this line of work is accentuated by the recent discovery by 
Professors Ramsay and Crookes, who find that the gas contained in the 
mineral cléveite, a variety of uraninite, is a mixture of helium and argon 
with a little nitrogen. If we look at the early anaylses of uraninite we 
find no indication of any gaseous constituent (apart from the oxygen), 
but in 1890 Hillebrand (to whom we are indebted for most of our know- 
ledge as to the chemical composition of uraninite)* showed that on fusing 
the mineral with an alkaline carbonate, or on treating it with a non- 
oxydising inorganic acid, a gas was liberated which appeared to agree in 
its properties with nitrogen. The discovery was regarded with much in- 
terest, because it was the first case in which nitrogen had been found in 
any mineral constituting the original crust of the earth. But now in the 
gas extracted from cléveite, Ramsay and Crookes recognise not only nitro- 
gen but argon and helium, a hypothetical element of the solar spectrum.” 
Other examples of the necessity of re-examining minerals whose com- 
position was long ago supposed to have been settled, are to be found in 
the case of the scapolites, many of which were shown by Dr. Adams to 

1 Am. Jour. Sci., July, 1894, p. 31. 
2 Am. Jour. Sci., Nov., 1890. 
3 The above is left as written. The fact that subsequent observations failed to 
confirm the presence of argon along with the helium in cleveite does not alter the 
argument. 
j \ 
