F888) 
XXXII. 
ON THE SPECTROGRAPHIC ANALYSIS OF A SPECIMEN 
OF COMMERCIAL THALLIUM. 
By JAMES H. POLLOK, D.Sc., Royal College of Science, Dublin. 
[Puats XXVII.] 
[Published Frsruary 26, 1909. ] 
Tue following spectrographic analysis was made to determine 
the mineral impurities present in a specimen of commercial 
thallium received from Professor Hartley, and is a convenient 
illustration of the general method of procedure in any analysis 
of the kind, and the details of the measurement and identification 
of the lines may be of use to others engaged in similar work. 
The difficulty in determining traces of impurities is, that the 
lines due to those impurities are necessarily faint and few in 
number, and there are always a great number of faint lines due 
to an element that are not recorded in ordinary tables of wave- 
lengths, or, if recorded, may very possibly be due to the same 
impurity in the photograph of the original observer. To 
distinguish between faint lines due to impurities and those due 
to the element itself, a photograph is first taken with a pure 
specimen of the metal, or a solution of some pure salt of the 
metal, using a long slit. The slit is then shortened to a third 
of its length, and a photograph of the metal under examination 
taken on top of the last; in this way the first photograph cancels 
all the lines due to the metal and electrodes, and only lines due 
to impurities show in the short spectrum, and can thus be very 
quickly picked out, measured, and identified. 
The first spectrum on Plate XX VII. shows a short spectrum 
of the sample of commercial thallium, superposed on a long 
spectrum of a solution of pure nitrate of thallium sparked 
between gold electrodes, and the lines of the impurities present 
