[ 49 ] 
IV. Researches on Spectrum Photography in relation to New Methods of Quantitative 
Chemical Ancdysis. —Part I. 
By W. N. Hartley, F.R.S.E., dc., Professor of Chemistry, Roycd College of Science, 
Dublin. 
Communicated by Professor Stokes, Sec. R.S. 
Received June 20—Read June 21, 1883. 
[Plate 3.] 
Contents. 
Introduction.'. 
On a method of photographing spark spectra of the elements with solutions of their compounds .. 
Alterations in the spectra of graphite and other electrodes caused bj the presence of liquids .... 
On the spectra of solutions of binary compounds... 
On the spectra of ternary compounds. 
On the spectra of dilute solutions and alterations caused by prolonged exposures . 
Of the presence of impurities in certain spectra ... 
Rage 
49 
50 
53 
55 
58 
60 
61 
Introduction. 
In the year 1878 I submitted to the Council of the Chemical Society a series of 
photographs executed with roughly constructed and imperfect apparatus, which 
showed variations in the number, length, and strength of the lines exhibited by some 
of the elements in the ultra-violet region. These photographs proved that by 
alterations in the exposure of the photographic plates certain impurities became 
visible in metals otherwise apparently pure, without any changes being noticeable in 
the spectra such as might be introduced by over-exposure. Thus by increasing the 
period of exposure three-fold the lines of iron were plainly seen in a spectrum 
regarded as that of pure aluminium. Gold obtained in as pure a condition as possible 
by parting gave evidence of the presence of silver. Two specimens of indium were 
examined : they both yielded the flame spectrum and the reactions usual with this 
metal; one of them showed no strong lines in the ultra-violet which could not be 
attributed to tin, lead, or cadmium ; the other contained tin and cadmium in such 
proportions that the lines of these metals were more prominent than those of indium. 
A third specimen, prepared by Professor Richter, of Freiberg, yielded no spectral 
lines which could be attributed to any foreign metal. 
MDCCCLXXXIY. H 
