50 
PROFESSOR W. N. HARTLEY ON SPECTRUM PHOTOGRAPHY 
A striking fact was noticed which, as will be seen, is exemplified in the case of 
magnesium, namely, that it is not always what appears to be the strongest and 
longest lines which first make their appearance when an impurity is discernible; and 
it was considered that many observations were necessary to ascertain which are the 
most persistent lines in the spectra of the elements. It was proposed to extend these 
observations, but their importance at that time did not seem to be appreciated. 
Almost all varieties of spark spectra have since been investigated, with the object of 
applying spectrum photography to the purposes of chemical analysis. From time to 
time opportunities have occurred when the process which has been gradually developed 
and rendered practical has been advantageously put into operation for the solution of 
questions upon which it would have been difficult to arrive at a decision by other 
means. For instance, I employed photographs of spark spectra in an examination 
of the rare earth bases separated from the mineral rhabdopbane, for testing the 
purity of certain cerium compounds, and for estimating the amount of beryllium 
contained therein (Journal of the Chemical Society, vol. xli., Transactions, 1882, p. 210). 
It may therefore be considered that the method is so far complete in detail that it 
maybe with advantage described for publication, notwithstanding that it is manifestly 
capable of improvement. 
At an early stage of the inquiry it was found necessary to devise some convenient 
method of examining solutions of salts, and ascertaining the best material from which 
to form electrodes. 
On a method of photographing spark spectra of the elements with solutions of their 
compounds. 
In the examination of substances for the purpose of chemical analysis it is un¬ 
desirable to use metallic electrodes even when this is possible, owing to the fact 
that alloys are not generally of a homogeneous character, and the sparks pass from 
only one or two spots situated at the surface of the electrodes. A better means of 
judging of the composition of a mass, whatever be its dimensions, is afforded by 
solutions, because the amount of metal volatilised by the spark from mere points of 
the substance is exceedingly small—according to Messrs. Lockyer and Roberts not 
more than 0*0001 gram. (‘Studies in Spectrum Analysis,’ Lockyer, p. 226). Messrs. 
Parry and Tucker (‘ Engineering,’ vol. xxvii., pp. 127, 429 ; and xxviii., p. 141) in 
the examination of iron and steel have made use of the charcoal points first used 
by Bunsen, and by dispensing with a Leyden-jar have photographed the lines 
characteristic of metallic solutions in such a manner as to render them comparatively 
free from the air spectrum. This method does not appear satisfactory to me, because a 
prolonged exposure is necessary, and with the reduced density of the spark, fluted 
spectra of air make their appearance while the metallic lines are too short. These 
bands or flutings are composed of a large number of fine lines, the distance between 
