326 
PROFESSOR W. N. HARTLEY ON SPECTRUM PHOTOGRAPHY 
Analytische Chemie, vol. xiv., pp, 297-301 ; also Journal Chem. Soc., 1876, p. 550, 
Abstract). Messrs. Liveing and Dewar have published notes on quantitative spec¬ 
troscopic experiments (Proc. Roy. Soc., vol. xxix., pp. 482-489). Observing the 
emission spectrum of sodium vapour, they sought to estimate the quantity of substance 
present in a given space by measuring the width of the sodium lines. 
It has been already shown that solutions containing the same element in different 
proportions emit different variations of the same spectrum, the lines differing in 
number, length, and intensity. I have also proved the converse of this—namely, that, 
under the same spark conditions similar solutions of the sarnie strength always emit the 
same spectrum. By similar solutions are meant solutions containing salts of the same 
metal, the non-metallic constituents of which cannot affect the length or strength of 
the metallic lines. Evidence of a quantitative nature in Part I. has established the 
fact that chlorides, nitrates, sulphates, and carbonates are salts of a similar character 
in this respect, and it has also been proved quantitatively in the case of chloride and 
sulphate of copper. It is, however, doubtful whether solutions of borates and silicates 
are strictly comparable with the solution of other salts, for the borates and silicates of 
the alkalies emit strong lines, due to boron and silicon, and it certainly appears, from 
the observations I have hitherto made, that the metallic spectra are modified by the 
strong lines of these non-metallic elements. 
On the strength and length of lines in the metallic spectra\ yielded by solutions of 
various strengths, and on the limit of sensitiveness of the spectrum reaction. 
The following investigation was made with the view of placing our knowledge of 
spectrum analysis on a proper basis, so as to admit of its being applied more systemati¬ 
cally than hitherto in the examination of minerals in quantitative analysis, and parti¬ 
cularly in the assay of certain metals, 
Experiments have been made chiefly on those elements with comparatively simple 
spectra; those, for instance, which yield a few prominent long lines and many short 
ones, for since it has been shown that short lines from metallic electrodes become long 
ones when strong solutions are used, and short again as the solutions are diluted, we 
may anticipate a greater variety of spectra and a greater range for the exercise of 
observation, or, in other words, a greater sensitiveness of the spectrum reactions, 
than with metallic points. As often as possible chlorides have been used in preference 
to other salts, because hydrochloric acid is the solvent employed most generally in 
analytical operations, and the chlorides are as a rule the most soluble compounds of 
the metals, and likely to yield the most highly concentrated solutions. In a few pre¬ 
liminary experiments trial or test-plates were made for use in quantitative analysis, 
each containing about fifteen photographs of the spectra of solutions of various 
strengths. It was frequently found desirable to largely increase the number of photo¬ 
graphs where gradual alterations in the lines accompanied gradual dilution. Subse- 
