55 
IN RELATION TO QUANTITATIVE CHEMICAL ANALYSIS. 
The same photographs are obtained when only a small jar is used and the electrodes 
are cold. 
From these experiments it is evident that cooling the electrodes intensifies the 
spark, and consequently lengthens the short lines, just as heating them causes the 
reverse effect. The reason why heating the negative electrode acts so powerfully 
upon the spark is because the high temperature of the pole discharges the jar 
(Guthrie’s ‘Magnetism and Electricity,’ p. 84). 
In addition to the lengthening of the carbon lines, hydrochloric acid and solutions 
of metallic chlorides yield a weak continuous spectrum, and also a series of closely 
placed lines or flutings, further details concerning which are given on page 59. The 
continuous rays are most noticeable in spectra of solutions of chlorides of the alkali 
metals and of aluminium chloride; the flutings, however, appear most prominent in 
concentrated solutions of aluminium chloride and of zinc chloride. 
Dry electrodes in air showed in some instances a total absence of the lines, with 
wave-lengths 4266*3, 3919*5, 3881*9, 3875*7, 38707, 3590*0, 3585*0, 3583*5, 2746*6, 
and 2640*0, while 2836*8 and 2836*0 were exceedingly faint. These lines are marked 
on the map with a star. No reason can be assigned at present for this difference in 
the spectrum, as the lines 4266*3, 3919*5, 3881*9, 38757, and 3870*7 are visible when 
the atmosphere surrounding the poles is carbon dioxide, and the photograph was 
perfect as regards the other part of the spectrum. A difference in the strength of the 
spark may occasion some similar alteration, but such a cause did not operate in this 
case. 
On the spectra of solutions of binary compounds. 
A complete series of photographs of metallic salts, chiefly chlorides, corresponding 
to the electrodes already enumerated and examined, was executed (Journal of the 
Chemical Society, vol. xli., p. 90). The reason for taking chlorides was that as a rule 
they are among the most soluble of salts. The solutions examined consisted of the 
chlorides of magnesium, zinc and cadmium, aluminium, thallium, iron, cobalt and 
nickel, arsenic, copper, strontium and tin. The cadmium, copper, and stannous 
chlorides were made by dissolving the electrodes used for the series of photographs of 
the elements. The zinc chloride was made from a very pure specimen of distilled 
zinc, the only impurity it betrayed being a trace of cadmium. The arsenic trichloride 
was obtained from a very carefully purified specimen of arsenious oxide. The ferric 
chloride was prepared from fine pianoforte wire. The thallous chloride was a 
particularly pure specimen made from the metal used as electrodes. 
Some preparations of barium and strontium chlorides made with great care were 
found to contain calcium. They were purified by treating saturated solutions with 
an equal volume of strong hydrochloric acid, whereby the chlorides are precipitated in 
a crystalline form, and by filtration through a plug of asbestos the acid liquid is 
separated therefrom, the crystals are washed with strong acid twice, after which 
