186 Scientific Proceedings, Royal Dublin Society. 
gold, it is necessary to employ fairly thick wire; but economy can 
be effected by using only a short piece attached to a thin platinum 
wire, of sufficient length to pass through the capillary. When 
sparking, the electrode in the solution must be made the negative 
electrode. ‘lhe upper or positive electrode may be held in any 
convenient manner, and the spark-gap should not be too long. 
A condensed spark should be used, and, if desired, a Hemsalech 
coil, to remove air-lines; a photograph is first taken with the 
electrodes alone, and a long slit, giving about one minute’s 
exposure ; the slit is then reduced to about one-third of its 
length, and the substance under examination also sparked for 
about one minute. ‘The lines of the substance are then seen as 
short lines between the long lines of the gold electrodes, and are 
easily identified by aid of the gold lines. In all other respects the 
work should be conducted as described by Hartley in the Journal 
of the Society of Arts, March 12th, 1886; see also ‘“‘ Thorpe’s 
Dictionary of Applied Chemistry,” vol. iii., p. 536. 
The first spectrum on Plate X. is that of cadmium with gold 
electrodes. Graphite and platinum electrodes can also be used, 
but are not so convenient for general purposes. Graphite has most 
of its lines concentrated in one part of the spectrum, so as to 
obscure any other lines there, and has too few lines in the other 
parts of the spectrum to assist in identification, and it is often 
very impure. Platinum has an inconveniently large number of 
fine lines; but where lines coincident, or nearly coincident, with 
gold lines are suspected, graphite or platinum can be used. The 
second and third spectra on Plate X. are those of platinum and 
carbon photographed on gold. The carbon spectrum is from 
Ceylon graphite, and shows the lines of silicon and magnesia, 
and many of the iron lines. The fourth spectrum shows the air- 
lines introduced by dispensing with the use of a Hemsalech’s self- 
induction coil. The fifth and sixth spectra show the silver and 
copper lines photographed on a gold spectrum with short lines. 
Silver is practically always present in gold, and a very minute 
quantity gives certain lines quite strongly, as will be seen by 
comparison with the previous gold spectra; but this is rather a 
help than otherwise, as these lines are useful for reference, and 
fill up blanks in the gold spectrum. Copper is the only other 
element likely to be present as an impurity in gold electrodes. 
