54 
PROFESSOR W. N. HARTLEY ON SPECTRUM PHOTOGRAPHY 
lines takes place as in the case of copper and graphite. The lines of iridium yielded 
by dry electrodes extend no more than one-fourth the distance towards the opposite 
pole. When one pole is a point of iridium and the other of copper, the two series of 
short lines are seen, but those of iridium only are visible when that metal con¬ 
stitutes the negative pole, though the copper lines are weak, but still visible, at the 
positive pole. 
When an iridium point, as the negative electrode, is partially immersed in water 
all the short lines are increased in length, so that they stretch four-fifths of the 
distance towards the copper ; if the copper be negative and immersed in water the 
copper lines are lengthened, and but few of those belonging to iridium are discernible. 
The same lengthening of short lines takes place when gold electrodes are moistened. 
We have but few facts which serve to indicate the constitution of the electric spark 
and the circumstances under which it is altered in character. It may, however, be 
affirmed, with some degree of truth, that an increased intensity of the spark, such as 
is gained by placing a jar in circuit, which is usually considered to correspond to an 
increase of temperature, causes an increased length in the short lines, and also that 
the more volatile the metal the longer the lines. It was considered necessary to 
ascertain the effect of heating the electrodes, and accordingly two points of iridium wire 
were taken as suitable. AVhen the spark passed at the ordinary temperature, a photo¬ 
graph which received five minutes’ exposure showed, beside the large number of short 
lines, a continuous band of rays extending through the whole spectrum. On the same 
plate, and immediately beneath this, another photograph was taken while the negative 
electrode was heated to the most brilliant incandescence by the oxy-hydrogen blow¬ 
pipe. The spark continued to pass during the whole period of five minutes, as was 
made evident by the scintillations visible at the cooler electrode. This latter was 
kept at a temperature sufficiently low as to show no signs of redness. The passage 
of the spark, but for the scintillations, was otherwise invisible and almost noiseless. 
On turning down the gas the spark appeared of a pale blue colour, the electrode being 
at a red-heat, and it now emitted a moderate crackling sound, similar to that caused 
when no jar is in circuit. On turning the gas out the spark grew brighter as the 
metal became colder, till finally the usual brilliancy and the sharp crackling noise 
were resumed. The second photograph, with the same exposure of five minutes, 
showed barely a trace of the iridium spectrum; some feeble rays emitted by the flame 
of the blow-pipe were visible, but the chief feature of the photograph was the 
beautiful group of nitrogen lines seen when the spark is passed between points 
without a jar in circuit. The wave-length of the line at the commencement of one of 
these groups was 3369, the commencement of the second group was a line with wave¬ 
length 3211’6, the spectrum terminating with a line with wave-length 3062’4. 
At a low red-heat the spark passes with only a moderate crackling sound, and the 
spectra photographed under these conditions consist of the two spectra of ail'—• 
namely, those of the first and second order superposed. 
