340 Scientific Proceedings, Royal Dublin Society. 
In drawing the curve, the spectroscope was adjusted to the 
best possible focus, and a photograph taken of the spark-spectrum 
of metallic cadmium, then of cadmium superposed on gold, and 
lastly of gold. The gold-cadmium spectrum was accurately 
measured, and the first pencil curve drawn from the cadmium 
measurement. The gold measurements were then dotted in 
very carefully with red dots, and the curve drawn accurately in 
ink to the gold dots. After drawing a curve, the particulars 
of the spectroscope used, the adjustments of the instrument, and 
the measurements of the fiducial lines, should all be carefully 
recorded on the first sheet of the curve, and an actual print 
of the negative, from the measurements of which it was drawn, 
pasted on. 
No matter how accurately a curve is drawn, it will be found 
practically impossible always to bring the spectroscope back to 
exactly the same adjustment; and even if the instrument is not 
touched, temperature changes, in both the spectroscope and the 
micrometer, will introduce errors in the measurement of lines 
of quite a serious character. So that, from one cause or another, 
with photographs of spectra of 4 or 5 inches in length, an 
apparent change of dispersion of ;}9 of an inch will be quite 
common, and would introduce a very serious error of a wave- 
length or so. It is, therefore, absolutely essential, in measuring 
spectra for purposes of identification of lines, to measure a number 
of the fiducial lines at the same time; and on every plate on 
which measurements may ever require to be made, a photograph 
of the gold or cadmium spectrum should appear at the top of 
the plate, if they do not occur on the plate as electrodes or 
otherwise. It is then always possible to verify measurements, and 
curve, from the photograph itself. In making measurements, it 
is very convenient to have foolscap sheets ruled and headed on 
which to record the measurements, intensities, and subsequent 
reduction to wave-lengths, and the final identification of the 
elements to which the lines are due. The Table at the end is 
an exact reproduction of the original sheet of measurements, 
corrections, and identifications of the lines of the first spectrum 
on the plate. 
Column one, headed “ Measurement, 7}, inch,’ gives the 
readings of the micrometer in hundredths of an inch, the line 
