Pottok—Spectrographic Analysis of Commercial Thallium. 389 
are seen as short lines between the long lines of gold and 
thallium. To facilitate identification, some of the strong lines 
of gold are dotted and numbered from 10 to 25, as in my “ Index 
of the Principal Lines of the Spark-Spectra of the Elements.” 
Some of the lines of the impurities are also marked by dots 
placed over them, the elements identified being lead, tin, copper, 
and aluminium. The second spectrum on the plate shows the 
thallium superposed on a short spectrum of gold. The third 
spectrum shows the thallium superposed on a long spectrum 
of gold. The fourth and last spectrum is the metallic spark- 
spectrum of the specimen of commercial thallium alone. The 
last three spectra are not absolutely necessary, but are useful 
to distinguish between the lines of the gold electrodes and the 
thallium, and in case of coincident, or nearly coincident, lines 
would enable such lines to be detected. 
In measuring the plate to determine the wave-lengths of the 
lines by use of an interpolation curve, and thus to identify them, 
only the position of the short lines need be noted, together with 
a sufficient number of the gold or thallium lines, for purposes of 
interpolation. To prevent confusion when the negative is dry, 
first dot the gold lines from 9 to 25, placing the dots just over 
the spectrum; then, if necessary, mark the lines it is specially 
required to identify by dots or other marks placed preferably 
under the spectrum, and not as shown in the Plate. 
The interpolation curve used was drawn on ruled paper divided 
into inches and tenths of an inch, on a scale of an inch to 
every zyo Of an inch measured on the spectrum, and 5, of an 
inch to each Angstrom unit of wave-length, and the position of 
the gold line No. 9 corresponded with the reading 283°85 on the 
curve. The measurements of the spectrum were made with a 
Hilgar micrometer reading to >54,> of an inch, and with a 
movable sub-stage to facilitate the adjustment of any line to a 
particular reading; and the gold line No. 9 was thus adjusted 
to the reading 283°85 on starting the measurement of the plate. 
In plotting and measuring, an inch-scale will be found much 
more convenient for accurate reading, and not nearly so trying 
to the eye-sight, as a millimetre-scale. 
1 Scientific Proceedings of the Royal Dublin Society, vol. xi., pp. 184-216. 1907. 
SCIENT. PROC. R.D.S., VOL. XI., NO. XXXII. 2Y 
