0287 4 
XIX. 
ON SOME DEVICES FACILITATING THE STUDY OF 
SPECTRA. 
By WALTER NOEL HARTLEY, D.Sc., F.R.S., 
Royal College of Science, Dublin. 
[PratE XX.] 
[Read, May 21; Received for Publication, May 24; Published, Aveusr 17, 1907.] 
Ir has been shown in previous communications that flame-spectra 
at high temperatures have a special value, inasmuch as minute 
traces of metallic and mineral substances may be readily detected, 
and their spectra photographed ; for instance, from iron ores and 
pig-iron as many as ninety lines of the element are photographed 
at one exposure.’ 
The source of heat being the oxyhydrogen flame, the tempera- 
ture lies between 1400° and 2000°; and as 1775° is about the 
melting-point of platinum, some other support than that metal 
must be used for solid substances. Thin slips of Donegal 
cyanite and ashless filter-papers have been used almost exclusively, 
and their use described in the publications quoted. The cyanite 
consists of 98:0 per cent. of aluminium silicate, according to an 
unpublished analysis made in my laboratory ; it merely softens in 
the flame, and it is useful for long exposures of half an hour or 
upwards. ‘The lines of sodium and lithium in the yellow and red 
are the only impurities which are photographed. The filter-papers 
are useful for rapid exposures of one to two minutes; they yield 
the sodium line oniy ; but atmospheric dust settles upon them, 
and consequently feeble red and green bands of calcium some- 
times appear, especially when ten filter-papers are used for one 
1 Flame-Spectra at High Temperatures,’’? Hartley, Phil. Trans., 1894, A, 
vol. clxxxv., part 1, pp. 161-211. 
‘‘ A Simplified Method for the Spectrographic Examination of Minerals,’ 
Hartley & Ramage, Chem. Soc. Trans., 1901, vol. Ixxix., p. 61. 
SCIENT. PROC. R.D.S., VOL. XI., NO. XIX. 2H 
