Band and Line Spectra of the same Metallic Elements. 87 
than when the spectrum consisted of lines. Opposed to this view was the fact 
mentioned by Kayser, that in the arc there must have been a larger quantity of 
silver than that contained in the oxyhydrogen flame, when the spectrum of the 
metal in each case was under observation; nevertheless, from the arc, lines 
_ without bands were photographed (Kayser’s Handbuch der Spectroscopie, vol. ii., 
pp. 258 and 286, 1902), and it was suggested that the band spectrum and the 
continuous background of rays* observed in many flame-spectra might be 
accounted for by the flame containing unusually large atom-complexes, or 
molecular aggregates. In the case of monatomic molecules the vapour- 
densities of which are known,}t this appears to be inconceivable; for it is 
difficult to imagine a substance such as cadmium or zine with a monatomic 
constitution, which is known in each case from its vapour-density, at a tempera- 
ture well above its boiling point, to consist of molecular aggregates and not of 
simple molecules. At the same time it must be admitted that the remarks con- 
cerning the vapour-pressures of silver in the flame and arc respectively were very 
difficult to answer satisfactorily : first, because there are good reasons for doubting 
whether the vapour-pressure in the are referred to by Kayser was actually 
greater than in that part of the flame where silver is being boiled freely and 
continuously, the vapour being condensed like a cloud, and deposited on 
surrounding objects; and again, the conditions under which the vapour is 
produced in the arc and in the flame may be very different. ‘The banded 
spectra both of tin and of silver were observed in the are by Liveing and 
Dewar, and, more recently, A. Hagenbach{ has shown that the are may be 
made to yield both band and line spectra, even the spectra of compounds such 
as chlorides. The researches of Kayser and Runge, also of Rydberg, have 
reduced the constitution of line spectra in definite groups of the elements to a 
mathematical expression, of which the respective atomic weights are a function ; 
but the spectra are of a comparatively simple constitution, the band spectra of 
the same elements being more complicated. Both bands and lines of the same 
element may often be obtained in the same flame; silver and copper afford 
examples of this. The bands are numerous and complex; but the lines, common 
to the flame, are, and spark, are not. Thus the cause of the difference between 
the band and line spectra would appear, in the absence of further evidence, to be 
mainly a difference in temperature. That the bands are the spectrum of the 
* For a recent investigation see Proc. Roy. Soc., 1906, A., vol. 78, p. 403, ‘* Note on the Continuous 
Rays observed in the Spark Spectra of Metalloids and some Metals.” 
{+ Von Wartenberg (Metallurgie, 1906, 39, 381) has shown that from the relative density of silver 
vapour at 2000°, the molecule is monatomic. 
{ ‘‘ Ueber Banden-Spectra.” Festschrift. Aachen. June 1905. B.G. Teubner, Leipzig. 
