528 Prof. R. Bunsen's Spectral- Analytical Researches. 



Next to the breadth of slit the intensity of the source of light 

 exercises the greatest influence upon the formation of the spec- 

 trum. Vapours, when exposed to the high temperature of the 

 Ley den-jar spark, often exhibit a great number of lines, which 

 either disappear or become too faint for recognition at the lower 

 temperature of the simple spark or non-luminous gas-flame. 

 With increased intensity of light continuous spectra are often 

 produced, in consequence of which many of the feebler lines, 

 which would otherwise be perfectly apparent by contrast, more 

 or less escape recognition. The fact, first announced by Kirch- 

 hoff, that the relative intensity of the individual lines does not 

 alter proportionately with increase in temperature of the source 

 of light, explains the other fact that these lines, which are 

 relatively feeblest in the flame-spectrum, sometimes become the 

 most marked in the spark-spectrum, as is especially the case 

 with the spectrum of lithium. On the other hand, the appa- 

 rently contradictory fact has been noticed, that the flame-spec- 

 trum of several substances is much sharper and richer in lines 

 than the spark-spectrum ; for instance, the lines which are so 

 characteristic in the flame-spectra of caesium and of rubidium 

 disappear, the former entirely, the latter to a less degree, in the 

 spark-spectra of the same elements. This, however, is easily 

 explained when we consider that the light proceeding to the slit 

 from the glaring mass of flame (which is of larger dimensions 

 although not of so high temperature as the spark) is of greater 

 intensity than that which proceeds from the train of sparks, 

 where the quantity of gas is very limited although of extremely 

 high temperature. Those substances whose spectra become 

 apparent at low temperatures are best examined in the gas- 

 flame, and not with the spark. The principal spectra belonging 

 to this class are those of the alkalies, of the alkaline earths, of 

 iridium, thallium, and some others. Such spectra are most 

 perfectly produced by bringing the little beads to be tested 

 supported on the unlooped end of a very fine platinum wire, 

 into the fusion-zone of a non-luminous gas-flame. 



From these considerations it follows that the practical appli- 

 cation of spectra necessitates the possibility of a variation in 

 the spark-temperature within not too wide limits, and the choice 

 of a breadth of slit, to be maintained throughout the observation, 

 by which a sufficient separation of the characteristic lines shall 

 be secured without such a diminution of the light as may 

 prevent a clear perception of the spectrum. In laboratories, 

 where one prism only is commonly used in the observation of 

 spark-spectra, these conditions are most easily fulfilled by nar- 

 rowing the slit until the two characteristic bands of yttrium 

 chloride in the red are resolved into a distinctly perceptible mass 



