1891.] Lines in the Spark Spectra of the Elements. 451 



tThe nimbus is evidently an expression of the quantity of matter 

 the spark, and the intensity of the chemical action which the rays 

 emitted by its ignited vapour are capable of exerting. 



(3.) On the Continuous Spectrum which forms the Background to the 

 Lines of certain Spectra. This must be caused by the ignition either 

 of some solid substance or of a vapour which is not that of an 

 element but an oxide. An examination of the spectra in which the 

 continuous background of rays is a conspicuous feature discloses the 

 fact that the metals which are not oxidisable do not possess it, for 

 instance, gold, silver, and platinum. Metals of the iron group show 

 it near the points of the electrodes when the non-volatile oxides are 

 formed. The very volatile metals with volatile oxides, such as 

 mercury, iridium, thallium, zinc, and cadmium, do not show it. 



I Spectra of the metalloids, such as tellurium, arsenic, antimony, and 

 bismuth, which are not only volatile but which form volatile oxides, 

 show it very strongly. Ordinarily, magnesium does not show it, 

 because the exposure necessary for photographing the spectrum of 

 that element is less by one-half the period of the others, and by one- 

 quarter that of tellurium. When a plate is long exposed to the rays 

 of magnesium, the continuous spectrum appears at the points of the 

 electrodes where the non-volatile oxide would be formed. It may 

 be considered that in the passage of the spark, the vapour of the 

 element fills the track, and this vapour, on cooling, forms, for a 

 minute period of time, an incandescent oxide, and, the spectrum of 

 this being a, continuous spectrum, its photograph appears as a back- 

 und to the rays emitted by the element. 



But it is nevertheless the fact that the continuous background 

 a very characteristic feature of the metalloids, though why the 

 pours of these oxides should produce this action more conspicu- 

 ously than those of the oxides of the volatile metals, there seems to 

 no sufficient or well-understood reason to be advanced at present, 

 may be that the vapours of the metalloids in cooling emit a con- 

 ons spectrum for a short period prior to oxidation. 

 On the Breadth of Lines. It is well known that, under identical 

 nditions, the principal lines in the spectrum of an element become 

 >nger and broader as the rays forming the spectrum proceed from 

 larger quantity of material, that is to say, form a denser radiating 

 ,yer. It is evident, then, that in any series of three or more 

 ments of similar character, the intensity and the breadth of the 

 es in their spectra will depend upon (1) intensity of chemical 

 ergy, (2) volatility and vapour density, and (3) electric conduc- 

 vity of the metal. 



In accordance with these conditions, the lines of cadmium are 

 >ader than those of zinc, aud the lines of zin ; broader than those 

 magnesium. 



