

450 Prof. W. N. Hartley. Physical Character* of the [Apr. Ifi, 



and mercnry spectra were measured on my enlargements. The prin- 

 cipal portion of the lines lying between point and point of the elec- 

 trode was 42 mm. in all spectra. The extension of the lines below wait 

 22 mm. to 25 mm., extension above, 9 mm. to 10 mm. As the exten- 

 sion is always sharp and well deBned, it is an important feature in 

 these spectra. Even concentrated solutions of the metal?, when photo- 

 graphed with graphite electrodes, exhibit this extension in their 

 principal lines. For instance a solution of beryllium chloride shows 

 a very remarkable extension above and below the points of the upper 

 and lower electrodes ; the dimensions of the principal line, wave- 

 length 3130'2, are as follows : between the points, 42 mm. ; below, 

 10'5 mm. ; above, 17'5 mm. It is at the upper or positive electrode 

 that the longest extension is observed, but at the lower or negative 

 electrode that it is strongest. In the case of the cadmium lines, the 

 extension is smaller, but strong at the side of the negative electrode, 

 and very fine and long at that of the positive.* The appearance 

 lines due to impurities or traces of metals in the spectrum of the 

 negative electrode only, I have attributed to the oscillation of the 

 spark discharge, and the fact that the negative electrode is the hotter, f 



(2.) The Nimbtis. The nimbus is not apparently dependent on the 

 volatility or the oxidisability of the vapour of the elements, though 

 these properties are connected therewith. 



By far the largest nimbus is that of magnesium ; those of cadmium 

 and mercury stand next in order ; the smallest are those of platinum, 

 gold, copper, and silver. It is thus evident that neither conductivity 

 nor vapour density controls it, for there is very little nimbus on the 

 lines of the thallium and iridium spectra; but volatility certainly 

 increases it. There is a considerable nimbus on some of the lines in 

 the spectra of arsenic, antimony, and bismuth ; also on a few lines of 

 tin and of lead. In the case of magnesium, the cause of the dense 

 and large nimbus is probably the intensity of the chemical action of 

 which the rays of the incandescent vapour are capable, together with 

 the large quantity of metal in the track of the spark, owing to its 

 volatility. 



The chemical activity of the zinc rays is less than that of the rays of 

 magnesium, but the effect of this is overbalanced by the density of the 

 vapour and the volatility of the metal being both greater ; accordingly 

 the lines of zinc have a large nimbus. The nimbus is somewhat larger 

 on the lines of cadmium than on those of zinc, the volatility and the 

 density of the vapour are both greater. 



* In a paper published in the ' Scientific Proceedings of the Royal Dublin 

 Society,' on the constitution of electric sparks, this does not appear in the litho- 

 graphed illustration, but I have carefully verified the fact by referring to the original 

 photographs. 



f Loc. cit., p. 373. 



