( 354 ) 



When we bear this in mind, many nntil now mysterious phenomena 

 will find a ready explanation. So e.g. the fact that Liveing and Dewar^) 

 saw the sodium lines strongly broadened each time when vapour 

 was vividly developed after bringing in fresh material, but saw them 

 become narrower again when the mass came to rest, although the 

 density of the vapour did not diminish. If by pumping nitrogen 

 into the evapoi'ated space the pressure was gradually increased, the 

 lines remained sharp; but if the pressure was suddenly released, they 

 were broadened. All this becomes clear as soon as one has recog- 

 nised in the lines dispersion bands, which must be broad when the 

 density of the absorbing vapour is irregular, but narrow, even with 

 dense vapour, if only the vapour is evenly spread through the space. 



Another instance. According to the investigations of Kayser and 

 RuNGE the lines, belonging to the second secondary series in the 

 spectra of magnesium, calcium, cadmium, zinc, mercury, are always 

 hazy towards the red and are sharply bordered towards the violet, 

 whereas lines, belonging to the first secondary series or to other series 

 are often distinctly more widened towards the violet. With regard 

 to the spectrum of magnesium they say:^) "Auffallend ist bei mehre- 

 ren Linien, die wir nach Roth verbreitert gefunden haben, dass sie 

 im RowiiANo'schen Atlas ganz scharf sind, und dann stets etwas 

 kleinere Wellenlange haben. So haben wir 4703,33, Rowland 4703,17 ; 

 wir 5528,75, Rowland 5528,62. Unscharfe nach Roth verleitet ja 

 leicht der Linie grössei-e WQllenlange zuzuschreiben; so gross kann 

 aber der Fehler nicht sein, denn die RowLANo'sche Ablesung liegt 

 ganz ausserhalb des Randes unserer Linie. Wir wissen daher nicht, 

 woher diese Differenz riihrt." Kayser has later ^) given an explana- 

 tion of this fact, based on a combination of reversal with asymme- 

 trical widening; but a more probable solntion is, in my opinion, 

 to regard the widened serial lines as dispersion bands. 



If we namely assume that, when we proceed from the positive 

 carbon point, which emits the brightest light, to the middle of the 

 arc, the number of the particles associated with the second secondary 

 series decreases, then rays coming from the crater and whose wave- 

 length is slightly greater than that of the said serial lines will be 

 curved so as to turn their concave side to the carbon point. Their 

 origin is erroneously supposed to be in the prolongation of their 

 final direction, so they seem to come from the arc, and one believes 



1) Liveing and Dewar, On the reversal of the lines of metallic vapours, Proc. 

 Roy. Soc. 27, p. 132-136: 28, p. 367—372 (1878—1879). 



2) Kayser und Runge, Über die Spektren der Elementc, IV, S. 13. 

 ^) Kavser. Hundbuch der Speklroskopie II, S. 366. 



