292 Emission and Absorption Lines in a Gas-Spectrum. 



considerations do not cover any case of the kind here appa- 

 rently presented : the action of a surrounding optically 

 inactive gas was not considered at all in our theory, and 

 even if it had been adopted it would have given no effect. 

 According to the explanation given by us, we require an 

 increase in the density of the incandescent vapour in the arc 

 to give the effects under discussion. 



It will, however, be noticed how well our results agree 

 with those obtained by Humphreys. Our shift of the lines 

 is equal in order of magnitude to that investigated by him. 

 His conclusions given in paragraphs 2, 3, and 4 are exactly 

 fulfilled by our theoretical shift, and we have found evidence 

 that the conclusion drawn in paragraph 5 may also be ful- 

 filled. It suits the case so well that I was induced to try 

 and connect the two. The question now is: can an increase 

 in the density of the incandescent vapour take place when 

 the pressure in the surrounding air is increased ? Certain 

 evidence can be brought forward to answer this question in 

 the affirmative. 



In some experimental work on this subject Haschek and 

 Mache* considered that they had obtained a demonstration 

 of a high pressure in the path of the metallic arc, which may 

 be considered as due to the metallic vapour thrown off from 

 the electrodes. A further result, and the important one 

 from our point of view, was the dependence of the pressure 

 upon the pressure of the surrounding gas. The experiments 

 were, however, conducted with pressures below that of the 

 ordinary atmosphere (550 mm. to 96 mm.). Numerically the 

 results obtained were rather irregular, but they indicate 

 decisively that the pressure in the spark increases with that 

 in the surrounding gas. The temperature of the arc being 

 considered as constant, this means that an increase in the 

 density of the metallic vapour is brought about by increasing 

 the pressure of the surrounding gas. 



Further and more direct evidence can also be obtained. 

 For instance, in Humphreys' experiments it was found that 

 the electrodes burn away much faster as the pressure is 

 raised, and thus there is a greater rate of supply to the arc 

 of the material to which the lines are due. This is equivalent 

 to saying that the density of the metallic vapour in the arc 

 is greater, or the temperature being practically constant, the 

 pressure is greater. Humphreys accounts in this way for 

 the increased intensity of the lines as described in con- 

 clusion 1. Again, Humphreys found that reversals are much 

 more pronounced and frequent under heavy than under light 

 * A stroph vsical Journal, 1879. 



