Discharge in Rarefied Gases. 245 



mator-axis. Now, if, the electric rays consist of gas-molecules 

 whose motion in the direction of the axis constitutes the 

 propagation of electricity, and if the electricity in the ray 

 propagates itself with the velocity c, the wave-length of 

 the rays of light of the kathode-light emitted by a must 

 appear smaller than the wave-length of the rays of light 

 belonging to b, in the proportion of 40,000 to 40,000 +c 

 miles. The lines of the spectrum of a, in comparison with 

 the corresponding maxima of b, must therefore appear displaced 

 towards the violet end of the spectrum ; or, if we suppose that 

 the light from a contains some molecules moving more rapidly 

 or less rapidly than the rest, then at least the lines of the 

 spectrum belonging to a must show an expansion towards 

 the blue. The observations were made on the bluish-green 

 line of the hydrogen-spectrum (F in the solar spectrum) ; and 

 the result was that, on interchanging the electrodes, there 

 was neither displacement nor broadening of the lines large 

 enough to be certainly observed ; or, more exactly, there was 

 neither a displacement nor a broadening of Hp (the F line) 

 amounting to a third of the distance between the two sodium 

 lines in the same apparatus. I was satisfied with this (for my 

 purpose) sufficient result, without making further measure- 

 ments of the dispersion of the apparatus. It indicates that 

 the velocity of displacement of the gas-molecules does not 

 amount to 14 miles per second (a more accurate knowledge 

 of the relative dispersion in the different regions of the 

 spectrum would, undoubtedly, considerably reduce this value) ; 

 the dispersion in the neighbourhood of F would be at least 

 1^ times as great as in the neighbourhood of D ; the velocity 

 of the gas-molecules may therefore be taken as less than 10 

 miles. 



Wheatstone observed the image of a vacuum-tube nearly 

 2 metres long, in a revolving mirror whose axis of rotation 

 was parallel to the tube, and observed whether the reflected 

 image of the tube placed itself obliquely to the axis of rota- 

 tion. Such a result would have led to a conclusion as to the 

 time required by the discharge to pass from one end of the 

 tube to the other. The tube remained parallel to the axis of 

 rotation, even with a velocity of 800 revolutions per second ; 

 and by determining the limit at which an oblique position of 

 the tube would have been noticed, the experiment gives as 

 the velocity of propagation of the electricity through a vacuum 

 at any rate a large multiple of 10 miles. 



I am at present occupied with the preparation of experi- 

 ments similar to those of Wheatstone's, and hope to be able 

 to make independent experiments in this direction. In the 



