M. A. J. Angstrom's Optical Researches. 335 



we know, not indeed the actual velocity of the spark, but at least 

 a limit value, under which it cannot fall. He found that a spark, 

 4 inches in length, passed in a less time than the millionth part 

 of a second. But a stream of light 6 feet long, in a glass tube 

 partially exhausted, showed itself also unchanged in the rotating 

 mirror ; and as, according to Masson, the spark cannot otherwise 

 propagate itself in vacuo than by the particles from the poles 

 themselves, it would follow that at least 3 feet were passed 

 over in less than the millionth part of a second. This gives a 

 velocity which in all probability far surpasses that of the double 

 stars. 



Now, if a spark be permitted to pass in an oblique direction, 

 according to Doppler's theory the colour of the particles which 

 stream out from one of the poles must be different from that of 

 the particles which stream from the other ; the more so, as they 

 move in opposite directions, the velocity above mentioned being 

 thus doubled as regards the production of the colour. Such a 

 difference, however, is not observed. The bright lines correspond 

 to each other perfectly as before. Hence, from both practical 

 and theoretical grounds, we are justified in concluding, that time 

 of oscillation and colour are independent of the velocity of the 

 medium from which the light emanates. 



11. In the foregoing pages I have shown that the spectrum 

 of the electric spark is really to be regarded as an overlapping of 

 two spectra, one of which belongs to the metal, and the other to 

 the air through which the spark passes. From this it necessarily 

 follows, that when the spark passes through a gas different from 

 atmospheric air, the spectrum must exhibit a corresponding 

 alteration. Partly to ascertain the nature of this alteration, and 

 partly to place the fact of a superposition of an air and metal 

 spectrum beyond all doubt, I have observed the phenomena 

 exhibited by the electric spark in different gases. The gases which, 

 up to the present time, I have had an opportunity of examining 

 are, — oxygen, carbonic acid, nitric oxide, hydrogen, carburetted 

 hydrogen and nitrogen. As the results obtained appear to me 

 to be not without interest, I take the liberty of adding them to 

 the foregoing. 



The apparatus made use of in these observations is shown in 

 fig. 4. It consists of a glass tube and two brass caps, the whole 

 forming a closed cylinder, in which the electric spark passes 

 between the two spheres of brass a and a. One of the brass 

 balls can be raised or depressed by means of a screw. The tube 

 // tenrea for the introduction of the gas, which then issues through 

 a hole in the opposite end of the cylinder. To ensure that the 

 cylinder contained no atmospheric air, the gas was generally 

 permitted to stream for an hour through the vessel before the 



