736 Prof. R. W. Wood on Hydrogen Spectra 



The Secondary Spectrum "Flash." 



This is one of the most interesting of the phenomena 

 shown by the long hydrogen tube. Ir. is shown at its best 

 bj bringing the tube to the bhick stage with a current of 

 say 20 amps, in the primary of the transformer. Viewed 

 through the green filter, the central portion of the tube now 

 appears very dark. If now we open the switch, stopping 

 the discharge, we find that on closing it again there is a 

 bright flash of very brief duration (observing with the green 

 filter). In other words, the secondary spectrum is very 

 strong at the moment when the discharge starts, but fades 

 away in a small fraction of a second. The phenomenon can 

 be repeated as rapidly as we can open and close the switch, 

 and if we do this very rapidly the tube is much brighter 

 (with green filter) than if the switch is kept permanently 

 closed. The phenomenon appears only when water-vapour 

 is present. The duration of the flash was determined by 

 photographing the tube end-on through a yellow screen on a 

 revolving photographic plate, which gave a series of short 

 dashes on a circular arc. With 2 amps, in the transformer 

 ihe flash was found to last for five half cycles of the current 

 that is the first five dashes were visible after closing the 

 switch : the image was too faint to record itself after 

 the flash was over. With 8 amps, the duration w as three 

 h ilf cycles, and with 20 amps, one half cycle, with a very 

 faint image of the second. 



With hydrogen at a higher pressure (1 mm. dark- space), 

 the duration of the flash was greater, as many as 10 faint 

 dashes having been recorded with a 20-amp. current. 



The spectrum of the flash was photographed in the 3rd 

 order spectrum to see whether any peculiarities came out. 

 This was very easily accomplished by arranging a small 

 spring switch which was closed for a small fraction of a 

 second by the handle of the Gaede pump. With an exposure 

 of two hours and a half the integrated effect of 20,000 flashes 

 was recorded. 



With 3 amps, in the transformer (on closed switch) the 

 spectrum of the flasli was much under-exposed, and nothing 

 of interest was apparent in the photograph, except that the 

 Balmer lines were much weaker in comparison with the 2nd 

 spectrum, than in the case of continuous operation with the 

 same current strength. The flashes were white in colour 

 and only H a and Hp could be seen with the prism. Con- 

 tinuous operation with the same current showed H y and H$ 

 at great intensity. This shows that on closing the circuit 



