394 Mr. A. LI. Hughes on the 



by thin quartz is X 1450, while some specially selected 

 specimens of fluorite transmit as far as X 1230. 



It was thought that if the ultra-violet light could be pro- 

 duced in the same vacuum as the plate, much shorter wave- 

 lengths might be obtained, since there would be nothing to 

 absorb the light. They would, assuming Ladenburg's law, 

 show their presence by producing electrons of greater speed 

 than have hitherto been obtained from surfaces illuminated 

 by ultra-violet light. If we take the velocities of electrons 

 emitted from a surface as an indication of the type of the 

 radiation falling upon it, we see that there is a large gap 

 between the softest X-rays and ultra-violet light of the 

 shortest wave-length known. The potential required to stop 

 the electrons produced by the former radiation is of the 

 order of 1000 volts, by the latter it is of the order of only 

 3 volts. All attempts to investigate this gap, either from 

 the X-ray side or from the ultra-violet side, appear to be 

 frustrated by the opacity of all known materials. 



The condition that the illuminated plate must be in a very 

 good vacuum and also in the same apparatus as the source 

 of light, practically restricts the choice of the source of light 

 to the mercury arc. The pressure of mercury vapour in the 

 arc may rise to several centimetres but falls off very rapidly 

 with the distance from the arc. The apparatus used was so 

 designed that the pressure of mercury vapour in the neigh- 

 bourhood of the illuminated plate was quite inappreciable. 

 A preliminary experiment showed that, after the are had 

 been produced several times, there was no further evolution 

 of gas from the mercury or the w r alls of the tube. 



The apparatus used is shown in fig. 1. The arc is struck 

 bv raising one of the two barometer columns at A. The 

 potential used is about 50 volts, and the current through 

 the arc is kept at the same value (2-4 amps.) throughout a 

 series of experiments. At a distance of 90 cms. from the 

 arc is the illuminated plate B. The velocity of the electrons 

 emitted by the plate B is measured by observing the (negative) 

 potential required by the surrounding case D to stop the 

 leak. The plate B and the case D are covered with soot by 

 means of an acetylene flame in order to reduce the reflexion 

 of light to a minimum. When the case D is at different 

 positive potentials, the leak increases with the potential 

 except when the vacuum is very high. Both Ladenburg and 

 Hull found that with these slow-moving electrons it was 

 necessary to obtain a very good vacuum, otherwise the leak 

 increased with the potential, indicating that ionization bv 

 collision was taking place, 



