High Frequency Spectra. 201 



minated by lighting the filament, but for the Gundelach tube, 

 light from a lamp was focussed on the target. The focal spot 

 was viewed through a horizontal telescope, placed at the same 

 •height above the table as the centre of the crystal face, and the 

 necessary adjustments for the central ray were easily carried 

 out. 



In the experiments described below it was necessary to adjust 

 the focal spot of the Coolidge tube when it was not possible to 

 use a telescope. In this case a mirror was used, and the images 

 of the focus, slit and crystal face were viewed from above. 



Experiment. 



Tube. — The target of the Gundelach tube consisted of a plati- 

 num button attached to a stout copper rod, on the external end 

 of which was a brass radiator. The potential difference applied 

 to the tube was produced by a Snook-Victor high tension rectifier. 



At the beginning of the research, the tube was too soft to excite 

 the K series of platinum. The maximum pressure that could 

 be applied to the tube was 60,000. volt. If the switches on the 

 high tension rectifier were turned to increase the pressure, the 

 only result was an increase in the current passing through the 

 tube. It was found, however, that by keeping the current low 

 — less than O.S.m.a. — that the tube gradually hardened. After 

 a fortnight's running, for several hours a day, a pressure of 

 85,000. volt could be applied to the tube. It was found diffi- 

 cult to maintain this potential difference, because fluctuations 

 in the current became very big, and, with the increase in cur- 

 rent there was the consequent drop in pressure. 



After a considerable amount of experiment the following pro- 

 cedure was adopted. The tube was allowed to harden, so that 

 there was practically no current, when a pressure of 95,000. 

 volt was applied. Then, to begin an experiment, the tube was 

 slightly softened, so that a current of 0.6. m.a. was obtained 

 at 85,000. volt. If the current became less, the pressure was 

 raised, and usually this gave an increase in current. The usual 

 experience was, however, that the current, with the attendant 

 fluctuations, increased after about iS.min., so that the corres- 

 ponding falls in pressure were less than 80,000 volt. This was 

 probably due to the target, as it became heated, giving off occluded 

 gas. The pressure was then cut oif from the tube for between 



10 



