738 Prof. R. W. Wood on Hjjdroaen Spectra 



the glass is charged with water-vapour, which is only slowly 

 driven off. After some time, however, the secondary 

 spectrum makes its appearance in the central portion of 

 the tube, and the mercury lines are very likely to appear 

 as well. 



The technique which I now employ to bring the tube 

 to the white stage is an improvement on that used in 

 the earlier work. As soon as the secondary spectrum has 

 developed at the centre to an intensity nearly equal to that 

 which it has at the ends, I allow the tube to run for about 

 an hour and a half with 20 amps, in the transformer, 

 without pumping (With l he palladium tube we operate 

 with the stop-cock to the pump closed, of co >rse.) The 

 Crookes dark- space is kept at about 5 mm. during this 

 operation by lighting the hydrogen flame which plays on 

 the Fd tube for a few seconds from time to time. This 

 may not be necessary, but the pressure is apt to fall in the 

 tube, and if the green fluorescence stage is reached the tube 

 is sure to puncture. At the end of the run the luminosity 

 of the tube is usually much less than at the start. This is a 

 good sign. We now stop the discharge and exhaust io 

 a high vacuum. Fresh hydrogen is now admitted until 

 the dark-space has shrunk to 1 mm. This will require the 

 flame on the palladium tube for a minute or more. The 

 switch may be closed for an instant from time to time 

 to note the condition, but it must not be kept closed until 

 the pressure is right; at the proper pressure the tube will 

 appear rilled with striae. If the tube is ready to go into the 

 white stage these stria) will be in rapid motion, usually 

 crowding towards certain spots in the tube. Watching the 

 discharge at the centre through the direct-vision prism, we 

 fix our attention on H^. If it fades gradually but does not 

 disappear, we must stop the discharge and immediately 

 exhaust the tube, otherwise H^ will begin to brighten 

 again (H^ reaches its minimum intensity in fifteen or 

 twenty seconds after starting the discharge). We now 

 admit fresh hydrogen to 1 mm. dark-space, and again 

 turn on the current, watching Hp as before- This time 

 it may sink in intensity to the level of the lines of the 

 secondary spectrum ; and on repeating the exhaustion and 

 refilling, it will probably disappenr entirely, while H a is 

 so much reduced that it appears of a dull brick-red colour. 

 I have never succeeded in getting H a to disappear entirely. 

 It is more than likely, however, that the white stage will 

 not come at the first attempt. If the colour remains a 

 vermilion- pink after the second filling with hydrogen, the 



