380 Messrs. W. E. Wilson and G. F. Fitzgerald. 



as the arc was very unsteady, and when the pressure was raised it 

 was almost impossible to keep an arc going, partly because the arc 

 broke when it was elongated the least bit, and partly because a com- 

 plete lantern of soot trees grew all round the crater, and seemed to 

 short-circuit the arc from time to time. 



The arc being very unsteady, no satisfactory reading of the voltage 

 and current was possible. At from 60 to 80 Ibs. pressure the voltage 

 varied from 60 80, and the amperes kept continually varying from 

 1520. At 40 Ibs. with 20 amperes the volts varied from 5060. The 

 crater was not well developed, so that the radiation observation, even 

 at low pressures, was not very satisfactory, while at high pressures 

 the arc was too short to see into the crater at all, and the lantern of 

 soot trees hid a considerable length, 3 or 4 mm. of the negative carbon 

 besides. The radiomicrometer gave 440 divisions with a good arc in 

 air, and 380 with the moderately good crater in hydrogen. But this 

 difference is no greater than would often occur with a good and 

 moderately good crater, so that there is not any proof of a difference 

 of temperature due to cooling power of hydrogen. These experi- 

 ments showed us that it was quite hopeless to get any measures of 

 radiation under pressure with hydrogen. 



We finally tried an atmosphere of carbon dioxide. We used a 

 cylinder of liquid C0 2 , which was connected to our arc box by a 

 copper tube and stop valve. The arc burned fairly well in this gas, 

 and, except for the difficulty of getting a sufficiently long arc at 

 pressures above 150 Ibs., some pretty satisfactory measures of radiation 

 were obtained. We found that whenever the pressure was suddenly 

 reduced, there was a fog formed in the box, which cut off the light 

 enormously. Also by looking down the steel tube, which is closed at 

 its end by a lens, we could see powerful convection currents in the 

 gas which scattered a lot of light. At high pressure the refraction due 

 to these currents prevented any sort of an image of the crater being 

 formed while the pressure was varying. While the pressure was steady 

 a good image could be formed. This tube is nearly 3 ft. in length, 

 and only \ in. in bore, and it would naturally take time for the gas 

 to settle down throughout its length. We propose to have this tube 

 removed, and the aperture in the box closed by a strong piece of plain 

 glass, and to form an image of the carbons by a lens placed at a suit- 

 able distance outside. This we expect will remove the difficulty 

 arising from these convection currents. 



The result of all these experiments so far is that it would require 

 more evidence than we have been able to get, to affirm that either 

 the temperature of the crater of the arc is raised or lowered by 

 pressure. We got some very concordant observations, which showed 

 the temperature to be lowered with pressure, and in which at the time 

 we could see no evidence of absorption by fog, but then, at other 





