68 THE ROYAL SOCIETY OF CANADA 



deduced by supposing that the calibration curve was rectilinear be- 

 3^ond 1070° C, the highest point of calibration. Platinum, platinura- 

 h-idium thermo-couples are not generally used in measuring tempera- 

 tures higher than 1100° C. or at most 1200° C. but in the present case 

 it was found that the couple still remained intact when an electro- 

 motive force of 196 X 10-^ volts was reached and this was taken from 

 the curve as representing approximately 1400° C. 



It is quite clear that with a platinum platinum-rhodium thermo- 

 couple still higher temperatures might have been recorded, but after 

 the maximum current of 10-6 amperes had been running for a short 

 time the tube cracked and the investigation was not carried fuither. 



Among other things the investigatibn shews: — 



(1) that with a moderate consumption of energy the luminous 

 vapour in the mercury arc may attain and easily exceed a tempera- 

 ture of 1400° C. and 



(2) that when the mercury arc is made to surround a silica tube such 

 as that inserted in the lamp shewn in Fig. I the temperatures at points 

 inside the silica tube are not nearly so high as that of the luminous 

 vapour surrounding it. 



The investigation suggests, too, that in all probability the temp- 

 eratures indicated by a thermo-couple when exposed directly to the 

 discharge are still very much below that corresponding to the mean 

 molecular kinetic energy of the luminous vapour. The most satis- 

 factory way, though a difficult one, to ascertain the tempera- 

 ture, would be to investigate the form and variation in width of a 

 selected spectral line when the consumption of energy in the arc is 

 varied. 



The Physical Laboratory,, 



University of Toronto, 

 May 15, 1913. 



