﻿Ionization produced by Carbon at High Temperatures. 195 



follow closely the platinum scale. The small amount of 

 absorption through the glass vessel was allowed for in all 

 cases where direct temperature readings were taken. This 

 absorption was estimated by heating a strip of platinum 

 with a constant current and reading the temperature 

 optically, firstly by sighting directly on the surface, and 

 secondly by allowing the light to pass through the globe. 

 In some of the latter measurements, after the resistance of 

 the carbon had become constant through graphitization, the 

 relation between temperature and current through the rod 

 was found and made use of for the temperature estimations 

 when the vessel had, after prolonged use, become coated 

 with a rather opaque film of carbon. The accuracy of the 

 temperature readings can probably be relied upon to within 

 35° at 1500° on the platinum scale and 75° at 2000° on the 

 optical scale. 



For the anode, a graphite rod (E) was soldered into the 

 water-cooled brass tube (T), the top being placed at 4*8 cm. 

 distance from the carbon (A). During the experiment, 

 a large difference of potential (usually 330 volts) was 

 maintained between E and A; the high-potential circuit 

 being connected through the galvanometers M and M/ 

 which were of different ranges, one giving a deflexion of one 

 scale-division for 4*1 x 10~ 9 amp. and the other of one 

 division for 6xl0~ 7 amp. For higher currents these were 

 replaced by a milliammeter reading to 0*5 amp. and for 

 measuring currents during discharge an ammeter reading to 

 30 amps, was used. By means of the two-way switch S 

 the galvanometers were thrown out of circuit except during 

 the moment when readings were being taken. 



Exhaustion of Apparatus. — Jn the early experiments use 

 was made of lithium for reducing the pressure. For this 

 purpose the metal was placed in an iron boat contained in a 

 tube of Jena glass (L). By surrounding this tube by an 

 electrical furnace, and nsing a thermo-junction pyrometer, 

 it was found that beginning with low pressures absorption 

 of the gas to within one per cent, of the volume took place at 

 450°, though at this temperature the lithium very slowly 

 sublimed and gradually reacted with, and finally cracked, 

 the glass. The exhaustion was afterwards effected with an 

 automatic JSprengel pump connected to P, which would 

 reduce the pressure to 0*0001 mm. when the apparatus was 

 cold. The Topler pump K (not drawn to scale), the volume 

 of which was \ that of the whole apparatus, was also used 

 for evacuating, and as a McLeod gauge for measuring the 

 pressure. For this latter purpose the volume of gas in K 

 was compressed in the capillary, of known volume, and the 



