﻿196 Dr. J. N. Pring and Mr. A. Parker on the 



length of this tube occupied by the gas when at 5 cm. 

 pressure was measured. The tap at P leading to the 

 Sprengel pump was turned off for a short time before each 

 reading of the pressure. The estimations of the pressure in 

 this work were of course subject to the same limitations that 

 apply in all cases where use is made of the McLeod gauge, 

 through the presence of condensible gases and surface 

 condensation on glass. 



Purification of Carbon. — The carbon used consisted of 

 amorphous retort carbon and contained about 2*5 per cent, 

 ash (oxides of iron, silicon, &c), and Q'2 per cent, of combined 

 hydrogen. The method of purification employed consisted 

 in heating for long intervals at temperatures varying from 

 1700° to 2000°, and maintaining very low pressures. In 

 this way impurities are gradually volatilized away and 

 occluded gases expelled. It has never been found possible 

 to purify carbon completely *, on account of the tenacity 

 with which it holds many elements in large quantities at 

 temperatures greatly above their points of volatilization. 

 (After use in the experiments below, the carbon was analysed 

 by combustion in oxygen and found to contain 0'0 ( J1 per 

 cent, of ash.) The graphite rod £ was occasionally heated 

 during the above treatment by causing a discharge to pass 

 between this and the heated carbon rod. Suitable conditions 

 of temperature of the rod and pressure were arranged, and 

 a potential difference of about 250 to 350 volts was applied 

 between the carbon and the graphite. When the discharge 

 was once started, a voltage of 40 to 60 was sufficient to 

 maintain a current of 20 to 30 amperes, and the graphite 

 was gradually raised to a bright white heat at the end, and 

 the inside glass cylinder (F) to the softening point. After 

 heating the carbon rod in this manner for some 20 to 

 30 hours it was found possible, at any time when the 

 apparatus was cold, to reduce the pressure in the vessel to 

 0"0001 mm. within 2 hours. On raising the temperature of 

 the rod, however, the pressure always rose immediately, and 

 amounted to about 0"01 mm. when at 1800°. This could 

 not be reduced any further by long continued heating and 

 exhausting. After the globe had become coated, with a thin 

 film of carbon, the pressures obtained when the rod was 

 heated were higher than before, and for a given temperature 

 of the rod remained constant, even after heating and pumping 

 for several days. This residual gas is probably due either 

 to carbon monoxide which would be formed by reaction 

 between the oxides of the glass and the deposited film of 

 * Cf. Pring, Trans. Chem. Soc. (1010) xcvii. p. 504. 



