RICHARDS AND BAXTER. — ATOMIC WEIGHT OF COBALT. 67 



free from higher oxides, was now heated in a hard glass tube sealed at 

 one end ; and the tube was exhausted to within two tenths of a milli- 

 meter of mercury by means of the Sprengel pump. When the pump was 

 stopped the tension of the gas slowly increased from four tenths of a 

 millimeter at 400° to one and five tenths millimeters at full redness, the 

 highest temperature obtainable with a Bunsen lamp. The pressure of 

 the oxygen resulting from the dissociation of the black oxide of cobalt 

 was determined to be in the neighborhood of six hundred and thirty 

 millimeters at the same temperature. Evidently, if the tension of the 

 oxygen remains below six hundred and thirty millimeters of mercury, the 

 higher oxides of cobalt cannot exist (except in an imprisoned state) ; 

 and, on the other hand, if the tension of the oxygen does not fall below 

 one and five tenths millimeters there can be no reduction of the mon- 

 oxide. In the next analysis the oxygen was removed from the tube as 

 fast as it was evolved until the tension of the gas remaining was about 

 one and five tenths millimeters. The pressure inside the tube was kept 

 at this point until heating ceased, when the tube was exhausted as com- 

 pletely as possible. In subsequent ignitions of this boat-load the tube 

 necessarily contained air, not oxygen, and allowance was made for this 

 fact by increasing the pressure five times. In this case also constant 

 weight was obtained only after many ignitions, although the differences in 

 the weight of the oxide were small. 7.74242 grains (in vacuum) of 

 cobaltous oxide yielded G. 09219 grams of cobalt, corresponding to an 

 atomic weight 59.0G8. 



Evidently even here traces of metal had been found during the heating 

 of the monoxide. This undesired reduction could have taken place only 

 during the cooling, for at all other times the tension of the oxygen was 

 greater than the dissociation pressure of the monoxide. Measures were 

 then taken to make this error impossible. 



On account of the size of both tube and furnace, the cooling was 

 necessarily very slow ; this fact is indeed an advantage, because other- 

 wise the life of a tube would be short. Assuming that the tube cooled 

 always at the same rate, the tension of the oxygen from the dissociated 

 oxide after any given interval of time from the cessation of ignition 

 should always be the same. A table embodying such tensions was pre- 

 pared by repeatedly exhausting the tube to a tension slightly below the 

 supposed point, and allowing the pressure inside the tube to reach a 

 maximum. This table, which was obtained by averaging three different 

 series of observations, is given below. The readings were obtained by 

 means of a McLeod gauge. 



