E. W. Morley — Volumetric Composition of Water. 231 



gas from the store in n. At p are two drying tubes intended 

 simply to prevent aqueous vapor produced at r from entering 

 o or s, and interfering with their action. At s is a Sprengel 

 pump. The mercury which operates it first falls through the 

 tube t, kept exhausted by a Toepler pump fused to u This 

 trap is made long as shown so that even when the gas in j?, q, 

 r is at atmospheric pressure, mercury falling into t has still to 

 pass through many centimeters of vacuum. At v is the jar in 

 which the gas is transferred to the measuring apparatus. It is 

 plain that, disregarding for the moment the apparatus r, we 

 can, on opening the valve o, transfer gas from the store in n to 

 the jar at v. But the apparatus at r has sometimes to perform 

 an office which has made it possible to obtain from impure 

 gases results as accurate as from pure gases. It will be 

 recalled that Humboldt and Gay-Lussac obtained, although 

 they worked with very impure gases, results whose mean 

 errors are comparable with those of Scott, whose gases were 

 far purer. This was owing to the fact that they devised 

 means for measuring the impurities of one gas by some 

 process other than the eudiometric explosions which determine 

 the total nitrogen in both gases, and also the volume of the gas 

 which has disappeared in the explosion. If Scott had suc- 

 ceeded in refining on the methods of Humboldt and Gay- 

 Lussac in this respect as much as in his methods of preparing 

 and measuring the gases, he might well have attained a degree 

 of accuracy satisfactory to himself and answering the present 

 demands of science. The apparatus r was designed to make 

 possible the determination of the nitrogen contained in the 

 hydrogen used. It consists of a hard glass tube filled with 

 copper oxide, and united to the rest of the apparatus by means 

 of a mercurial seal. The apparatus to the right of n was dupli- 

 cated for oxygen, except that the tube r contained copper in- 

 stead of the oxide, and no drying tubes pp were needed. 



To get copper oxide which could be heated in a vacuum 

 without giving off gas cost much trouble. Electrotype copper 

 was oxidized in a current of oxygen ; but when it was heated 

 in a vacuum, it gave off a gas for a long time. In some cases, 

 heating for several days answered the purpose ; but some 

 samples had to be rejected and the heating recommenced with 

 a new sample. The gas given off was mostly absorbed by 

 potassium hydrate, but as yet has not been further studied. 



The mano-barometer m served partly in preparing for weigh- 

 ing the gas in the globe <?, and partly in preparing for deter- 

 mining the amount of nitrogen in the hydrogen by means of the 

 apparatus r. For the last purpose, an accuracy permitting 

 readings to a thousandth of an atmosphere was sufficient ; the 

 apparatus will therefore not be described at present. 



[To be continued.] 



