( 641 ) 



In order that the required acciirac}' may be obtained it is neces- 

 sary tliat the pressure shall not fall below 0,5 Atm. 



To determine the theoretical normal volume measurements are 

 made at t\A'o or if possible at three positions with the same quan- 

 tity of gas. Whenever the third virial coefficient C (cf Comm. N°. 71 

 June '01 and N". 74) ^) does not come into account at the pressures 

 considered and to the accin^acy required, the same value of the coef- 

 ficient B must be found by a combination of any two of the three 

 measurements. This gives directly the deviations from the la^v of 

 Boyle and hence the theoretical normal volume. 



§ 21. Tlie mLvimj apparatus. On PI. II tig. 2 the volumenometer 

 described in § 20 is slio^vn connected with other apparatus for the 

 preparation of accurate mixtures of gases and for the investigation of 

 their compressibility. 



The drawings do not require much explanation. The mixing vessel 

 F and the reservoir (r have about 2 liters capacity. 6r is provided 

 Avith a three wa} cock and is particularly useful when a numbei* of 

 mixtures are required with a small proportion of one component 

 which can be contained therein. 



The connecting tubes between i\, i\, i\ and from E to r„ are nar- 

 row so that the nncertainty of their temperature may have no influence 

 on the accuracy of the measurements, the other tubes are large enough 

 to allow free connection between the various parts and the pump 

 and to help rapid exhaustion. 



When a mixture of given composition is required, and the mercury 

 stands above the closed clamps Cl^, 6'4, C/,, C/^, the whole apparatus 

 is pumped out through )\ and is filled through i\, while )\ is shut, 

 from say the gas apparatus connected to r.-^, after being washed out 

 with this gas. 



Then i\ is shut and ?v so turned that the gas is shut up in G 

 while F is brought into connection with the pump through i\. The 

 cocks )\, i\ and }\ are then opened and E and F and the connec- 

 tions pumped out again. By raising Kr and opening CL^ widely and 

 67i (to prevent air entering from J/) only pai'tially the mercurj^ 

 is caused to enter E and to fill first Eba^ without enclosing any 

 gas, then Eij u[) to the required position, when Cl-^ is shut. The 

 space between }\, )\, i\, )\ is repeatedly washed out by a small 

 quantity of gas from (r through the cock i\, fi-oin here also gas is 



1) Livre jubilaire dédié a J. Bosscha ; Archives Nóerlandaises, Ser. II. T. VI. 

 p. Sli—S^S. 1901. 



