r^ 



-a 



\ 



cooling. One can, of course, using printed data, calculate the con- 

 traction that a given hydrocarbon would sustain under the effect of 

 pressure and cooling. But it would be exaggerated to add these two 

 contractions : at high pressures the coefficient of thermal expansion is 

 smaller than at ordinary pressure. Now on this subject the relevant 

 literature is almost silent. We 

 thought it wise to make direct ex- 

 periments and with the engineers 

 of the Société Sécheron at Geneva 

 we perfected a new method at the 

 time that I was working on the 

 bathyscaphe for the French Navy. 

 Into an iron tube with thick 

 walls (see Fig. 2) 7-87 in. long 

 and 0-787 in. in diameter, con- 

 taining a certain quantity of d 



mercury c, there was introduced -i 



a glass tube, filled with petrol b, 

 which had a small orifice at its 

 lower end. The iron tube is closed 

 at both extremities and by a 

 flexible lateral tube d it is con- 

 nected with a high-pressure pump. 

 One begins by heating the whole 

 apparatus to 86° F. : the petrol 

 expands and the excess escapes by 

 the orifice. Then the whole ap- 

 paratus is cooled down to 32° F. 

 and through the tube d oil is 

 introduced under pressure. The 

 petrol contracts and a corres- p^^^ ^^ Apparatus for measuring the 

 ponding quantity of mercury variations of volume of the petrol 

 penetrates by the orifice into the according to temperature and pressure 

 glass tube. At the moment when, 



at the same time, the pressure reaches 400 atmospheres and the tem- 

 p^ture is at 32° F. the apparatus is inverted and opened, and the 

 volume of mercury which has entered the graduated part in the glass 

 tube is measured. After a few small corrections necessitated by the 

 contraction during and expansion after the experiment of the glass 



[ 31 ] 



--C 



