516 SCIENCE PROGRESS 



the required coefficients deduced by a " parabolic extra- 

 polation " of which no account is given. Their results were 

 as follows : 



105 • A4OT 105 • A2m IOS * Aj 



NH 3 . . . J J53I J I526 1518 1521 



S ° 2 ' ' ' { 2380 } 2384 236 ° 2379 



That their method of extrapolating gave too high results is 

 highly probable since their values of Ao are greater than those 

 deduced from their measurements by Berthelot's method and 

 given in Table VI, results which it has already been shown are 

 probably high. Moreover, their results are not sufficiently 

 exact to justify the extrapolation ; the difference between the 

 two values for A^ in the case of ammonia, is actually greater 

 than the difference between the values they adopt for A^ and 

 A££. Jacquerod and Scheuer mention one source of un- 

 certainty in the results, namely, that due to condensation of gas 

 on the inner walls of the containing vessel, an error which was 

 experimentally determined and allowed for in Gray and Burt's 

 experiments on hydrogen chloride. 



The uncertainty attaching to Jacquerod and Scheuer's 

 extrapolated values may be explained by reference to fig. 2. 

 These experimenters determined the position of the three 

 points F, D and B only, with the object of finding E and A ; 

 and their method consisted in determining the relative positions 

 of F and D in one experiment with a certain mass of gas and 

 in determining the relative positions of D and B in another 

 experiment with a different mass of gas. Hence, assuming the 

 point F to be correctly placed, D may be in slight error with 

 reference to it, from the first experiment ; while B may be 

 slightly in error with reference to the (already slightly incorrect) 

 position of D, from the second experiment. There remains the 

 error incurred by assuming a parabolic relationship between pv 

 and p. From the graphical point of view, Berthelot's method 

 consists in determining the position of A from the known 

 positions of only two points (E and G or E and C, as the case 

 may be) and an assumed relationship between pv and p. 



Another method has also been used in arriving at the 

 requisite compressibility values. It is obvious that at constant 

 temperature the density of a gas which follows Boyle's Law 



