Variation in the Pressure of Saturated Vapours. 87 



unreliability of the data may also be remarked in calculating 

 x after the first method ; it evinces itself in the irregularity 

 of the resultant values. But it is somewhat difficult to deter- 

 mine the extent and locality of the errors : to do this it is 

 necessary to compose a table of the differences of different 

 series of data for the pressures ; if there are errors, irregu- 

 larities then appear in the second or third of the differences. 

 Such are, for example, the vapour-pressure of ether at 40°, 

 907*04 (Regnault, vol. xxvi. p. 393) ; of bisulphide of carbon 

 at 95°, 2966-34 (p. 402) ; of propionic acid at 120°, 389*0, and 

 at 135°, 650*2 (Landolt's Tables, p. 55). All these data are 

 erroneous, but I have not been able to determine the extent 

 of the error. The vapour-pressure of methyl alcohol at 80°, 

 given as 123£*47 (Regnault, vol. xxvi. p. 460), should be 

 1235*47. The vapour-pressure of acetone at 50°, 602*86 

 (p. 472), should be 60S*86 ; of mercury at 380° (p. 521), 

 113p-65 should be 1136*65 ; at 260° it should be 96*£s> and 

 not 96*75. The value 5072*43 at 480° is also very doubtful. 

 In Landolt's Tables (p. 57) the vapour-pressure of benzene, 

 given as 5685*03 should be 578p*03. This misprint is, how- 

 ever, mentioned by Regnault, but is inserted in his tables 

 without correction. Lastly, in Regnault's tables (vol. xxi.) 

 of the vapour-pressure of steam, there are several errors which 

 have been reprinted without correction in many works. 



Conclusion. — The pressure of a vapour in a state of satura- 

 tion is expressed by equation (5). The values of c— c ± and of 

 r Q may be taken from experimental data, but are better calcu- 

 lated, looking upon them as arbitrary constants, from tables 

 of vapour-pressures, because the latter can be determined from 

 experiment with far greater accuracy than the former. How- 

 ever, the values obtained for r by either method differ but 

 slightly, with the exception of those cases when considerable 

 errors may be suspected in the values of the vapour-pressures 

 or r determined by calorimetric measurements. The values 

 obtained for c — c 1 differ far more. This is due, as has been 

 already explained, to the fact that even slight errors in c and 

 Ci produce an immense effect upon the value of c — C\. If we 

 carefully follow out Regnault' s method for measuring c : , we 

 inevitably come to the conclusion that the specific heat of a 

 vapour may contain a very large error, because it is calculated 

 from the difference between two quantities of heat which are 

 exceedingly great compared to the magnitude of their differ- 

 ence. Consequently, the error in c — c x may be so great that 



