200 Drs. Ramsay and Young on 



It is evident that the agreement here is very satisfactory, 

 and that any differences are well within the limits of experi- 

 mental error, and of error in reading from curves. 



2. Ethyl Alcohol. — The data for ethyl alcohol are published 

 in the ' Philosophical Transactions/ 1886, part i. p. 123. 

 In constructing isothermals for the liquid, which are published 

 in plate iii. of that memoir, isobars were constructed from 

 the isothermal curves drawn through the experimental points; 

 these isobars were made use of in finding isochoric lines at 

 which the relations of temperature and pressure could be 

 compared. This method applies to volumes not greater than 

 3 cub. centim. per gram. Between volumes 3 and 12 cub. 

 centim. per gram the isochoric lines were constructed from 

 the diagram from which plate vi. was copied (the smaller 

 divisions are omitted on the printed plate). For volumes 

 above 12 cub. centim. per gram the actual observations were 

 made use of which are detailed on pp. 144-151; the slight 

 alteration in volume caused by the expansion of the measuring- 

 tube was here neglected; at the utmost the error introduced 

 is 0*15 per cent, of the total volume. 



The only satisfactory manner of comparing the found and 

 calculated pressures for liquid alcohol is to reproduce the 

 diagram showing the compressibility of the liquid, denoting 

 the calculated pressures by crosses, the actual experimental 

 observations being represented by circles. Although there is 

 apparently only one volume at which a comparison is possible 

 on each of the lower isotherms, yet it must be remembered 

 that the values of a and b are deduced from the isobaric curves, 

 to construct which numerous points on each isotherm were 

 made use of. 



Table III. on page 202 gives the values of a and b for 

 definite volumes. These were read from three overlapping 

 curves, on which values of b were abscissae, the reciprocals of 

 the volumes being ordinates ; the actual values of b in the 

 region of the critical volume were, for reasons previously given, 

 interpolated. The form of this curve is remarkable. It is 

 equivalent to the first term in Clausius's and Van der Waals's 



TJT 



formulae, -r~ — ^ : but it is evident that its form cannot be 



represented so simply. Its general form is represented in 

 the accompanying woodcut. 



The same general form is to be noticed in the case of ether. 

 The curve obtained in plotting a against volume is of similar 

 form. 



The comparison of actual and calculated results at volumes 

 below 3 cub. centim. per gram is difficult. As it is most 



