different Liquids at Equal Pressiires. 511 



This is identical with the equation given by Dr. Ramsay 

 and myself (Phil. Mag. Jan. 1886, p. 33) for certain closely 

 related substances such as ethyl chloride and bromide orchloro- 

 benzene and bromobenzene, and I have shown that it holds 

 good for the four halogen derivatives of benzene. 



For bodies which are not closely related we proposed an 

 equation which may be written 



R=Ro + c£, 



where R is the ratio of the absolute temperatures of the two 

 substances at any pressure p, R is the ratio at such a pressure 

 that the temperature of the second liquid =0° C, and t is 

 the Centigrade temperature of the same liquid at the pressure p. 



The second terms of the two formulae (B in that of M. Colot 

 and ct in the one just mentioned) are not equivalent, but they 

 both serve to bring the calculated results pretty well within 

 the limits of experimental error, though not in all cases for 

 an indefinite range of pressure. 



I have calculated the constants for both formulae for eleven 

 pairs of liquids of which the boiling-points, at a number of 

 equal pressures, are given in Table I. 



The constants for the eleven pairs of substances are given 

 in Table II. 



It will be noticed that the constants B and c are always of 

 opposite sign, and that as regards magnitude they both fall, 

 with one exception, in the same order. 



In Table III. I have given the differences between the 

 calculated and observed temperatures for each pair of liquids 

 and for both formulae. I have in each case taken the greatest 

 available range of pressure to make the test as searching as 

 possible. 



It will be seen that when the constants Bore are very 

 small, both formulae give good results, and either of them 

 may be employed for a very wide range of pressure. For 

 the equation B J ='R -\-ct the agreement is good in the case of 

 methyl acetate and benzene, although c has a comparatively 

 high value, but when one of the two liquids is water, an 

 alcohol, or acetic acid, the differences are decidedly greater. 



The applicability of the formula t = A6 -f B seems to depend 

 more directly on the magnitude of the constant B; there is, 

 indeed, a rough proportionality between the constants and the 

 mean differences given at the foot of Table III. It is quite 

 clear that in the last five cases the relation between the 

 boiling-points is represented by a curve and not a straight 

 line, and a slight tendency to curvature is noticeable even in 

 some of the other comparisons. 



2 N 2 



