THE BKISTOL PEKIOD 79 



elusion that the molecules of a liquid are more complex than 

 those of the same substance in the gaseous state, and that even 

 above the critical temperature there is a mixture of the two 

 kinds of molecules. The method of heating must, however, 

 have been unsatisfactory, for the observed critical temperature 

 and pressure of benzene were far too high. (Observed 291*7 C. ; 

 60*3 to 60'5 atm., instead of 288'5C., and 47 '9 atm.) The 

 pressure does not appear to have been corrected for the deviation 

 of air (in the manometer) from Boyle's law, but even when this 

 correction is introduced the critical pressure is only lowered to 

 59 '2 to 59*4 atm, (It is possible, of course, that the benzene 

 contained some more volatile impurity which would raise the 

 critical pressure.) 



Ramsay invited me to join him in a continuation of these 

 researches. He had employed a vapour bath in his work on 

 specific volumes, the liquid in a small bulb being heated by the 

 vapour of the same substance, boiling under atmospheric pressure. 

 The method of heating in a vapour bath was found to be very 

 satisfactory and was evidently capable of wide application. We 

 decided to adopt it, making use of a series of pure liquids, as 

 already mentioned, and altering the pressure, as required, to 

 give a range of temperature for each substance. The experi- 

 mental tube could now be placed in a vertical position and the 

 reading of volume was thereby rendered easier and more accurate. 

 The method of filling the experimental tube with liquid in such a 

 manner as to expel the last traces of air was also improved, and 

 curves were drawn from Amagat's data for the correction of 

 pressure for the deviation of air from Boyle's law. 



It was perhaps unfortunate that, with the exception of ether, 

 the substances we selected for investigation were hydroxyl 

 compounds alcohols, acetic acid and, later, water for these 

 substances are now known to behave abnormally in many respects. 

 Still, the comparison of their properties with those of a large 

 number of normal substances which I afterwards examined has 



