Generalizations of Van der Waals. 1 83 



obtained agree very well with those calculated from the 

 experimental results. The other substances will be considered 

 in their order. 



Halogen Derivatives of Benzene. — Accurate determinations 

 of the volumes of liquid have been obtained only up to 280°, 

 but the ratios are very nearly constant and are practically 

 identical at corresponding temperatures and pressures ; the 

 extrapolation may therefore be considered justifiable. In the 

 case of chlorobenzene a few approximate determinations have 

 been made near the critical temperature ; they give slightly 

 higher ratios than those at lower temperatures, and agree 

 rather more closely with the mean ratio of the absolute 

 temperatures at corresponding (equal) pressures. 



The molecular volumes of the saturated vapours give 

 generally higher ratios than those of the liquid; this may 

 be partly due to experimental error; but as the deviations are 

 nearly all in the same direction, it seems hardly justifiable to 

 attribute them entirely to this cause. It is evidently impos- 

 sible to make use of these ratios in determining the critical 

 volumes ; but the mean ratios of the molecular volumes of 

 liquid may probably be relied on to give fairly accurate results 

 (Trans. Chem. Soc.'lv. p. 517). 



Carbon Tetrachloride. — As the observations extend to within 

 a few degrees of the critical temperature, there does not seem 

 to be much room for error. 



Stannic Chloride. — The molecular volumes of saturated 

 vapour cannot be made use of, as the results at the highest 

 temperatures are a little doubtful owing to slight decomposition 

 of the substance ; considerable extrapolation would also be 

 necessary. The ratios for the liquid state give, however, very 

 nearly straight lines, which may therefore probably be extra- 

 polated without much error. 



Ether. — The four curves would evidently be very close 

 together at the critical temperature, but as the results with 

 the saturated vapour are somewhat irregular it seems best to 

 rely ouly upon those with the liquid. 



Methyl Alcohol. — In this case the curves for the liquid state 

 do not seem to agree well with those for the saturated vapour; 

 I have taken the mean value, but it is evidently open to some 

 doubt. 



Ethyl and Propyl Alcohols. — The agreement is better with 

 these alcohols, but the volumes of saturated vapour are some- 

 what irregular near the critical point. 



Acetic Acid. — The four curves meet very satisfactorily at 

 the critical temperature. 



The mean ratios of the molecular volumes to those of floor- 



