Law of Molecular Attraction. 491 



Third. — I further found * that, if Crompton's equation 



dP 

 was true, the -^ at the critical temperature was equal to 



2R/V C . I give the corresponding values at the critical 

 temperature of the constant so obtained above. This relation 

 I later fcund had earlier been noted as an empirical fact by 

 Dieterici f. 



So finally, excepting the values at 0° C. (30° for bromo- 

 benzene and iodobenzene) and near the critical temperature, 

 and di-isobutyl, every single divergence greater than 2 per 

 cent, from the mean value of the constant for the remaining 

 25 non-associated substances measured by Young and his 

 co-workers was found to be due to errors and removed. The 

 divergences at 0° and near the critical temperature and for 

 di-isobutyl were satisfactorily explained. 1 now have the 

 satisfaction of knowing that, within the limits above stated, 

 8 of the 25 substances show not a single divergence of the 

 constant greater than 1 per cent, from the average value, 

 and that there are only 17 divergences over 1*5 per cent, 

 out of a total of 431 comparisons, and these divergences 

 occur where known errors of observation and calculation are 

 greatest. 



I have been asked to believe that this equation is an 

 accident. My answer is that years of labour have shown that 

 where there were originally material discrepancies between 

 the equation and the measurements, these discrepancies were 

 not due to the inaccuracy of the equation. The equation 

 stands today in complete accord with the facts. 



I would remark that in not a single instance was any 

 error corrected, or any measurement smoothed, or Biot 

 formula recalculated, except by Dr. Young himself; and 

 probably when the revised values were sent me by Dr. 

 Young he has never in a single instance (except where 

 mere misprints were corrected) been able to foresee whether 

 his revisions would benefit or injure the " constant " that I 

 was to obtain from their use. In revising the volumes 

 (densities) of the saturated vapour, and in recalculating some 

 of the Biot formulae, Dr. Young had other reasons than 

 those given above for undertaking the revision, and the fact 

 that I had previously in some instances suspected that such 

 revision was desirable was only incidental. Dr. Young has 

 himself collected and republished J I the complete revised 

 data that I have used. 



* Journ. Phys. Chem. vol. ix. p. 402 (1905). 



t Ann. der Physik, xii. p. 144 (1903). 



J Sci. Proc. Koy. Dublin Soc. xii. No. 31 (1910). 



