Matliias and the Critical Density. 



297 



Again, all the differences at 0° on the one side of normal 

 pentane are negative, and with one exception (fluorbenzene) 

 all on the other side are positive. 



That there is a relationship o£ the values of a to the ratios 



IX 



D' 



is clearly seen by the diagram (fig. 2). 

 Fisr. 2. 



D, 



4-0 



3-9 



3-8 



3-7 



3-6 







c 



c 



° o 



« n O 







3 °o 



o 



o 



oc 



o <- 



o 

 o 



lO 



o 







c 











0-85 0-90 



0-95 1-00 



a 



1-05 MO 



With regard to the differences at 



M. Mathias notices 

 p. 19) that at low temperatures the substituted 



(loc. cit 



ammonias give values of a that increase as the temperature 



falls, but he regards this as being due to the approach to the 



solidifying-point. 



It appeared to me more probable, however, that in most 

 cases the mean densities would be more accurately repre- 

 sented by the equation 



than by the simpler one D t = T> Q + ixt, and I therefore cal- 

 culated the constants D , a, and /3 for all the substances. 

 For the alcohols, however, it was necessary to introduce a 

 fourth term, yt' 3 , to get a satisfactory agreement. 



I am indebted to Miss E. C. Fortey, B.Sc, for much 

 valuable assistance with these calculations. 



The constants are given in Table II., and here it will be 

 seen that whilst the values of a are without exception negative, 

 those of /3 on the one side of normal pentane are all positive, 

 and, with three exceptions, those on the other side are 

 negative. These exceptions are : fluorbenzene, which was 

 before found to behave abnormally ; hexane, for which /3 = ; 

 and methyl alcohol, for which y has a very high negative 

 value. 



