94 Scientific Proceedings, Royal Dublin Society. 



In 1894, Walker ('* Journ. Chem. Soc." lxv., 193) proposed the formula 

 T = aM b , where T is the absolute boiling-point, M the molecular weight, and 

 a and b are constants depending on the series ; and in 1899 E. Boggio-Lera 

 (Gazzetta, xxix., 1, 441) brought forward the formula T=a^/M+c. Each 

 author gave values of the constants for a number of homologous series ; but 

 it may be noted that when, as in most cases, the number of members of 

 a series is small, and different constants are used for each series, it should 

 not be difficult to get good agreement between the calculated and observed 

 results if accurate data are available. The normal paraffins afford an 

 excellent test of the utility of a formula, for the series is complete to Ci 9 H w , 

 and there is no fear of complications due to molecular association. 



For the normal paraffins the best results appear to be obtained by taking 

 b " 0*5 in Walker's formula and c -= in Boggio-Lera's ; the two formulae, 

 therefore, become identical, and for a the value 37'38 may be adopted. 



The calculated and observed values are given in Table I, and it will be 

 seen that while the agreement is satisfactory from C 8 H 18 to CuH 30 , or perhaps 

 C 15 H 32 , the deviations increase rapidly at lower and higher temperatures, and 

 are very large for the lowest members of the series. 



Bamage, in 1904 (" Cambridge Fhil. Soc. Froc," xii., 445), proposed a new 

 formula, T = a [M (1 - 2"")]*, in which a = 37'3775, and has therefore the same 

 value as the a in Walker's formula, and n is the number of carbon atoms in 

 the molecule. This formula gives much better results at the lower tempera- 

 tures, as will be seen from Table I, but for the highest members of the 

 series 2'" is negligible, and the formula becomes identical with Walker's. 



Last year a new formula, log0 = K(\ogM) s , was proposed by Ferguson 

 ("Phil. Mag.," xxix., 599). This formula is of the same form as Walker's, but, in 

 place of the absolute temperatures and the molecular weights, the logarithms 

 of these values are taken. For the normal paraffins, Ferguson takes 

 K= L929 and s = - 4134. The agreement between the calculated and 

 observed temperatures is generally satisfactory from C 6 Hh to C 17 H 38 , and the 

 formula gives better results than those of Walker and Boggio-Lera or 

 of Kamage, but the deviations become considerable below C 6 Hi 4 and above 

 CLL* (vide Table I). 



[Table I. 



