6 CLARKE 



acter appears. All the computations rest upon Thomsen's data, 

 adjusted in the manner already described. In fixing the value 

 of «, actual atomic unions only are counted ; double or triple 

 carbon bonds are reckoned as one alone, the extra bonds being 

 in abeyance and thermally inoperative. In C2H2, for instance, 

 w = 3 ; in ethylene, n = <, ; and in dipropargyl n = 11. In all 

 chain molecules, therefore, n equals the number of atoms in the 

 compound less one. The fundamental fact throughout is the 

 union of an atom with an atom, the mode of combination, the 

 structure, being an entirely subordinate affair. 



To trimethylene, C3H,., if it be regarded as a ring compound 

 with nine unions in the molecule, the formula does not apply. 

 As measured by Thomsen, its heat of combustion, per gramme- 

 molecule, is 499430 calories. Adjusted to water as gas, this 

 becomes 468 151, and 4^= i 872604. Then, if n = 9, the quo- 

 tient is 14857, a quantity considerably too high. With n = 8, 

 as in the case of propylene, the value of the constant is 13985, 

 a result more in harmony with the others. According to Ber- 

 thelot and Matignon ^ the heat of combustion of trimethylene at 

 constant volume is 505 600 calories ; a determination still higher 

 than that of Thomsen, and therefore still less satisfactory. This 

 compound, then, is exceptional, and its thermal significance is 

 yet to be interpreted. As for the aromatic hydrocarbons, they 

 require a modified formula, and will receive separate considera- 

 tion later. 



Berthelot and Matignon also give data for several other of 

 the hydrocarbons studied by Thomsen, and these may well be 

 included here for sake of comparison. They all refer to gases 

 at constant volume ; and with them we may include two earlier 

 measurements by Berthelot,^ relative to methane and acetylene. 

 Taking the quadrupled heats of combustion, we now have the 

 figures given in the following table. The agreement with 

 Thomsen is fairly good, except for the last two compounds. 

 The difference between the quantities measured at constant 

 pressure and those which represent constant volume, is evi- 

 dently not large. 



'Ann. Chim. Phjs. (6), 30, 547. 



*Ann. Chim. Phjs. (5), 23, 178 and 180. 



