Atomic Laws of Thermochemistry, 41 



ment of to C plus the heat of formation of OH and of its 

 attachment to C to produce the CO OH characteristic of acids ; 

 and from formic, acetic, and propionic acids the mean value de- 

 rived for t is 120. As ^ + s = 119*6, Thomsen writes q + s = t, 

 and proceeds from this equation to deduce a heterodox structure 

 for the aldehydes ; first, he assumes the ordinary structures of 

 the ketones and acids, namely, those denoted above by their 

 general formulae, that is, that COOH in the acids is of the 

 form 0:C*OH, and CO in the ketones of the form 0:C :, 

 then it follows from the equation q + s = t that in the COH of 

 the aldehydes the and H must be united as hydroxyl HO, 

 and the Conust have two ^unsaturated bonds ; thus in place of 

 the usual form 0:C*H Thomsen declares for HO*C:. 

 Before . discussing these results in our notation we may as 

 well take the related results for the esters. 



6. Esters. — A difficulty arises in connexion with the values 

 of p + ad Tor the esters, because for the three isomers, propyl 

 formate, methyl propionate, and ethyl acetate, Thomsen 

 obtains the three values 253*7, 258*5, and 265*9, although to 

 be in harmony with the general principles holding for most of 

 the carbon compounds these ought to be more nearly equal. 

 As Thomsen took special pains to check the value for ethyl 

 acetate he considered that as the value for propyl formate 

 takes its right place in the series for the esters, the dis- 

 crepant value for ethyl acetate points to the fact that ethyl 

 acetate has a structure quite different from that usually 

 assigned, namely, that it contains the hydroxyl group, in short 

 that it is CH 3 *CH(OH)'CO*CH 3 instead of CH 3 *OOOC 2 H 5 . 

 A piece of heterodoxy like this has naturally made many 

 chemists dubious about the value of all Thomsen' s analysis of 

 thermochemical data. 



From the data for methyl formate, methyl acetate, ethyl 

 formate, propyl formate, and methyl isobutyrate it appears 

 that the fixing of the two atoms of in the molecule is 

 accompanied by the evolution of 105*4 kcal. For dimethyl 

 carbonate and diethyl carbonate the values of p + adare 251*5 

 and 341*2, whence, for the fixing of the three atoms of in 

 their molecules, we find a mean evolution of 162*1 kcal. We 

 can now state and discuss altogether in our notation these 

 relations for the aldehydes, ketones, acids, anhydrides, and 

 esters. But in the case of the acids we must first introduce 

 a correction for the polymerization that has been proved to be 

 associated with their abnormal vapour densities. Accord- 

 ing to the measurements of Berthelot and Ogier (Ann. de 

 Ch, et de PL 5th ser. xxx.) the heat required to dissociate a 



