Atomic Laws of Thermochemistry. 43 



in the acids, 



/(C:0)+/(O0 acid) +/(0H) = 115 ; (39a) 



in the esters, 



/(C : 0) +/(C ' -acid C) = 105; . . (40) 



in the anhydrides, 



2/(0: 0) +/(0-0- acid C) = 165-9; . (41) 

 in the carbonates, 



/(C:0)+2/(G'0-acidC) =162-1. . . (42) 



It can be seen that the reason for the occurrence of the 

 three mean values given above for /(C:0) is, that in the 

 equations for the aldehydes and ketones the term /(C'O acid) 

 occurs no times, in the equations for the acids, esters, and 

 anhydrides occurs once, and in the equation for the car- 

 bonates twice ; so that if, instead of taking /(C'O acid) and 

 /(CO- acid C) to be the same as/(O0) and/(OOC) in the 

 alcohols and ethers, as we did in solving the equations (34) to 

 (39), we assume/(C-Oacid)=/(C-0) + 15 and /(C'O' acid C) 

 =/(C*0'C) + 15, then the three values found above for/(C:0) 

 would become 52 in the aldehydes and ketones, 53 in the acids, 

 esters, and anhydrides, and 58 in the carbonates. 



This difference between /(C'O) and /(C'O acid) appears as 

 the first pronounced exception to our fundamental hypothesis 

 that the heat evolved on the junction of two atoms depends 

 only on the two atoms, and not on the other atoms with which 

 either may be associated. In the case of the junction denoted 

 by (C'O acid) the most characteristic feature is that the C 

 atom is joined by two bonds to another atom — that is, 

 /(C'O acid) is always associated with /(C:0) ; and this asso- 

 ciation produces an increase of thermal effect amounting 

 to 15. Thus we have a case of atoms affecting one another 

 without being directly connected ; and it is interesting to find 

 this phenomenon occurring in connexion with the profound 

 distinction that exists between the hydroxyl of acids and the 

 hydroxyl of alcohols. 



We can see that the difference between the 50*4 for/(C:0) 

 in the aldehydes and the 54*2 in the ketones is not sufficient 

 to warrant the peculiar change of constitution suggested by 

 Thomson for the aldehydes. 



Returning to the ^value for /(C:0), which we have found 

 to be about 53, let us consider again our early assumption (7a) 

 that /(C0 2 ) = 2/(00) = 135-3. Now/(C0 2 ) = 2/(C: 7 0) ; mark- 

 ing the tetrad character of the carbon ? while in carbonic oxide 



