1232 
C-atoms are again entirely surrounded by other atoms or atom groups. 
But for ethylene and iso-amylene, where double bonds are found 
— so that two carbon tetrahedra adjoin along a side instead of by 
the angular points — half the fundamental value is found for C, 
i.e. 1,55. In each of the compounds H,C = CH, and ree == CHE 
there are #vo such tetrahedra, which therefore freely expose part 
of their surfaces — without shadowing atoms or atom groups — to 
the attractive (cohesive) action to the outside. For the other atoms 
of iso-amylene C remains therefore = 0, because {hese remain sur- 
rounded on all sides. (single bindings in the angular points of the 
tetrahedra). 
For acetylene there is triple binding, i.e. the tetrahedra adjoin 
each other by an entire side plane, so that now the whole central 
body is exposed to the attractive action to the outside. Accordingly 
we duly find C = 3,1 as for the above considered anorganie substances. 
For C,H, and its homologues we have 6 atoms with a double 
binding, so that here we have 6 > 1,55. But in the aliphatie sub- 
stitution groups CH, with single bindings we find again duly C=0. 
For naphthaline with 10 double bindings we have also 10> 1,55, 
and for eyelohexane with only single bindings C is again = 0. 
From the above table it appears how close the agreement is 
between the calculated and the found values (for C,H,, C,H,, 
i-C,H,, among other compounds this agreement is even perfect); 
only for C,H, and cyclohexane a discrepancy exists of 8 a 9°/,, 
probably to be attributed to inaccurately known critical data *). 
Table d., see following page. 
The agreement is again satisfactory. Only CH,F deviates in a 
similar way as for hb, which may be ascribed to inaccuracy in the 
critical data. 
In acetone the C-atom bound directly to O, just as that of the 
group COOX for the compound esters, is = 3,1 — in accordance 
with CO, CO,, CS, ete. 
1) We should be careful not to transfer in our thoughts the deviations in Wa 
(calculated and found) doubled to a itself as a standard. An error of 30/, in Ka 
would of course give rise to an error of 60/, in a; but then we should overlook that 
(RT)? occurs in the formula for a, on the other hand RT in that for Va, so 
that an error in Tx is transferred to Va wnenlarged, but doubled to a. Not the 
deviations between the values of a. but between those of Ka are therefore to be 
considered as standard of accuracy. Indeed, @ is always a product of two separate 
factors. And these separate factors must only be taken into account and are com- 
parable with the quantity 5 
