112 



DIRECT PROPERTIES OF MOLECULES 



307 



as the atoms are considered at rest. But this conclusion 

 loses its justification if we suppose the gaseous molecule 

 to be in motion and in brisk rotation about its centroid, 

 which lies in the atom of carbon ; for the univalent atoms 

 are driven by the centrifugal force into the equatorial plane 

 in which they rotate about the carbon atom. This con- 

 ception seems to be juster than the other, at least for the 

 state of gas, though the other may, perhaps, better suit the 

 case of the liquid or solid state. 



It is not surprising to find that the benzol atom, which 

 we are accustomed to consider ring-shaped, also shows 

 itself to have a plane structure. According to Landolt 

 and Bernstein's tables, the experiments on diffusion by 

 Winkelmann 1 and those on viscosity by Puluj 2 give 

 the values 93000 (W.) and 80350 (P.) for benzol, whereas 

 the value calculated from the formula C 6 H 6 is 89300. For 

 aceton (C 3 H G 0) the same tables give 68000 from an 

 experiment by Puluj 3 on viscosity, the calculated number 

 being 68800. 



We find 4 also a tolerable agreement for the following 

 group of alcohols : 



Values of Q. 



but a less satisfactory agreement for the ethers : 



Methyl ether 

 Ether 



C 2 H ti O 

 C 4 H 10 



59000 

 98900 



43500 G 

 89700 

 80380 P 



1 Wied. Ann. 1884, xxiii. p. 203 ; 1885, xxvi. p. 105. 



2 Wien. Sitzungsber. 1878, Ixxviii. Abth. 2, p. 279. 



3 Ibid. 1878, Ixxviii. Abth. 2, p. 279. 



4 All the ' observed ' numbers in these four tables are obtained from 

 Winkelmann's experiments on diffusion in the memoir last cited except 

 those marked P and G, which are deduced respectively from Puluj 's experi. 

 ments on viscosity, or Graham's on transpiration. 



x 2 



