DR A. CRUM BROWN ON ISOMERIC COMPOUNDS. 713 



while that of Friedel's alcohol is 



(H) © 



(^ © d) 

 © 



As Friedel's alcohol is identical with that obtained by Berthelot from propylene, 

 it is highly probable that the same difference exists in the case of all the other 

 members of the two series. In the same way the iodide of propyl is 



(h) (a) (H) 

 ©-©-©-©-© 



and the hydriodate of propylene 



(h) 



©-(^-©--(^-© 



(b d) © 



14. As to the aromatic alcohols and the isomeric acids, we know too little of 

 their constitution to speak definitely with regard to their relation. They may, 

 as suggested by Erlbnmeyer (Zeitschrift, vii. 12), be absolutely isomeric, or, on 

 the other hand, they may be related to one another as the two series of alcohols 

 last mentioned are. 



Having enumerated the bodies which may without hesitation be called 

 absolutely isomeric, I shall now consider the bearing which the existence of such 

 substances has upon the theory of atomicity. 



If we examine the fundamental definitions of that theory, we shall see that 

 there is a point of importance left undecided. We define a multequivalent atom 

 as an atom having two or more equivalents, by means of which it may unite with 

 the equivalents of other atoms, but it is not decided whether these equivalents are 

 similar to one another or not. On the former supposition, there can be only one 

 substance corresponding to each constitutional formula, and absolutely isomeric 

 compounds are impossible. It must therefore be rejected, as such compounds 

 exist. We must then assume that some of the equivalents of at least some 

 multequivalent atoms are different from other equivalents of the same atoms. 



This assumption may take one of two forms, — 1. We may suppose that the 

 difference is an essential and unchangeable one ; that, for instance, the two equi- 



