2 REPORT 1859. 



Gerhard t's classification, like every classification which rests on chemical 

 principles, was a system of rational formula. It is very important, there- 

 tore, for our present purpose, to understand clearly at the outset what his 

 formulae were intended to express. As he constantly repeated, they were 

 not attempts to represent the arrangement of the atoms of chemical com- 

 pounds, but to represent the groups or atoms, which, in the double decom- 

 positions by which compounds are formed or destroyed, replace, or are re- 

 placed by, other groups or atoms. His types were selected as being the 

 simplest or best known bodies which could be the agents or products of 

 double decompositions similar to those of the substances classified as deri- 

 ving from them. Gerhardt's formulae are, therefore, in the strictest sense 

 chemical, and, as such, ought to be clearly distinguished from formulae 

 which are intended to express the molecular arrangement of compounds, 

 formulae which, speaking strictly, are physical, not chemical. The nature 

 and importance of the distinction to which we refer will perhaps be made 

 clearer if we recall to the recollection of the Section a recent instance in 

 which it appears to have been overlooked by one of the ablest of living che- 

 mists. Gerhardt had given two different formulae for aldehyde, namely, 



P 2 H 3 1 

 C 2 H 3 O. H and „ I O, each of which expresses accurately the chemical 



nature of aldehyde in relation to a particular set of reactions. Kopp, 

 however, found that the specific gravity of aldehyde, calculated from the 

 formula C 2 H 3 O.H, according to a rule which he had deduced from the 

 examination of a considerable number of substances, agreed with the specific 

 gravity found by experiment, but that the specific gravity calculated from 



C 12 IT 3 1 



the formula „ [ O did not agree with experiment. He therefore con- 

 cluded that the first formula was more accurate than the second. Assuming 

 that the rule we have referred to was founded on a sufficient num- 

 ber of accurate observations, such a conclusion would doubtless be correct, 

 were the formula? intended as expressions of the molecular constitution of 

 aldehyde so long as it remains such, that is to say, so long as its chemical 

 characters do not come into account; but the facts in question have no bear- 

 ing on the relative accuracy of formulae which have reference solely to the 

 reactions by which aldehyde can be formed or decomposed*. 



The idea of polyatomic radicles and molecules naturally arose out of the 

 attempt to represent polybasic acids according to types of decomposition. 

 The first chemist who used formulae expressing the replacement of more 

 than one atom of hydrogen by a single atom of a compound radicle was 

 Professor Williamsonf . The views which he had expressed were extended, 

 and the expression of them in chemical formulae greatly facilitated, by the 

 introduction, by Dr. OdlingJ, of a special mode of notation. But the most 

 numerous and most remarkable examples of polyatomic compounds hitherto 

 known, have been furnished by the researches of Berthelot§ and of Wurtz||. 



In order to explain the nature of polyatomic compounds and the meaning 



of polyatomic formulae, we cannot take a better illustration than the formula 



fC 3 H 5 V" 1 

 for glycerine proposed by Wurtz^l, tt3^ \ O 3 . This formula represents 



glycerine as deriving from three atoms of water by the substitution of the in- 

 divisible triatomic radicle C 3 H 5 for three atoms of hydrogen ; that is to say, 

 as a hydrate, but a hydrate which differs from ordinary hydrates, just as 



* Comp. Kekule, Ann. Chem. Pharm. cvi. 147, note. 



t Chem. Soc. Quart. Journ. iv. 350. $ Ibid. vii. 1. 



§ Ann. Chim. Phys. [3] xli. 2] 6. || Ibid. Iv. 400. 1f Ibid, xliii. 493. 



