TRANSACTIONS OF SECTION B. 571 



The compound atoms were units which had p:rown out of the atomic theory. 

 Each of them was the smallest quantity of a compound, which (consistently with 

 the results of analysis) could be represented as built dualistically of its constituent 

 atoms. 



Chemical combination was viewed as a process of juxtaposition of simple or 

 compound atoms, little account being taken of the disturbance of the previous 

 arrangement of those compound atoms. It was when a constitution, similar to 

 that attributed to salts, was imagined for other compounds not saline in their 

 character, that the dualistic theory broke down. Thus chlorocarbonic acid was 

 represented as a compound of carbonic acid with carbonic chloride, and was 

 accordingly designated as carbonate of carbonic chloride, while the formula was 

 made to contain the formulae of those bodies. Chlorosulphuric acid and chloro- 

 chromic acid were in like manner represented as compounds of sulphuric and 

 chromic acid respectively with imaginary hexachlorides. 



Careful investigations of the reactions in which chlorocarbonic acid takes part 

 showed, however, that in each of them it behaves as a compound containing only 

 two atoms of chlorine. It was found that the commonest and best-known car- 

 bonates and sulphates have a fundamentally similar constitution. Thus potassic 

 carbonate may be represented as a compound in which the two atoms of chlorine 

 in phosgene are replaced by two atoms of the radical K ; and oil of vitriol, as a 

 compound of two atoms of hydroxyl with the same group, S O.,, which in chloro- 

 sulphuric acid is combined with two atoms of chlorine. Chlorochromic acid has 

 not been examined to as great an extent as the above compounds, but all we know 

 of it points clearly to its having a molecular constitution similar to that of chloro- 

 sulphuric acid, viz. CI,, Cr Oj, for not only do their vapour-densities agree, but the 

 chromates in their constitution and crystalline forms exhibit a clear analogy to the 

 sulphates. 



Moreover, the simpler molecular formulae, which a fuller knowledge of their 

 chemical behaviour suggested for these bodies, were found in all cases to agree with 

 the volume belonging to the molecule of every pure substance known in the state 

 of vapour. 



A difficulty of another kind had been foreseen by the great founder of the 

 dualistic system, and it was by the investigations in organic chemistry that it 

 assumed serious proportions. 



Carbon compounds were discovered possessing definite and specific properties, 

 and presenting the characteristics of pure substances, but of which the results of 

 analyt-is did not agree with any simple proportion between the numbers of their 

 constituent atoms. Their empyrical composition could not be decided by the aid 

 of the so-called law of multiple proportions, for two or more atomic formulae 

 required percentages of the constituents differing so little from one another that 

 analysis could not decide which was the true one. 



In order to select the true molecular formulae of such complex substances from 

 among those which approached most nearly to the results of ultimate analysis, and 

 to determine with certainty their empirical composition, it was necessary to find 

 other methods for the determination of molecular weights. It was necessary to 

 study the various properties of compounds of known composition, and of others 

 which could be prepared iu a state ot purity ; to determine the vapour densities 

 and rates of diffusion of those which could be obtained in the gaseous state with- 

 out decomposition ; to determine boiling points and melting points ; to examine 

 crystalline forms of pure compounds and of mixtures ; to determine solubilities 

 and densities of solids and of liquids; but above all it was necessary to collect 

 fuller and more accurate knowledge of the chemical changes which take place in 

 the mutual reaction of molecules. 



A vast amount of accurate and careful work of these kinds has been done, and 

 has been subjected to rigid and often hostile scrutiny during the various stages of 

 its progress. We now know that couipound atoms, or molecules as we call them, 

 which can be identified by their geometrical, mechanical, and other properties, are 

 the same as the compound atoms indicated by the most comprehensive chemical 

 evidences of composition and reactions. The molecular constitution of matter was 



