114 Dr. E. J. Mills on the Atomic Theory. 



no longer dealing with fact but with a crude conjecture ; and 

 the well-known rules given by Dalton for ascertaining the com- 

 plexity of a combination, and intended to develope his system of 

 ratios, are more conjectural still : — If only one combination of 

 two elements exist, it must be presumed to be binary ; when 

 three combinations are obtained, we may expect one to be a bi- 

 nary, and the other two ternary, and so on*. Berzelius himself 

 felt enamoured of the theory, and occasionally turned aside from 

 the path of induction to guesswork of a similar kind : — Strong 

 bases consist of a radical taken once, combined with oxygen taken 

 oncef; and, "if a combustible radical J unite with oxygen in 

 several proportions, these are compared, and the result is reduced 

 to the smallest number of atoms possible." With the introduction 

 of the word equivalent by Wollaston, the symbolic unit acquired a 

 new implication; and the historical reader is familiar with the con- 

 fusion produced and maintained for very many years by this inser- 

 tion of a peculiarly dynamical idea in the ungenial soil of contempo- 

 rary speculation. Writers of text-books failed to understand their 

 position ; and the science was still struggling underground when 

 the splendfd discovery of Dulong and Petit gave it vigour and fo- 

 liage. This discovery, indeed, was left to other hands to improve 

 or perfect; but by the distinct relation it evinced between the sym- 

 bolic standard and specific heat, decision and certainty were able 

 to displace conjecture. Another important step was to adopt a 

 uniform understanding as to the relation of composition to specific 

 gravity of vapour, this understanding having been based upon 

 formulae which had been determined by the previously existing 

 canons. The indications afforded by isomorphism are of minor 

 value. But throughout these various processes — establishing 

 by analytical comparison certain standard units and ratios, dis- 

 covering that certain of these were connected with the specific 

 heat by a simple law, that to these a uniform vapour-volume 

 may be assigned, and that similarity of chemical function fre- 

 quently involves similarity of form (another function) — where 

 does indivisibility suggest itself? Still less is the existence of 

 an atom proved. 



I now pass from the singular relations of formulae to consider 

 them as existing in equations. Let it be required, for instance, 

 to find the formula of benzol. The values C = 12, H = l are 

 given. The analytical result, in terms of these values, is C H. 

 Recourse is then had to chlorination, which occurs, or may be 

 taken to occur, in successive stages ; the product at each stage 

 is analyzed (Cl = 35*5), and the corresponding minimum equa- 

 tions are written thus : — 



* New System, p. 214. 



f Traite de Chimie [Esslinger], vol. iv. p. 604. 



% Theorie des Proportions Chimiques (1819), p. 11?. 



