Manchester Mei>ioirs, Vol. liv. (1910), No. 1. 11 



That Berthollet felt the weight of this refutation of his 

 teaching is shown by the fact that he thought it necessary 

 to repeat Wollaston's experiments." 



Further, Gay-Lussac's Memoir on the combining 

 volumes of gases, published in 1809, afforded numerous 

 examples amongst gases of combination in fixed propor- 

 tion. Berthollet, however, had declared that this was 

 likely to occur amongst gases, a fact which greatly dis- 

 counted the possible effect on chemists of Gay-Lussac's 

 discoveiy — of tending to lessen their confidence in Ber- 

 thollet's doctrine of variable proportion. Besides, at that 

 stage of chemistry, Gay-Lussac was himself reluctant to 

 abandon this doctrine, and still held that, in general, 

 mass-action must produce compounds in all proportions. 

 He maintains the "great chemical law, that whenever two 

 substances are in presence of one another, they act in 

 their sphere of activity according to their masses, and 

 give rise in general to compounds with very variable 

 proportions, unless these proportions are determined by 

 special circumstances." " 



Indeed, Dalton's doctrine of combination in definite 

 and multiple proportions was victorious only in process of 

 time and in conseciuence of the efforts of J. J. Berzelius. 

 Yet it is worth noticing how much less complete Proust's 

 answer to Berthollet was, than the answer tacitly con- 

 veyed by Dalton's doctrine. Berthollet held that affinity 

 tends to combine elements in all proportions, and that 

 the composition of the oxides of a metal at maximum and 

 minimum depended on accidental factors, physical con- 

 ditions opposed to affinity, such as cohesion and elasticity. 

 Dalton showed that a beautifully simple relation exists 

 between the composition of one oxide and another, so 



"» Mem. a'Aicetiil, vol. 2, p. 470, 1809. 

 "2 op. cit., pp. 232—233. 



