40 HISTORY OF CHEMISTRY. [LECTURE III. 



nomena which are denominated, according to Bergman's views, 

 by the affinities " in the wet and dry ways." Thus, for example, 

 dissolved silicates are decomposed by almost all acids, whereas, 

 on the other hand, silicic acid drives most acids out of their 

 salts at a red heat. 



But Berthollet goes still further. Cohesion determines not 

 merely the nature of the compound formed, but also the pro- 

 portions in which combination takes place. His conception of 

 a chemical compound is not associated with the constancy of 

 proportions which had been assumed prior to his time. On 

 the contrary, there exist, in his view, chemical compounds with 

 all possible proportions ; 21 and it is only special reasons, such, 

 for example, as considerable condensation on combination (that 

 is, change in the cohesion of the constituents) which occasion 

 constant proportions. Thus hydrogen only combines with 

 oxygen in a definite proportion because water, the product of 

 the combustion, is liquid, and the contraction which takes 

 place presents too great an obstacle to the production of other 

 compounds. 22 But if, on combination, there is no change in 

 cohesion, or only a slight change, then compounds are formed 

 with variable proportions. As examples of these he mentions 

 alloys, glasses, and solutions. He says that in these cases the 

 limits are determined solely by the quantities required for mutual 

 saturation, but that, between these limits, the most varying 

 proportions occur. 23 



It will be observed that Berthollet thus classes solutions 

 and alloys amongst compounds, and it will now be understood 

 how he was able to distinguish amongst them some with vary- 

 ing proportions. But it is much more remarkable that Ber- 

 thollet also assumes variable proportions amongst the oxides. In 

 one of his essays upon the laws of affinity, 24 in which he speaks 

 of metallic precipitations, he assumes, in accordance with his 

 principles, that the two metals distribute themselves over the 

 oxygen : there are thus formed, according to him, oxides con- 



21 Stat. Chim. i, 373 ; E. I, 282. ti Ibid. I, 367 ; E. I, 276. * Jbid, 

 I* 373-374 5 E- I, 282-283, 24 Ann, Chim. 37, 221. 



