LECTURE III.] HISTORY OF CHEMISTRY. 43 



he shows that many metals form only a single oxide ; that 

 many, however, form two ; and that in those cases where three 

 oxides exist, the intermediate one can be regarded as a com- 

 pound of the other two. 31 In the same way, he endeavours to 

 detect the error in Berthollet's view regarding the existence of 

 sulphur compounds with variable proportions of sulphur. In 

 all these investigations he emphasises the distinction between 

 mixtures and compounds. He says that the latter are charac- 

 terised by quite definite proportions, which hold for compounds 

 occurring in nature as well as for those obtained in the labora- 

 tory, and that this pondus nature lies just as little within the 

 discretion of the chemist as does the law of affinity, which 

 governs all combinations. 32 



But Berthollet also responds by means of facts. He examines 

 the carbonates of the alkali metals, 33 and finds that when the 

 base is saturated by carbonic anhydride under pressure, crys- 

 tals are obtained which differ in composition from the carbonates 

 previously known. He shows that these give up carbonic 

 anhydride on dissolving and heating, and yield salts differing 

 again in composition. He contests the fact asserted by Proust, 34 

 that by leading a trace of carbonic anhydride into an alkaline 

 solution only a few molecules are saturated, while the others 

 remain uncombined. According to Berthollet, a solution of 

 this kind evolves carbonic anhydride on the addition of a drop 

 of hydrochloric acid, and consequently contains a souscar- 

 bonate^ that is, he considers that the trace of carbonic anhy- 

 dride present is distributed over the whole quantity of the base. 

 Rendered cautious by Proust's rejoinders and excellent re- 

 searches, Berthollet now no longer assumes all possible propor- 

 tions between oxygen and the metals as actually occurring. 

 He limits these to a few ; yet, in his examination of the oxides 

 of lead, he asserts that he has isolated four different stages of 

 oxidation which are attained by heating the metal in air. 30 

 All the same he has thereby moved a step nearer to Proust. 



31 Journ. de Phys. 59, 260. 32 Ann. Chim. 32, 31. 33 Journ. de Phys. 

 64, 168. 34 Ibid. 59, 329. 35 Ibid. 64, 181, 3 Ibid. 6l, 35 2 - 



