344 Researches respecting 



4th, But to ascertain the degree of saturation to which 

 each of these substances can attain^ a separation must be 

 made, which can only be effected by the elasticity, crystal*- 

 lization, precipitation, or action of a solvent : but we have 

 seen that these different means ought to be considered as 

 foreign forces, which alter the results, and which determine 

 the combinations formed, without a possibility of our measur- 

 ing their effect so as to disengage from it that of the elective 

 aflinity; so that the separations which in articles I. II. III. 

 were considered only as an effect of elective affinity and of 

 proportions, are really the effect of a concourse of several 

 forces, as is proved by the observations which followed these 

 articles. 



When it is necessary, for example, to ascertain the quan- 

 tity of the sulphat of potash and of soda formed, as the force 

 of cohesion of these two sulphats does not much differ, it is 

 probable that it would not occasion much change in the pro- 

 portions of the two salts which would crystallize, but it would 

 be necessary to separate the excess of the alkali by means of 

 alcohol in order to obtain the whole crystallization j but al- 

 cohol not acting with equal force upon potash and soda, 

 would produce a new change. To these considerations we 

 may add, that a change of proportions would not only make 

 the force of the soda and the potash (No 1.) to vary, but 

 also that of the alcohol, not to mention the affinity of the 

 water, which serves as a solvent. 



If barytes were to be compared with either potash or soda, 

 we should have changes still more considerable : in that case, 

 the force of cohesion of the sulphat of barytes would be such 

 that it would leave to the alkali but a very small quantity of 

 sulphuric acid, which would be the expression of the ratio of 

 the force of the cohesion of the sulphat of barytes to that of 

 the solvent, rather than of the affinity of the barytes to that 

 of the alkali. 



This is so certain, that if, with the view of comparing the 

 affinity of the acids for barytes, we should begin by treating 

 the barytes with an excess of sulphuric acid, it would be 

 almost entirely precipitated, unless the acid were highly con- 

 centrated 3 it would even be impossible to distinguish the 



combined 



