OF ARTS AND SCIENCES. 217 



This experiment was suggested by the fact that the basic filtrates 

 obtained by adding one or more molecules of KOIi, and removing the 

 precipitate, yield a second precipitate on standing, the length of time 

 apparently depending on the degree of basicity. This second precipi- 

 tate is also produced in a few minutes by boiling or by adding a con- 

 siderable quantity of water. Exp. E^' is therefore interesting as 

 showing that time does not affect the equilibrium while the original 

 precipitate remains in the system. 



Conclusion. 



It is clear, from the evidence herein set forth, that in the reaction 

 which has been the subject of investigation we have another instance of 

 chemical equilibrium oi* mass reaction, and the question at once arises, 

 What is the cau.?e of it ? Without attempting to answer this definitely 

 at present, we suggest the probability that it may be due to the pro- 

 gressive dissociation by water of basic compounds of AI2O3 and SO3 on 

 the one hand, and of potassic aluminate on the other. It would 

 appear from Exp. S', S", and S'", that AX-f)^ . 3 SO3 must itself be 

 partly dissociated in water solution, and it is not improbable that the 

 matter may be still further complicated by the dissociation of another 

 substance formed from the j^'oducts of decomposition of the first. I 

 hope to obtain definite evidence upon this point, by continuing a line 

 of investigation already begun. Tl'.at the basic sulphates of alumiua 

 are dissociated, is apparently indicated by the fact ah'eady cited, tliat 

 the filtrates in experiments B io H yield a second precipitate on 

 heating, diluting, or standing some length of time. While the 

 crystallized potassic aluminate to which Fremy assigns the formula, 

 K2O . Al.Oa . 3 H.,0, decomposes on the addition of water with the 

 separation of aluminic hydrate. 



However that may be, it is evident, (1.) that the reaction between 

 alum and KOII is not by any means as represented in what may be 

 called the ideal equation cited in the beginning of this paper, but that 

 it is much complicated by secondary reactions which render it entirely 

 unsuitable for the determination of the formation heat of the sulphate 

 and other compounds of alumina, although the numericd error intro- 

 duced by its use may not be great ; and (2.) that these secondary reac- 

 tions are of the nature of cliemical equilibriums or mass reactions, the 

 distribution of the constituents of the system about the line of solu- 

 bility being a function of temperature, volume (i. e. dilution), and total 

 mass of constituents. 



