Spring took two non-hygroscopic salts, potassium nitrate, KNO3, and well-dried sodium acetate, C2H3NaO2, and left a mixture of their powders for several months in a desiccator. An interchange of metals took place, as was seen from the fact that the resultant mass rapidly attracted the moisture of the air, owing to the formation of sodium nitrate, NaNO3, and potassium acetate, C2H3KO2, both of which are highly hygroscopic.28
When Berthollet enunciated his doctrine the present views of atoms and molecules had yet to be developed, and it is now necessary to submit the matter to examination in the light of these conceptions; we will therefore consider the reaction of salts, taking M and N, X and Y as equivalent to each other—that is, as capable of replacing each other ‘in toto,’ as Na or K,, ½Ca or ½Mg (bivalent elements) replace hydrogen.
And since, according to Berthollet's doctrine, when mMX of one salt comes into contact with nNY of another salt, a certain quantity xMY and xNX is formed, there remains m - x of the salt MX, and n - x of the salt NY. If m be greater than n, then the maximum interchange could lead to x = n, whilst from the salts taken there would be formed nMY + nNX + (m - n)MX—that is, a portion of one only of the salts taken would remain unchanged because the reaction could only proceed between nMX and nNY. If x were actually equal to n, the mass of the salt MX would not have any influence on the modus operandi of the reaction, which is equally in accordance with the teaching of Bergmann, who supposed double reactions to be independent of the mass and determined by affinity only. If M had more affinity for X than for Y, and N more affinity for Y than for X, then according to Bergmann there would be no decomposition whatever, and x would equal 0. If the affinity of M for Y and of N for X were greater than those in the original grouping, then the affinity of M for X and of N for Y would be overcome, and, according to Bergmann's doctrine, complete interchange would take place—i.e. x would equal n. According to Berthollet's teaching, a distribution of M and N between X and Y will take place in every case, not only in proportion to the degrees of affinity, but also in proportion to the masses, so that with a small affinity and a large mass the same action can be produced as with a large affinity and a small mass. Therefore, (1) x will always be less than n and their ratio
x / n
less than unity—that is, the decomposition will be expressed by the equation, mMX + nNY = (m - x)MX + (n - x)NY + xMY + xNX; (2) by increasing the mass m we increase the decomposition—that is, we increase x and the ratio
x / (n - x) ,
until with an infinitely large quantity m the fraction
x / n will equal 1, and the decomposition will be complete, however small the affinities uniting MY and NX may be; and (3) if m = n , by taking MX + NY or MY + NX we arrive at one and the same system in either case : ( n - x )MX + ( n - x )NY + x MY + x NX. These direct consequences of Berthollet's teaching are verified by experience. Thus, for example, a mixture of solutions of sodium nitrate and potassium chloride in all cases has entirely the same properties as a mixture of solutions of potassium nitrate and sodium chloride, of course on condition that the mixed solutions are of identical elementary composition. But this identity of