6i6 



NA TURE 



[April 29, t8S6 



In the present paper, however, I shall confine myself mainly 

 to the subject of solution, for two reasons. First, because it 

 seems to me that if we can satisfactorily account for this part of 

 the subject, the remainder will be easily and naturally explained ; 

 and, second, because from a study of Thomsen's researches on 

 thermo-chemistry, as given in Muir and Wilson's recent work 

 on that subject, I have obtained data which seem to me almost 

 to demonstrate the truth of niy views on the subject of solution. 



In my paper of 1881 I explained solution to be due to the 

 affinities of the constituent elements of the body dissolved for the 

 constituent elements of the solvent. Thus NaCl dissolves in 

 water because of the affinity of the Na of the salt for the O of 

 the water, .and of the CI of the salt for the H of the water. 

 These affinities not being strong enough to cause double decom- 

 position, an indefinite compound is formed having the properties 

 of what we call a solution. If this explanation be correct, we 

 should expect that the relative strengths of the aftinities of Na, 

 O, CI, and H should have an effect on solution, and that if we 

 substituted another metal for Na whose affinity for CI and O 

 was greater or less, we should have a corresponding change in 

 the solubility of the salt. If, for instance, this other metal had 

 greater affinity for CI and less for O, we should expect the salt 

 to be less soluble, because the CI would be held more firmly, 

 .-nd could not act so energetically on the H of the water, while 

 I he action of the metal on the O of the water would also be 

 less. Now it may be admitted generally that the heat evolved 

 in similar chemical operations is a measure of the chemical 

 affinities of the elements concerned, or at least of their relative 

 affinities. With these explanations, let us consider some of 

 Thomsen's results. He finds as the result of numerous experi- 

 ments that as the atomic weight of the metal increases in 

 similar compounds of Mg, Ca, Sr, and lia, 



(i) Heat evolved in production of MCI., inctrasus 



(2) ,, ,, MO 'decreases 



(3) ,, solution of MCI., ,, 



(4) Solubility of MCI; in water ' ,, 



These results apply to the alkali metals also. 



To make these considerations plain, consider the following 

 table :— 



Now the order of solubility of these chlorides is exactly as 

 these results would lead us to expect, MgCI, being most soluble, 

 and BaClo least so, while CaClj and SrCl., are intermediate ; 

 and the whole result is exactly what we should e.xpect if my 

 explanation of solution be correct. The chloride of the metal 

 whose affinity (as measured by the heat evolved) is greatest for 

 CI, and least for O is the least soluble. Again, let us take the 

 following series of elements in which the heats of combination 

 with equivalents of either CI or O gradually decrease (excepting 

 Na for O), but that of CI much faster than that of O, so that 

 after passing Al the heat of combination of one equivalent of O 

 is greater than that of CI, and consider the action of the 

 chlorides towards water : 



NaCl Soluble. 



MgCIo ,, but decomposed on evaporation. 



.\ICI3 ,, forms crystals with water, and decomposed 



on further heating. 



ppj ^ / In contact with water, more or less quickly decom- 

 ■^^■^ \ posed. 



SCI I 



op, ! Instantly decomposed in contact with water. 



CI gas Dissolves in water, which it slowly decomposes on 

 exposure to sunlight. Also at low tempera- 

 ture combines with water, forming compound 

 CI5H2O. The behaviour of this gas is interest- 

 ing, becouse it brings gases under the same prin- 

 ciple of solution. 



KCl Soluble. 



We have thus a regular gradation of change, from simple 

 solution through double decomposition to solution again, ac- 

 cording as the affinities of the elements for CI and O vary. Thus 

 solution appears as a periodic function of the elements. 



Let us take another case, where the affinities of the elements 

 for CI and O again regularly decrease, but where the affinity for 

 the O diminishes faster than that of the CI, and note the result. 

 The following group of elements represents this case : 



NaCl Soluble. 



CuCl Insoluble, or nearly so. 



CuClj Soluble. 



AgCl Insoluble. 



AuCl 



AuClj Soluble. 

 Now it will be observed that Cu, Ag, and Au have a small 

 affinity for O compared to what they have for CI ; especially is 

 this the case with Ag, and when these elements combine with 

 only one atom of CI, the chloride is insoluble ; the affinity of 

 the single atom of CI for the H of the water, combined with the 

 small affinity of the metal for the O of the water, is too weak to 

 produce solution ; but as soon as one or more atoms of chlorine 

 are taken up, the compound becomes soluble through the 

 increased affinity acting on the water. 



All this is quite consistent with the view of solution I have 

 proposed. In truth, this view is the most simple and natural 

 explanation of the facts. There are not sufficient data to be 

 obtained to trace these actions through all the various groups 

 and series of elements, but there are numerous indications that 

 they are regular recurring phenomena. I need only mention 

 the analogous behaviour- towards water of the chlorides of P, As, 

 Sb, Bi and S, Se and Te. 



I have chosen the chlorides to illustrate the principles of 

 solution because of the simplicity of their composition and 

 action, and also because the data are more complete for them 

 than for other salts, but the same principle can be traced 

 through all. 



Let us consider, however, for a moment the solution of the 

 oxides. In this case, as the metal is already combined with O, 

 it is not likely to have so much effect on the O of the water. 

 We see, however, the same action taking place. Thus take the 

 following group and their heats of combination and solution : 



,,, . , Heat of formation Heat of solution 



^'='='1 of oxide ofox.de 



Mg ... ... 146,000 2,960 



Ca 130.930 18,330 



Sr 128,440 29,340 



Ba 124,240 34.520 



Now again we find that the oxide which has the greatest heat 

 of formation is the least soluble, because the elements are held 

 more firmly together, and less affinity is left to act on the water 

 elements. Thus MgO is almost insoluble, while BaO is the most 

 soluble of the group, and CaO and SrO intermediate. This 

 order is the inverse of that of the chlorides, as the heat of 

 formation is also inverse. 



Many examples of the close relationship between the heats of 

 formation and the heats of solution might be given. For 

 example, if we neutralise the hydrates of the above oxides with 

 a solution of HCI, we find the heat evolved approximately the 

 same in all cases, being about 27,600 units. At first sight this 

 appears very curious, seeing their affinities (as measured by heat 

 evolved), for CI and O vary so much and inversely ; but the 

 explanation is obvious enough. The heats of solution of the 

 various compounds decomposed and formed exactly compensate 

 for the variations in the heats of combination ; what is lost in 

 one way is gained in another. 



Closely connected with solution is the subject of crystallisation. 

 This also is most satisfactorily explained, if the principle I 

 contend for be admitted, and regular structure necessarily follows. 

 For instance, in such a compound as BaCL 61I„0, the atoms of 

 the molecule must be arranged somewhat after the following 

 fashion : 



Hj O 



O H., 



11" 



EO— Ba— CI„— So 



I r " 



O H., 

 H., O" 



There are many other points connected with this subject, but 

 as the data I can obtain are fragmentary, I shall content myself 

 at present with indicating the direction of my inquiries. 



( I ) The affinities, measured as before, of the series Na, Mg, 



