366 ON TWO NEAV CRYSTALLINE COMPOUNDS 



the fact that the crystals of SbZna, which correspond most closely to it, are so very im- 

 perfect, that they would hardly be recognized as crystals did they not form the lower 

 limit of a series. Several other facts pointing in the same direction might be added, 

 but sufficient, it is thought, has been said to show that the variations of composition 

 described in this paper cannot be explained either by mechanical impurities in the 

 crystals or by the mixture of isomorphous compovmds. 



In the absence of any kno^^-n principle of chemical science by which the remarkable 

 variations of composition that have been demonstrated in this memoir can be explained, 

 the conclusion is almost forced upon us, that zinc and antimony are capable of uniting 

 and producing definite crystalline forms in other proportions than those of their chemi- 

 cal equivalents ; in other words, that the law of definite proportions is not so absolute 

 as has been hitherto supposed. The explanation, then, of the variation of composition 

 which I would offer is, that it is due to an actual perturbation of the law of definite 

 proportions produced by the influence of mass. I suppose, for example, that in the 

 crystals of SbZuj containing 55 per cent of zinc, the zinc and antimony are united in 

 exactly the same way as in those containing 43 per cent, or, in other words, just as if 

 the equivalent of zinc were increased to 52.57, that of antimony remaining the same. 

 In support of this position I would offer two considerations. The first is, that, if the " 

 variation is not caused by mechanical impurities or by the mixture of isomorphous com- 

 pounds, we can conceive of no other explanation for the phenomenon than the one of- 

 fered. This, of course, is merely negative cA-idence, for although science as yet presents us 

 with no principle for explaining variations of composition other than those which have 

 been discussed, and although we can conceive of none others, it does not follow that 

 others may not exist, or may not hereafter be discovered ; but, nevertheless, this con- 

 sideration is important, inasmuch as it meets an obvious objection, which Avould be 

 urged against any new doctrine which conflicts with a generally received canon of 

 chemical philosophy. The second consideration has the character of demonstration. 

 It is that the curve of variation is evidently generated by a second force counteracting 

 directly the chemical force. This second force, as has been shown, is exerted by the 

 excess of one or the other elements present in the menstruum, and it may therefore be 

 appropriately termed the force of mass. While the chemical force tends to make the 

 curve a straight line parallel to the axis of ordinates, the force of mass would reduce it 

 to a straight line, making an angle of 45° with the axis. Under the influence of both 

 these forces, it follows the arc of a circle between the two. Now, I urge that the char- 

 acter of this curve proves that the chemical force has been directly influenced by what 

 we have called the force of mass, in the same way that the irregularities of the orbits 



