354 ON TWO NEW CRYSTALLINE COMPOUNDS 



Several specimens of them were analyzed, but no regularity in their composition could 

 be detected, except that they all contained a very much larger amount of zinc than the 

 alloys in which they formed. This irregularity and the imperfection in the crystal- 

 lization seem to be caused by the interference of SbZug, that is, by a tendency to form 

 SbZug, wiiich exhibits itself in a proneness of the crystals of SbZug to an excess of zinc. 

 The line k i has been continued with dots, in order to show that the influence of 

 SbZua extends as far as the alloy of 42.8 per cent of zinc. On returning to the alloy 

 of 31.5 per cent of zinc, and adding an excess of antimony, it was found that the crys- 

 tals formed continued to have the theoretical composition of SbZuj until the amount of 

 zinc in the alloy had fallen to 21 per cent, so that the tendency towards the theoretical 

 composition was so great, that in the alloys between 31.5 and 27 per cent of zinc, crys- 

 tals were formed having very nearly this composition. On still further increasing the 

 amount of antimony in the alloy, the composition of the crystals gradually approached 

 that of the alloy, and from the alloy of 20.2 per cent of zinc very imperfect crystals 

 were obtained, having almost the same composition as the menstruum. At the same 

 time the crystals became less and less perfect, and finally disappeared altogether in the 

 alloys below 20 per cent of zinc. 



The portion of the curve k m n h is the most important result of this investiga- 

 tion, and therefore deserves especial notice. It has been shown that crystals of the 

 form of SbZn^j, or at least crystalline scales of the same character, are formed in the al- 

 loys between 20 and 43 per cent of zinc, the first per cent corresponding to SbZn, and 

 the second to SbZng. Half-way between these two points, that is, the alloy of 31.5 

 per cent, is the point where crystals having the calculated composition of SbZnj 

 are first obtained. Were the variation in the composition of the crystals of SbZuo ex- 

 actly proportional to the excess of zinc or of antimony in the alloy, as is the case with 

 SbZus, then the curve of variation would be the straight line formed by the continuation 

 of the line a b. From this line (b h) the course of the curve is deflected by the force 

 which determines the union of th^ elements in definite proportions, and which, for 

 the want of a special term, I will call the Chemical Force. This is so strong, that the 

 curve runs parallel to the axis of ordinates through the distance k m. Beyond this 

 point the influence of the excess of antimony in the alloy becomes stronger than the 

 chemical force, and the curve gradually bends towards the line h b, which it finally 

 meets at Ji. lu the portion h n of the curve the analyses are best represented by the 

 arc of a circle, of which the radius equals /* e, one half of h b, and to which the 

 line k m is tangent. In the portion n m the points determined by analysis may also 

 be connected by the arc of a circle, of which the radius o' n equals the difference be- 



