366 Mr. 8. U. Pickering on the Densities 



£3 



actually isomeric, yet they may not really be so, for the com- 

 position of such hydrates cannot be determined very exactly. 

 It may be remarked that the break at the tetrahydrate can 

 be more easily bridged over than any of the others : that at 

 12 per cent, is also a comparatively feebly marked break. 



There was some doubt as to whether the nonohydrate was 

 really isolated or not : the crystallization of it, if it occurred, 

 took place just at a point where the freezing-point figure 

 undergoes a great general change, where the crystallization 

 of water ceases and that of the hydrates begins, and it is 

 interesting to notice that this is just the point at which the 

 irregularities of the density-figure begin to be \ery prominent. 

 The values for the densities minus (1 + 'OlOSjp) show this fact 

 rather more clearly than those plotted in the diagram. 



It might be expected that at the point where a break occurs, 

 and where there must be present a larger amount of the 

 hydrate of the composition indicated by the break than there 

 is in a solution of any other strength, the volumes of the 

 solutions would show a comparative minimum, and the densities 

 a comparative maximum. But this is not necessarily the case, 

 unless the formation of the hydrate is always accompanied by 

 contraction, and the present results show that contraction does 

 not always occur : indeed, most of the breaks seem to occur 

 at points where the densities are at comparative minima (in 

 the diagram a relative increase in the densities will be shown 

 by the figure rising towards the top of the page). 



I have examined the first differences of the densities and 

 give the values for them in Table III., but they do not form a 

 figure nearly so suitable for examination as those already 

 mentioned. This is in part due to the smallness of the differ- 

 ences between many of the consecutive determinations. All 

 that can be said of the nature of the differential figure is that 

 it is rectilineal within the limits of experimental error between 

 any two breaks. In this it resembles the results with potash, 

 calcium chloride, and calcium nitrate, and confirms Men- 

 deleeff's statement as to the general nature of densities ; 

 though it must be remembered that in the two cases on 

 which this statement was chiefly founded — sulphuric acid and 

 alcohol — the first differences are conspicuously rectilineal 

 (see Mendeleeff, " Solutions," and Chem. Soc. Trans. 1887, 

 p. 779 ; Pickering, Chem. Soc. Trans. 1890, p. 79, and Zeit. 

 f. physikal. Chem. vi. p. 10). The theoretical considerations 

 which led Mendeleeff to conclude that the first differences 

 must be absolutely rectilineal are, no doubt, mistaken ; they 



