It will be well to observe that a number of other methods, especially from the province of those physical properties which are clearly dependent on the magnitude of the atom (or equivalent) or of the molecule, may lead to the same result. I may point out, for instance, that even the specific gravity of solutions of the metallic chlorides may serve for this purpose. Thus, if beryllium he taken as trivalent—that is, if the composition of its chloride be taken as BeCl3 (or a polymeride of it), then the specific gravity of solutions of beryllium chloride will not fit into the series of the other metallic chlorides. But by ascribing to it an atomic weight Be = 7, or taking Be as bivalent, and the composition of its chloride as BeCl2, we arrive at the general rule given in Chapter VII., Note 450. Thus W. G. Burdakoff determined in my laboratory that the specific gravity at 15°/4° of the solution BeCl2 + 200H2O = 1·0138—that is, greater than the corresponding solution KCl + 200H2O (= 1·0121), and less than the solution MgCl2 + 200H2O (= 1·0203), as would follow from the magnitude of the molecular weight BeCl2 = 80, since KCl = 74·5 and MgCl2 = 95.
The specific heats here given refer to different limits of temperature, but in the majority of cases between 0° and 100°; only in the case of bromine the specific heat is taken (for the solid state) at a temperature below -7°, according to Regnault's determination. The variation of the specific heat with a change of temperature is a very complex phenomenon, the consideration of which I think would here be out of place. I will only cite a few figures as an example. According to Bystrom, the specific heat of iron at 0° = 0·1116, at 100° = 0·1114, at 200° = 0·1188, at 300° = 0·1267, and at 1,400° = 0·4031. Between these last limits of temperature a change takes place in iron (a spontaneous heating, recalescence), as we shall see in Chapter XXII. For quartz SiO2 Pionchon gives Q = 0·1737 + 394t10-6 - 27t210-9 up to 400°, for metallic aluminium (Richards, 1892) at 0° 0·222, at 20° 0·224, at 100° 0·232; consequently, as a rule, the specific heat varies slightly with the temperature. Still more remarkable are H. E.