Influence of Atomic Weight. 307 



recently M. Raoul Pictefc (Conpt. Rend, lxxxviii. p. 855, 

 April 28, 1879) has fixed this relation still more exactly; for 

 he concludes : — (1) The higher the melting-point of a solid, 

 the shorter are the oscillations of its molecules. (2) The 

 melting-points of solids correspond to equal lengths of oscilla- 

 tion ; and therefore the product of the length of oscillation by 

 the melting-point is constant for all solids. Or, expressed in 

 the form of an equation, 



= c; 



where t = melting-point, reckoned from —273, u = coefficient 

 of expansion, cl = specific gravity, p = atomic weight, and 

 c = constant. 



If we can prove that one or more certain properties all vary 

 regularly with a certain other property, and if we can offer 

 some explanation of this last property, we are in a fair way of 

 being able to explain the other properties. Thus it has already 

 been pointed out that, the higher the melting-point of an 

 element, the less is its coefficient of expansion ; and it can be 

 easily shown that both these phenomena are dependent on the 

 attraction between the molecules. It is generally allowed that 

 in the solid condition the forces which draw the molecules 

 together are greater than those which tend to drive them 

 apart. In the liquid condition the opposing forces are more 

 equally balanced, and the molecules move freely over one an- 

 other. In the state of gas, on the other hand, the repulsive 

 force is stronger than the attracting, and the body fills the 

 whole space in which it is placed. If, then, a substance melts 

 when the repulsive force = attracting force (and the repulsive 

 force, we know, is due to heat), then it follows that the greater 

 the attraction between the molecules of a body the greater will 

 be the heat required to melt it. Consequently, if we compare 

 a number of bodies at the ordinary temperature, their melting- 

 points will be the higher the greater the attraction between 

 their molecules, because the greater the heat required to over- 

 come that attraction. We may therefore say that the melting- 

 point of a body is proportional to the attraction between its 

 molecules, when there is no repulsive force acting, i. e. at the 

 absolute zero of temperature. If this be true, then we can 

 easily see that the melting-point of a substance ought to bear 

 some relation to its coefficient of expansion, hardness, &c, 

 since these properties are evidently dependent on the attraction 

 between their molecules ; for the greater the attraction between 

 the molecules of a body, the greater will be the heat required 



