484 Mr. W. Sutherland on the 



trihydrol is continually dissociating into dihydrol, and some 

 of the dihydrol associating to trihydrol ; but the most striking 

 dissociation of trihydrol into dihydrol is that which accom- 

 panies the melting of ice. Under a given pressure this 

 occurs with remarkable definiteness at a corresponding 

 temperature. 



The thermodynamical discovery and explanation of the 

 lowering of the melting-point of ice by pressure, though so 

 justly famous in the history of the science of heat, give no 

 clue to the mechanical or molecular cause of this remarkable 

 phenomenon. In a solid body we assume the molecules so 

 to hinder one another's motion, that each is kept vibrating 

 within narrow limits round a certain point. With rise of 

 temperature the mean distance between two neighbours 

 increases, and when the distance increases to such an extent 

 that the molecules can no longer hem one another in, they 

 all start migrating, and the solid is said to melt. Now at the 

 melting-point an increase of pressure, by forcing the mole- 

 cules nearer to one another, ought to necessitate an increase 

 of temperature to cause melting, for the increase of tem- 

 perature would neutralize ihe effect of the increase of pressure. 

 This is the molecular explanation for the usual fact that 

 increase of pressure increases the melting temperature. But 

 in the exceptional case of ice, although it expands with heat 

 as long as it is ice, and shrinks under pressure, yet though 

 pressure brings the molecules nearer together, they escape 

 by melting at a lower instead of a higher temperature. 

 Moreover ice at 0° is a solid of very considerable rigidity, 

 though it is on the very verge of melting. Its actual 

 rigidity has not been measured, but as it has a Young's 

 modulus between 20,000 and 90,000 (probably 70,000) kgms. 

 weight per sq. cm., that of glass at 15° being 500,000, and of 

 silver 760,000, it is evident that ice on the verge of melting 

 still enjoys a remarkable degree of rigidity. The melting of 

 ice is of an entirely different character from that of an 

 ordinary physical melting. We have proved it to be accom- 

 panied by considerable dissociation. A simple conception of 

 the dynamical cause of this dissociation can be obtained by 

 following the lines of investigation of " A Kinetic Theory of 

 Solids " and " Further Studies on Molecular Force " (Phil. 

 Mag. [5] xxxii. & xxxix.). In these papers it is assumed 

 that in solid compounds atomic motions are more important 

 than molecular, because of Joule's and Kopp's law that each 

 atom contributes a definite atomic heat. The atoms in the 

 solid are still held by the chemical forces in their proper 

 arrangement to form the molecule, but the molecules are so 



