34: Prof. J. H. Poynting on 



this curvature of surface is a difference of pressure in the 

 liquid; and I suppose the variation of vapour-tension to be due 

 to the difference of pressure. A proof is given of Sir W. 

 Thomson's formula, which seems to bring out more clearly the 

 connexion of the phenomenon with the pressure, and which 

 seems to apply to solids as well as liquids. According to this 

 formula, the steady state (the melting-point) may be reached 

 at any temperature if the pressure can be so adjusted that the 

 vapour-tensions in the two states at that temperature and 

 pressure are equal. The resulting lowering of the melting- 

 point by pressure agrees in amount with that given by the 

 well-known formula of Prof. J. Thomson. 



It follows from this mode of regarding the subject, that, if 

 in any way the ice can be subjected to pressure while the water 

 in contact with it is not so subjected, then the lowering of the 

 melting-point per atmosphere is about 11^ times as great as 

 when both are compressed. I give the results of some expe- 

 riments which I have made to test this,, and which certainly 

 seem to indicate that the fall of melting-point is much greater 

 than the amount usually supposed if the ice alone be compressed. 



The isothermals for ice-water are then discussed. It has 

 been supposed that, if we could employ a sufficiently low tem- 

 perature and high pressure, then ice would pass continuously 

 into water ; that is, the isothermals would have no horizontal 

 part corresponding to a mixture of ice and water, and we 

 should have a critical point. Assuming, however, that a mix- 

 ture of ice and water completely freed from foreign gases can 

 be subjected to great negative pressure or tension, it seems 

 probable that there is another critical point at a temperature 

 above 0° and at a high negative pressure ; that is, the water- 

 ice line is a closed curve. We know that below 0° the water 

 isothermals can be prolonged below the horizontal portion, 

 since water is unfrozen in certain cases, — and that the ice iso- 

 thermals can be prolonged above the horizontal portion; for ice, 

 at 0° say, can be suddenly compressed without melting in the 

 interior. This suggests that the true form of the isothermals 

 is a continuous curve, of the nature which Prof. J. Thomson 

 has suggested in the case of liquids and their vapours. 



If we suppose that the curves are continuous in the same 

 manner for ice-water above 0°, then Prof. Carnelley's " Hot 

 Ice" would seem to be represented by the prolongations up- 

 wards of the ice isothermals beyond the horizontal line to 

 where they meet the line of no pressure. The critical point, 

 which certain assumptions roughly fix at about 14° C, would 

 then be an upper limit, or rather above the limit, to the tem- 

 perature of hot ice in a vacuum. 



