Change of State : Solid-Liquid. 47 



On considering the isothermal s below 0°, it will be noticed 

 that the water-isothermals, at least as far as that for —20°, 

 can be prolonged downwards past the horizontal line to meet 

 the line of no pressure; for Despretz succeeded in cooling water 

 to —20° in thermometer-tubes without freezing. These pro- 

 longations are represented by a af, b ¥, c c f (fig. 2). Similarly 

 the ice-isothermals can be at least slightly prolonged upwards 

 past the horizontal line. For, suppose we take a block of ice 

 at 0° and suddenly subject it to great pressure. Since it ex- 

 pands on heating, then sudden compression produces, if any 

 thing, a slight rise in the temperature. At the same time the 

 melting-point is lowered, and the ice begins to melt at the 

 surface, and in time the whole will be lowered to the new 

 melting-point. But just at first, and until it falls to that 

 temperature, we have the ice on the prolongation of the iso- 

 thermals upwards as at A A' or B W in fig. 2. In a certain 

 sense, then, we have "hot ice." 



Since, then, the water-isothermals may be prolonged down- 

 wards and the ice-isothermals upwards, we may probably here 

 adopt Prof. J. Thomson's suggestion as to the true shape of 

 the isothermals in the case of liquid-and-gas mixtures (Brit. 

 Assoc. Beport, 1871, p. 30 ; Maxwell's ' Heat,' p. 125). This 

 is indicated by the dotted line for —2° in the figure. If the 

 isothermals also have this shape above 0° (as indicated by the 

 dotted line for the 4° isothermal), then at first the ice-isother- 

 mals will be prolonged upwards to meet the line of no pres- 

 sure, as, for instance, that of 4° at H. This seems to be the 

 place where we must put Dr. Carnelley's " hot ice," on the 

 diagram, if its temperature be really proved to be above 0°. 



But if the critical point for the higher temperature exist, it 

 is evident that, before this temperature is reached, the prolon- 

 gations of the ice-isothermals will cease to reach up to the line 

 of no pressure, and the limit to the temperature of hot ice in 

 a vacuum is that of the last isothermal which touches the line 

 of no pressure. To obtain ice at still higher temperatures, it 

 would apparently have to be subjected to great tension. If 

 the above calculation for the critical point is at all near the 

 truth, then the highest temperature possible for ice in a vacuum 

 is something below 14° 0. 



The view here advocated as to the nature of the melting of 

 ice, would show that its fixity is as much a " constant acci- 

 dent " as the fixed boiling-point of water. If we have a piece 

 of ice at any temperature and allow no water to form on its 

 surface, then I see no reason why it should melt if heat be 

 supplied to it by conduction from bodies which, when melted, 



