342 



NATURE 



[Fed. lo, 1 88 1 



288° and re-solidifies at [27o°-275°, i.e. at a temperature 

 identical with that at which it solidifies under diminished 

 pressure as above described. 



After the above experiments had been made the 

 investigation had to be unavoidably deferred, and was 

 not resumed till last autumn, when a large number of 

 determinations were made of the boiling-points of several 

 different substances under various pressures, and from 

 these was drawn the general conclusion described in a 

 letter to Nature (vol. xxii. p. 434), in September last, 

 viz. : " In order that any solid substance may become 

 liquid it is necessary that the pressure be adoi>e a certain 

 point, called the critical pressure, otherwise it cannot be 

 melted, no matter how great the heat apphed." Assuming 

 the truth of this conclusion, I set to work to apply it in the 

 case of ice, as it would undoubtedly have the greatest 

 interest in connection with that substance. On this 

 account my experiments since the end of August have 

 related almost solelv to ice. 



The problem to be solved was whether ice could be 

 prevented from melting by maintaining the pressure 

 below its critical pressure, /.(■. the tension of its vapour at 

 the melting-point, and that whatever the intensity of the 

 heat applied. Now the theory of critical pressure gives 

 us no information as to whether the ice, on non-fusion, 

 would or would not rise above its ordinary melting-point 

 when strongly heated, but as this result had been pre- 

 viously attained in the cas,e of mercuric chloride it 

 appeared not impossible that the ice might become hot. 



The Cjuestioii as to the rise of temperature of the ice 

 above o" , though at first but a si'ie issue of the investiga- 

 tion, became from its more especial interest the chief 

 object of inquiry, and the experiments which have been 

 made and those which are at present in hand relate almost 

 solely to this point. 



The great difficulty to be overcome was to maintain the 

 pressure in the containing vessel below 4"6 mm., zV. the 

 tension of aqueous vapour at the freezing-point ; for it 

 will be easily understood that if the ice be but slightly 



heated the quantity of vapour given off would soon be 

 sufficient to raise the pressure above that point. After 

 several fruitless attempts, the following plan, involving 

 the principle of the cryophorus, was adopted : — A strong 

 glass bottle, such as is used for freezing water by means 

 of Carre's pump, was fitted with a cork and glass tube C 

 (Fig. 2) and the cork well fastened dov/n by copper wire. 

 A and c were then filled with wet mercury (the water 

 facilitating the removal of the air-bubbles) and C con- 

 nected with the end of the tube D e by means of the stout 

 india-rubber tubing B, a thermometer having been pre- 

 viously attached by the wire .i' to the lip of the tube at B. 

 The tube d e was about one inch diameter, and about 

 four feet long from the bend to the end E ; after connec- 

 tion with c it was completely filled with mercury and the 

 whole inverted over the mercurial trough f, as shown in 

 the figure, when the mercury fell to o, the ordinary height 

 of the barometer. The mercury was run out of A by 

 tilting up the bottle and inclining the tube D E. By this 

 means a Torricellian vacuum was obtained from A to O. 

 D was next brought to the vertical, and the bottle A placed 

 in the trough p. A tin bottle G without a bottom was 



Flame 



of 

 Bunsenls 

 Burner \ 



Fig. 3 



1 Flame 

 of 

 Bunsen's 

 Burner 



Mercury 



fitted with a cork, so that it might slide somewhat stifHy 

 along i) e. . 



To begin with, the tin bottle was placed in the position 

 c and filled with a freezing mixture of salt and ice. Some 

 boiled water was then passed up into the tube D E, suffi- 

 cient to form a column at M about two inches deep. The 

 thermometer H had been previously arranged sothat itsbulb 

 might be one or two inches above the surface of the water M. 

 The bottle a was next surrounded by a good quantity of 

 freezing mixture, in order that any vapour given oft" from 

 the water at M might be condensed in A as fast as it was 

 formed, and thus the internal pressure might never be 

 more than about ro to I'S mm. When A had been suf- 

 ficiently cooled, which required about fifteen minutes, the 

 tin vessel G was slid down the tube d e, and its freezing 

 mixture removed. The water at M had then solidified to 

 a mass of ice, which on heating with the flame of a Bun- 

 sen's burner, melted either wholly or partially, and the 

 liquid formed began at once to boil. The fusion com- 

 menced first at the bottom of the column of ice, whereas the 

 upper part fused only with difficulty, and required rather 

 a strong heat. The fusion in this case was probably due 



