1909] on Experiments at High Temperatures and Pressures. 551 



Will a state of affairs be reached in which it is no longer possible to 

 distinguish between the liquid and its crystalline form ? Will there 

 be, in fact, a sort of critical point at which the melting-curve will end ? 

 At present we can only say that no indications of such an occurrence 

 have been observed experimentally, and Prof. Tammann takes the 

 point that it is highly improbable that anything in the nature of con- 

 tinuous transformation can take place, because a crystal has different 

 properties in different directions related to its axes, and there is thus 

 a much greater qualitative difference between crystals and liquids than 

 between Uquids and gases, both of which are isotropic. I must admit 

 that this argument does not appeal to me very strongly. If it be 

 possible to compress a substance till it reaches a state in which, at one 

 and the same temperature, the liquid has the same density as the 

 crystals, presumably the mean distance of the molecules will be the 

 same in both cases. I see nothing monstrous in the view that under 

 these circumstances crystallisation may set in gradually, and that it 

 may not be possible to say exactly when the liquid ceases to be a fluid 

 and becomes a crystalline solid. There are no theoretical or other 

 grounds for supposing that the phenomena of crystal grow^th as 

 observed when there is a change of volume accompanying the crystal 

 formation, will necessarily hold when no such change of volume occurs. 



If we refer to the theory of the change of m.p. by pressure 

 it is obvious that if either the change of volume or the latent heat 

 of melting vanish at any temperature or pressure on the melting- 

 curve, then in the neighbourhood of this pressure the curve must 

 degenerate to a point — or small pressure changes will not affect the 

 m.p. It was pointed out, however, that there is a term or terms 

 depending on the square of the pressure, and if these were relatively 

 important the only thing we should notice would be a change of 

 curvature at the point under consideration. It does not follow that 

 there is no maximum or minimum to the melting temperature of any 

 particular substance because the term in P- may be vanishingly 

 small : it may be (and generally is) of opposite sign to the term in P, 

 and in this case it is only a question of the relative importance of 

 the terms where the maximum or minimum melting-point lies. 

 Damien's empirical formula expresses precisely the effect to which I 

 refer. The practical result which is of importance in questions 

 affecting the condition of the inner layers of the earth is that we are 

 not entitled — in fact, it is wrong— to suppose that pressure must 

 necessarily go on raising the melting-point indefinitely ; everything 

 depends on the substance under consideration. It is therefore 

 necessary to make such experiments as those of Tammann at vastly 

 higher temperatures and pressures than those we have been consider- 

 ing, up to probably over 10,000 kilograms per sq. cm. (or 63*5 tons 

 per sq. inch). 



In 1893 some experiments were described by Parsons * in which 



* Phil. Mag. xxxvi. 304. 



Vol. XIX. (No. 103) 2 o 



