448 Royal Society : — Prof. J. Thomson on the 



is necessary here as an introduction for rendering intelligible what 

 is to follow. 



Taking into consideration any substance which we can have in 

 the three states, gaseous, liquid, and solid, we may observe that, 

 when any two of these states are present in contact together, 

 the pressure and temperature are dependent each on the other, so 

 that when 6ne is given the other is fixed. Then, if we denote geo- 

 metrically all possible points of temperature and pressure jointly by 

 points spread continuously in a plane surface, each point in the 

 plane being referred to two axes of rectangular coordinates, so that 

 one of its ordinates shall represent the temperature and the other 

 the pressure denoted by that point, we may notice that there will 

 be three curves — one expressing the relation between tempera- 

 ture and pressure for gas with liquid, another expressing that for 

 gas with solid, and another expressing that for liquid with solid. 

 These three curves, it appears, must all meet or cross each other 

 in one point of pressure and temperature jointly, which may be 

 called the triple point*. 



The curve between gas and liquid, which may be called the boil- 

 ing-line, will be a separating boundary between the regions of the 

 plane corresponding to the ordinary liquid and those corresponding 

 to the ordinary gaseous state. But by consideration of Dr. An- 

 drews's experimental results (Phil. Trans. 1869), we may see that 

 this separating boundary comes to an end at a point of tempera- 



* In making this statement, that it appears that the three curves must all 

 cross each other in one point, I would wish to offer it here (as I previously did in 

 the 1871 British-Association paper) subject to some reserve in respect of condi- 

 tions not yet known with perfect clearness and certainty. I have to suggest 

 that we might not be quite safe in assuming that, within a cavity containing 

 nothing but pure water-substance partly gaseous, the melting temperature and 

 pressure of ice solidified from the gaseous state would be the same as the melting 

 temperature and pressure of ice frozen from the liquid state, and in making other 

 suppositions, such as that the same quantity of heat would become latent in the 

 melting of equal quantities of ice formed in these two ways, and in neglecting 

 conceivable but, I presume, as yet imperfectly known distinctions of capillary 

 conditions between ice amply wet with water and ice only moistened with the 

 last vestiges of water before the whole liquid may be either evaporated or frozen. 

 It might be a question in like manner whether we can be sure that there can be 

 theoretically a condition of repose in a cavity containing only perfectly pure 

 water-substance in which the three states are present together, each in contact 

 with the other two, so that there would be ice partly wet with water, and partly 

 dry in contact with gaseous water-substance, or steam as it may be called, while 

 the water and steam were also in contact with each other. I offer these remarks 

 by way of caution, as they force themselves into notice when we attempt to 

 sketch out the features of the three curves under consideration, and because they 

 may serve to suggest questions for experimental and theoretical investigation 

 which may have been generally overlooked before. In the present paper, how- 

 ever, I proceed on assumptions, such as are usually tacitly made, of identity in 

 the thermal and dynamic conditions of pure ice solidified in different ways, 

 assumptions which, so far as is known, may be, and probably are, perfectly true ; 

 and I proceed on the supposition that there can be theoretically the condition 

 of repose here alluded to, of the solid, liquid, and gaseous states, present together 

 each in contact with the other two — and consequently that the three curves would 

 meet or cross each other in one point, which I have called the triple point. 



