66 THE ROYAL SOCIETY OF CANADA 



Here we find that there is no permanent increase in the decrement 

 as there should be if diffusion were the only cause operating. Figure 2 

 shows a sHght decrease in the decrement after the vapour has been 

 heated to a high temperature and Figure 10 shows that the decrement 

 increases at first and then decreases when the temperature is main- 

 tained constant. The initial increase in the latter case is due evidently 

 to diffusion; but the latter decrease points to a permanent change of a 

 different character, a change in the substance. This change may be 

 brought about by raising the temperature or by protracted heating 

 at constant temperature. Probably it occurs more rapidly at higher 

 temperature. 



Evidently, then, the curves for rising temperature, for the liquid 

 side at least, are always too low, as the process above mentioned is 

 going on all of the time during the heating. The values for falling 

 temperatures are too high. Apparently maintaining at constant 

 temperature on the initial rise results in a decrease in the decrement 

 and on the fall gives an increase. Experimental results obtained 

 thus far seem to indicate that at the same temperature the final results 

 are the same. On the falling temperature curve we are evidently 

 dealing with a substance which if not homogeneous is thoroughly 

 mixed and diffusion does not obscure the result. Even below the criti- 

 cal temperature diffusion will not be important, as it is noticed that 

 after raising to a high temperature condensation at the critical 

 temperature occurs equally in both legs of the tube. 



It will be evident from the foregoing that there are real changes 

 in the nature of the substance (at least for these tubes) near the 

 critical temperature, that these changes are slow near the critical 

 temperature but progress as the time goes on. The change is not 

 confined to the immediate region near the critical temperature but goes 

 on at higher temperature. It seems that there is a sort of equilibrium 

 for any temperature above the critical temperature, which condition 

 is not reached instantly but after the lapse of considerable time. 

 When the temperature is rising the decrement is constantly falling 

 toward the equilibrium value, and when the temperature is falling 

 the change is taking place in the reverse direction. The more the 

 temperature is raised the lower is the value of the decrement when the 

 temperature is falling and the farther it is from equilibrium at any 

 temperature. 



Study of the salient point on the curve for falling temperature 

 (Figure 9) shows that there is a decided change at this temperature 

 which we may call the critical temperature. Below this point return 

 to equilibrium is more rapid but requires more time if the temperature 

 was raised higher before the falling temperature readings were taken. 



