62 BRIDGMAN. 



features of the transition velocity curves at each point of the transi- 

 tion curve. The following discussion of the transition velocity curve 

 will make plain what the essential features for comparison are, and 

 how it is possible to represent each curve by a single point. 



The four transition velocity curves of Figures 2, 3, 4, and 5 are 

 typical of all. These curves are all similar in appearance except for 

 the hook on the rising pressure branch for AgNOs. This hook is 

 connected with the formation of nuclei of the new phase; it is obvious 

 that immediately after formation of the nuclei the growth of the new 

 phase is less rapid than it is after the surface of separation has had a 

 chance to become fully developed. It is probable that all the velocity 

 curves have this hook in the initial stages, but in most cases it was not 

 possible to observe it. In this discussion we confine our attention to 

 the parts of the curve beyond the hook. ^Yith this restriction, all 

 four curves are essentially similar. It is in the first place evident that 

 the speed of reaction becomes rapidly greater at pressures increasingly 

 remote from the equilibrium pressure. One could not of course, expect 

 otherwise. This means that some special convention is necessary to 

 give any meaning to the term "the velocity of a transition." Such 

 a convention is suggested by the curves themselves. What we mean 

 by a rapid transition is one which increases greatly in speed for a slight 

 shift of pressure away from the pressure of equilibrium. In this sense, 

 therefore, a transition is more rapid if it is represented in Figures 2-5 

 by a steeper curve. Or stated conversely, a "rapid reaction" in this 

 sense is, paradoxically, one that stops rapidly. This so-called " speed" 

 is really the pressure acceleration of speed. It is natural, therefore, 

 to take the slope of the reaction velocity curve as the measure of the 

 speed of the transition. Throughout the rest of this paper, the acceler- 

 ation, measured in this way, w^ill be taken as the " speed." This would 

 give a perfectly definite result if the curves were straight lines, but 

 this is not the case. What is more, it does not seem possible to set up 

 a single type of equation which shall be satisfied by all the curves. 

 The most interesting feature of the curves, however, and that least 

 affected by accidental properties of the rest of the apparatus, is the 

 limiting slope when the velocity of transition becomes zero. This 

 limiting slope is evidently to be obtained by extrapolating the curves 

 until they cross the axis, and drawing the tangent at the point of 

 crossing. The numerical value of the limiting slope of the tangent 

 (the acceleration or "speed") expressed as fractional change per 

 minute per kgm. per cm.^, is one of the essential features referred to 

 above. Corresponding to curves like those of Figures 2-5, there are 



