42 Mr. W. Sutherland on a Kinetic Theory of Solids. 



The readiest way of clinching the whole demonstration now 

 will be to draw the parabola y=l — x 2 from x=0 to as=sl, 

 and with the mean values of N/wi or ~N, calculating w/N as y 

 and 0/T as a, to mark the separate points so determined. 

 This is done in the figure. 



For zinc Tomlin son's values of the rigidity are used, and 

 for iron the melting-point is taken as 1650. To avoid ex- 

 cessive crowding of points on the figure all the points except 

 one for iron, and all for platinum and magnesium are omitted. 

 The course of the curve among the points proves that as 

 regards the connexion between rigidity and melting-point the 



