Kinetic Theory of Solids. 543 



good test of the kinetic theory of solids. The best way to 

 apply the test will be to calculate for each metal from the 

 observed values of q and the theoretic values of k the values 

 of Q, — these should come the same. For example, according 

 to my experiments on zinc the values of qlO~ Q at absolute tem- 

 peratures 289°, 333°, and 376° are 821, 758, and 682, and 

 the corresponding values of Q10" 6 are 1280, 1320, and 1320, 

 which satisfy well the requirement of constancy. Kiewiet*s 

 values for zinc, found by the method of bending at 283°, 325°, 

 and 353°, are 1031, 975, and 924, which give for Q10" 6 the 

 values 1710, 1800, and 1870, which do not satisfy the con- 

 dition of constancy. It will be seen from equation (13) that 

 a good deal depends on a correct absolute value of q, as well 

 as on the relative values at different temperatures. For the 

 other metals I will simply give the values of Q10 -6 calcu- 

 lated from 3 values of q at temperatures from 289° to 376° 

 on the absolute scale. For tin 880, 890, and 990, and ac- 

 cording to Kiewiet 740, 750, and 720 ; lead 326, 332, and 

 339 ; magnesium 439, 443, and 433 ; copper 1590 and 1540 ; 

 aluminium 818 and 839 ; aluminium unannealed 735, 746, 

 and 752 ; and iron 2253, 2241, and 2226. It is evident that 

 experimental uncertainties do not at present let us expect any 

 better agreement between theory and experiment than that 

 just shown. But there is another form in which the same 

 comparison can be made. One of the results of equation (13) 

 is that Young's modulus at absolute zero is 3 times the 

 rigidity. Now by means of (13) we can from the mean 

 values of q at 15° C. given in Table VII. calculate the values 

 of Q, and compare them with the values of N, given in 



Table V. 



Table VIII. 

 10~ 6 Q or 10~ 6 times Young's modulus at absolute Zero. 

 Cu. Ag. Au. Mg. Zu. Cd. Al. 



Q 1390 860 840 490 1490 1040 840 



Q/N... 3-1 2-9 2-9 3-0 3'5 3*2 



Sn. Pb. Fe. HI. Co. Pd. Pt. 



Q 700 270 2170 2460 1970 1130 1690 



Q/N... 3-5 2-3 2-8 3-2 2*6 



With the exception of platinum, the only metals for which 

 the departure of N/Q from the value 3 is beyond the limits 

 of experimental error are zinc, tin, and lead, the metals for 

 which the results of different experimenters disagree most, as 





