.Electrons concerned in Metallic Conduction. 



261 



its value at 18° C, a temperature which cannot have been 

 far from that of the surroundings in Drude's experiments. 

 The temperature coefficient of the conductivity is also very 

 small. We may conclude, finally, that when the proper law 

 of velocities is used in the calculation, the value of p calcu- 

 lated from vie will only have a possible error of 4 or 5 per 

 cent. 



Consider now the values of p 1 and K furnished by the 

 equal velocity hypothesis. In the following table some of 

 these are calculated from the formula (4:7), the best con- 

 ductors being selected for convenience of calculation. 



Table V. 



Metal. 



Pi- 



P- 



K. 



Met at,. 



Pi- 



P- 



K. 



Nickel ...... 



Cobalt 



Silver ,. 



•302 

 •427 

 •743 

 •267 

 •455 

 1-36 



2-07 

 2-70 

 2-81 

 298 

 2-85 

 356 



6092 

 80-57 

 38-65 

 75-51 

 4621 

 31-01 



Platinum ... 



Cadmium ... 



: Tin 



•681 

 1-64 

 2-17 

 1-29 

 135 



350 

 4-79 

 5*56 

 551 

 6-66 



60-24 

 4584 

 38 92 

 85-06 

 1013 



Copper 



Gold 



: Zinc 



Aluminium . 



Magnesium . 



This table appears to be decisive between the two alterna- 

 tives. The values of K just calculated are too large to be 

 accepted. An approximate calculation for the other metals 

 shows that they furnish no exception to this statement. 



When these values are regarded side by side with those of 

 Table IV. it seems that we may conclude, with some cer- 

 tainty, that only the Maxwellian distribution of velocities 

 can be admitted, and that it suits the facts very well, and so 

 well that we may proceed further with our conclusions. A 

 very small deviation from this law in the direction of that 

 of equal velocity would cause a rapid increase in the necessary 

 value of K. The figures for silver, the subject of the most 

 careful experiments, are very convincing as regards the 

 perfectly satisfactory nature of the selected law. 



Our ultimate conclusion is, accordingly, that the values of 

 p in Table II. cannot, in the most unfavourable cases, with 

 the possible exception of tin, magnesium, and cadmium, 

 differ from the true values by more than about 5 per cent. 

 But when these numbers are examined, it is seen that, except 

 in the case of cadmium, thev are all very close to integers or 



