¢ LAAD 
In fact, at O°C. the temperature coefficients of the resistances of 
pure metals are, as a rule, greater than 0.00367, and, for platinum, 
gold, silver and lead they lie in the neighbourhood of 0.0039 
and 0.0040. 
And lastly, the formula also expresses well the fact that the 
diminution of resistance diminishes in quantity at hydrogen tempe- 
ratures, and that in greater degree for substances of high melting 
point than for those of low melting point. 
An accurate numerical equation, however, such as to determine 
even the bend in the curve that represents the resistance as a func- 
tion of the temperature can be obtained only on the assumption 
that smaller values of » and, therefore, of a come in the front at lower 
temperatures. Definite values, indeed, cannot be ascribed to r. Einstun”), 
for instance, deduces from its elasticity a =— 200 for silver (for he 
gives 4= 73.10 + cm. for the wave-length in vacuum corresponding 
to r), while Nernst’) from the specific heat, deduces the value 
a = 162 corresponding to 2 = 90.10-4, a number, however, which 
is not of itself sufficient to represent the whole behaviour of silver. 
For lead, Nernst gives a = 58, while Einstein gets a = 104 from the 
elasticity. Moreover according to the elasticity « should increase 
somewhat at lower temperatures, while from the specifie heat, it 
would appear that the change should take place in the same sense 
as that in which the resistance changes This, too, shows that the 
theory is still far from perfect. 
As there exists so much uncertainty, and as it is more a question 
of showing that the introduction of vibrators leads to a qualitative 
explanation of the sense in which the observed change of resistance 
deviates from proportionality to temperature, [ have taken for a 
one half of each of the values obtained by Eisrui from the elasti- 
city. In this way we obtain for a: ; 
tor Pt AAT Ae 400 Aur DA ehh 
It appears, therefore, that there is indeed a qualitative correspon- 
dence *). Before we can attach any greater importance to it, howe- 
1) A. Ersrein Ann. d. Phys. (4) 34 (1911) p. 170. Since the address delivered 
in the December meeting was only ready for printing in the number of the Dutch 
Proceedings for February 1911, I have been able to add then the following calcu- 
lations from the elasticity to what 1 communicated in December. 
2) Cf. also Mapetune, Gött. Nachr. 1909, p. 100, who was the first to calculate 
the period of molecular vibrations. 
5) The numbers are all taken from the Leiden observations (KAMERLISGH ONNES 
and Cray |. c.) and they refer to the purest of the wires, while the probable 
negative correction for the influence of admixture and for the results of treatment 
during manufacture necessary for its expression in terms of the pure metal have 
