( 449 ) 
Shida: 
Consequently, the coefficient of conductivity has the value 
8alAa 
— EA Lea Geena MoS (5) 
Qh? (28) 
8 EN 
rei Finan, add) 
Drupe’s result for this case is 
or 
1 
k==I Nau. 
3 
The ratio of the two conductivities is by my formulae 
k 8 z 
w=5(*)r- vir SEREN 
0 9 \e 
k >(<) 
—==| =| 7. 
0 3 \e 
Here again, the difference between the two formulae consists 
merely in the numerical coefficients. 
and by those of Drupe 
k 
Just like Drepe we may therefore conclude that the value of — 
o 
does not depend on the nature of the metal and that it varies pro- 
portionately to the absolute temperature, consequences that have been 
verified with a certain approximation in the case of many metals. 
It need hardly be observed that these conclusions could only be 
arrived at because we have neglected the mutual encounters between 
electrons *). Im fact, these would tend to diminish the conductivity 
for heat, but not that for electricity, since they cannot have an 
influence in a phenomenon in which all electrons move in the same 
way. It is clear that, under these circumstances, a value of “ inde- 
pendent of the nature of the metal could hardly be expected. 
Let us next consider the absolute values. 
ia al 
The value of that can be deduced from those of / and o and 
é 
for which, using (28), I find 
yp 9k, 
eee De shat ie BEES) 
e 86 
1) See Tuomson, l.c., p. 146. 
Proceedings Royal Acad. Amsterdam. Vol. VII. 
