5o8 
PHYSICS: P. W. BRIDGMAN 
Proc. N. a. S. 
of crystalline structure in determining the variations of resistance. On 
the other hand the coefficient of lithium is abnormal in being positive in 
both the liquid and solid states. (The melting curve of Li was measured 
under pressure and found to be normal in that the solid expands on melt- 
ing.) The magnitude of the coefficient of the solid is less than that of the 
liquid. Gallium is normal in the solid as well as in the liquid, and the 
coefficient of the liquid is much larger than that of the solid. The coeffi- 
cient of liquid mercury is slightly less than that of solid mercury. The 
coefficients of liquid sodium and potassium are less than those of the 
solids, but by only small amounts. The relative decrease in the coeffi- 
cients of these solids with pressure may be greater than the decrease of 
the liquid. Liquid potassium has an abnornal change in the sign of the 
variation of temperature coefficient with pressure and the pressure coeffi- 
cient with temperature. 
All of the six metals measured in the solid and liquid state agreed in 
showing a relatively small change in the ratio of the resistance of liquid 
to solid along the melting curve from low to high pressure. Thus the 
change in the ratio of the resistance of liquid to solid potassium was 
from L56 to 1.55 under a pressure increase of 9700 kg. This pressure is 
sufficient to decrease the difference of volume between liquid and solid 
to 0.31 of its initial value. It seems to be universally true that the tem- 
perature coefficient of the liquid is less than that of the solid, and the 
change of resistance on melting is in the direction of the change of volume. 
Theoretical Implications. — The following remarks are only two of many 
that might be suggested by the data. 
It is probable, because of the peculiar nature of the atomic structure 
of lithium and the fact that its pressure coefficient of resistance is pos- 
itive, that the picture of the mechanism of electrical conduction given 
by Wien and Lindemann as a space lattice of electrons sliding in the 
channels of a space lattice of atoms may have much of truth for this 
particular element. 
The new evidence now at hand does not make it necessary to modify 
the former point of view^ that for most elements the most important 
single factor in determining the variations of electrical resistance is the 
amplitude of atomic vibration. 
This investigation was assisted in large part by a generous grant from 
the Rumford Fund of the American Academy of Arts and Sciences. 
1 Bridgman, P. W., Proc. Amer. Acad., Boston, 52, 1917, (573-646). 
2 Bridgman, P. W., Physic. Rev., Ithaca, N. Y., 9, 1917 (269-289). 
