ELECTRICAL RESISTANCE UNDER PRESSURE. 141 



instantaneous pressure coefficient increases with rising temperature 

 and decreases with rising pressure by amounts which, for the range 

 of this work, may amount to a factor between 2 and 3. The tem- 

 perature coefficient of resistance decreases with increasing pressure 

 by very perceptible amounts, and decreases much more for potassium 

 than sodium. 



The decrease in temperature coefficient at the higher pressures 

 is especially significant. One might perhaps expect that at higher 

 pressures the metal would be compressed into an approach toward its 

 behavior at 0° Abs under atmospheric pressure, since the volume 

 may be reduced by pressure to less than its value at 0° Abs. Now as 

 the absolute zero is approached at atmospheric pressure the tempera- 

 ture coefficient of resistance becomes much greater than the reciprocal 

 of the absolute temperature; this is the exact opposite of the behavior 

 found above at high pressures, the temperature coefficient becoming 

 less. The efEect of increasing pressiu-e is seen to be merely that of 

 making the part played by temperature of less and less relative import- 

 ance, which is after all not unnatural from a certain point of view. 

 In the absence of specific information to the contrary it is natural to 

 connect the unusual behavior of sodium and potassium with the large 

 change of volume, and to expect that other metals will show the same 

 sort of behavior imder correspondingly increased pressures. 



The non-metals Si and P would not be expected to agree in behavior 

 wuth the metals, but it is interesting, nevertheless, to summarize their 

 behavior. The magnitude of the mean coefficient of Si is about the 

 same as that of lead. The coefficient may increase very largely with 

 increasing temperature, however, and also may apparently increase 

 with increasing pressure. This is quite contrary to expectations, 

 and would seem to indicate an approach to some sort of instability 

 at high pressures; perhaps as the atoms are pushed more closely 

 together there is an approach to metallic conductivity. The varia- 

 tions of the temperature coefficient of silicon are also abnormal. 

 Initially the coefficient is normal in sign, but small numerically; as 

 pressure is increased it reverses in sign. This reversal in sign of the 

 temperature coefficient, unlike the l^ehavior of the pressure coefficient, 

 does not indicate an approach to metallic conductivity. Too great 

 weight should not be attached to these results, because the silicon was 

 impure. HoweA'er, it is evident that there are some interesting possi- 

 bilities here, and the measurements should be repeated when it is 

 possible to obtain purer material. 



Black phosphorus is remarkable for the great magnitude of the 



