CONDUCTION OF ELECTRICITY IN MOLYBDENITE. 2 2 9 



tions (it varied by 2 or 3 degrees for the three highest 

 temperatures but by not more than a degree in the others), 

 and the sign + indicates that the direction of current was 

 from A to D while the sign — indicates a flow on the 

 opposite direction. The numbers along the x and y axes 

 must be multiplied by the same constants as in figure 2 to 

 give amperes and volts respectively. In the case of figure 

 3 we see that the total resistance, proportional to the ratio 

 of ordinate to abscissa, decreased to a very marked extent 

 as the temperature rose, but on cooling back to 17° O. after 

 the final heating to 183° O. the curve 17 (II) was obtained, 

 showing a very slight recovery, the resistance being per- 

 manently reduced by the heating. It was found that, even 

 if heated through but a few degrees, on cooling back to the 

 original temperature the resistance was changed, usually 

 being less than before the heating took place. As the total 

 resistance is due mainly to contact resistances at A and D 

 it is not surprising to find that large changes occur on 

 heating, and also that the contacts are not restored to their 

 original condition on re-cooling nor is it surprising that the 

 effects of temperature changes are sometimes very different 

 at the two junctions. The curves in figure 3 show only 

 readings obtained with very small currents ; at some tem- 

 peratures readings were also taken for considerably larger 

 currents and in these cases the full curve shows the same 

 general effect as +AD or - A D in figure 2. 



In figure 4 the curves for BO are all straight lines and 

 a progressive decrease in resistance is seen as the temper- 

 ature rises, the change being relatively much smaller than 

 with the whole resistance AD. On cooling, a line 17(11) 

 is obtained, having a smaller slope than the original line 

 17 indicating a permanent change in conductivity as a 

 result of the heating. This result is more unexpected than 

 is the change in A D, but in practically every test made, 

 even when the change of temperature is but 5 or 10 degrees, 



