230 O. U. VONWILLER. 



it was found on cooling that the resistance between B an$ 

 O did not return to the original value. The results too, 

 were very irregular, sometimes the resistance was lowered, 

 at others raised as a result of heating and subsequent cool- 

 ing. Even when the same specimen was repeatedly heated 

 and cooled over the same range of temperature, with or 

 without the passage of current at the higher temperature, 

 it was found that sometimes for two or three trials the 

 resistance would decrease after each cycle and after the 

 next it would increase. In all cases, however, while the 

 temperature was constant at any value the curve for B O 

 was a straight line. In the hope of obtaining a value for 

 the temperature coefficient trials were made with widely 

 different conditions; sometimes very thick pieces of the 

 material were used and at others very thin strips, less than 

 0*1 mm. thick; the contacts were altered in nature, con- 

 sisting sometimes of clamped metal strips, at others of 

 wires with rounded ends pressing against the molybdenite, 

 and sometimes the molybdenite was plated with copper 

 deposited electrolytically at the regions where the contact 

 was to be made, but consistent results could not be obtained. 

 It is probable that the change in conductance between B O 

 after going through a temperature cycle is not due to a 

 real change in resistivity but is caused by a mechanical 

 change. The material has a lamellar structure and in the 

 expansions and subsequent contractions a certain disturb- 

 ance may occur resulting in a change of contact between 

 the sheets making up the strip, and if there is discontinuity 

 in some of these the effective area of the section and there- 

 fore the conductance may be changed. 



Inset in figure 4 is figure 5 showing the relation between 

 apparent conductance and temperature for B C and A D. 

 The ordinates in the case of AD have been multiplied by 

 five and the numbers given on the y axis are merely pro- 



