294 C. G. KXOTT. 



Curves 1 and 2 luno already l)een mcntii^ried. Tliey sliow tlie 

 mareli of resistance with temperature as measured on a mercurial 

 centigrade thermometer. \\i Xo. 3, tlie ])latiiiiim resistances are vir- 

 tnally used as temperatures, and form the al)scissa?. The ordinate« 

 are tlie corresponding cohalt resistances. The points belonging to 

 the various experiments are distinguished by special mark. 



It will he seen at a glance tliat in ovie ])articular cobalt behaves 

 very like iron and nickel. There is a ra])id increase in the steepness 

 of the curve at the higher temperatures. Fn iron and nickel this 

 rapid increase is followed at still higher temperatures l)y a distinct 

 decrease, the curves bending so as to present a concavity toward the 

 temperature (or platirium resistance) axis, Ta.ble IV. gives no hint 

 of such a tendency in cobalt. The curves all iDCCome steeper with 

 rise of temperature, if we except the distinctly irregular indications of 

 Experiment TA'. 



It Avill be seen from Table TV. that Experiments T. and H. are in 

 fiir agreement throughout ; arid that all four experiments point to the 

 existence of a critical temperatm-e, at which the resistance begins to 

 increase rapidly with rise of temperature. This critical temperature 

 is al)out the stage 1.5, which corresponds approximately to 350° C. 

 Tlie same conclusion may l)e drawn from Table IE and expressed in 

 these terms. Between the temperatures 400° and 700° C. the resist- 

 ance of a collait strip increases on the average at a rate nearly twice as 

 great as the average rate of increase loetween 0° and 300° C 



2. The Thermoelectric Positions of 

 Cobalt and Bismuth. 



So fir as I know, the only satisfictc^ry determination of the 

 position of the Cobalt line on the thermoelectric diagram was made by 



