ELECTRICAL RESISTANCE UNDER PRESSURE. 123 



normal. The coefficient is negative, and becomes greater numerically 

 at the higher temperature. The departure from linearity with pres- 

 sure is in the normal direction, the coefficient being less at the higher 

 pressures. The departure from linearity is s\^mmetrical, a second 

 degree curve accurately reproducing the results, so that it is not 

 necessary to draw the de\'iation curves. The deviation becomes 

 greater at the higher temperatures, as is normal. 



I can find no tabulated values for the specific resistance of neo- 

 dymium. The value for the specimen above, obtained from micro- 

 meter measurements of its dimensions, was 107 X 10"^ ohms per 

 cm. cube, about twdce the value for lanthanum. 



Carbon. These results on carbon must be regarded as of an 

 entirely tentative and orienting character. Present technical methods' 

 are not yet sufficiently perfect to permit of the manufacture of carbon, 

 either amorphous or graphitic, of specifiable or reproducible properties. 

 Any massive form of carbon always contains at least a slight amount 

 of binder of unknown properties, amorphous carbon is always impure 

 with a slight amount of graphite, and the purest graphite contains a 

 small and unknown amount of amorphous carbon. 



Experiments were made on three samples of carbon. The first of 

 these was supposedly amorphous cai'bon, an arc carbon made by an 

 unknown German firm. The second was Acheson graphite cut from 

 a piece of graphite furnished by the Acheson Co. for a resistance 

 furnace, and presumably not made with any unusual precautions. 

 The third specimen of graphite was also Acheson graphite, furnished 

 by the Acheson Co., in response to a special request for graphite of 

 the greatest obtainable purity. It was stated by them to contain 

 the minimum of binder, and to have been graphitized with unusual 

 thoroughness, but otherwise its properties w'ere not known. 



In x'lew of the unreproducible character of the results it will not 

 pay to give them in great detail. 



Two sets of readings were made on the gas carbon, a complete run 

 to 12000 kg. and back at 30°, and a few readings at 96°. This speci- 

 men was about 3.5 inches long, and 0.154 inches diameter. Measure- 

 ments were made by the potentiometer method, using the three 

 terminal plug as usual. Connections were made to the carbon with 

 spiral springs snapped into grooves filed around the surface of the rod. 

 A preliminary seasoning was made to 6000 kg., but there was very 

 little permanent change of resistance. At 30°, the resistance decreases 

 with rising pressure, and the direction of curvature is normal, that is, 

 the proportional eftect becomes smaller at the higher pressures. The 



