Resistance of Carbon under Pressure. 263 



fourth of its length from the end a groove was filed round it, 

 and around it was bound tightly the end of a clean thin copper 

 wire. It was placed vertically upon a piece of copper in an 

 upright frame, so that pressure could be applied longitudinally 

 at the top ; a flat piece of copper was placed upon the top of 

 it and pressed lightly upon it. The point at which the thin 

 wire was bound round it was then about 18 centim. from 

 the lower end. The upper and lower portions were then 

 connected up with a Wheatstone's bridge, provided, as in 

 Kirchhoff's pattern, with a wire of German silver stretched 

 over a divided scale, the resistances of the two parts of the 

 wire right and left being (when balance was obtained in the 

 galvanometer) proportional respectively to the resistances of 

 the two portions of the carbon rod. The relative, not the ab- 

 solute, resistances of the two portions were therefore being- 

 measured. Two bichromate cells supplied the requisite cur- 

 rent, the galvanometer being a short-coil astatic instrument 

 of simple form. 



On trying the resistances, it appeared that the resistance of 

 the longer part, which was uppermost, was somewhat greater 

 in respect to that of the lower and shorter part than would 

 have been expected from their relative lengths ; the ratio of 

 their resistances being 81 : 19, or about 4-25 : 1, whereas 

 their lengths were as 3 : 1 almost exactly. After taking the 

 rod out of its place and putting it back again under light 

 contact at the top as before, the ratio was found to be 82 : 18, 

 or 4*55 : 1. The rod was thus removed and replaced several 

 times; and the ratio of the resistances was found to differ some- 

 what every time, the figures varying from 4 - 7 : 1 to 3'92 : 1. 

 Pressure was now applied at the top of the rod, and the ratio 

 of the resistances was again measured. With a load of 5 kilo- 

 grammes (as much as it was judged the rod would bear with- 

 out risking breaking it), the ratio of the resistances was found 

 to be much more constant and much nearer to the ratio of the 

 lengths of the two portions, being 75'3 : 24 - 7, which is not very 

 different from 3:1. It was therefore clear that the previous 

 values had been greatly affected by the differences in contact 

 at the two ends; the lower contact having less resistance than 

 the upper, in consequence of the superincumbent weight of 

 the rod and connexions — about 19 grammes in total. 



§ 3. Another rod of Carre's carbon was next examined, and 

 its actual resistance measured in ohms. Its length was 42*6 

 centim., its diameter 0-48 centim. To prepare it for the ex- 

 periment, it was electroplated with a thin coating of copper 

 to the length of about 1 centim. at each end, the extremities 

 being afterwards scraped bare of copper so that end-contacts 

 Y2 



