Producible by Attrition and Contact of Metals. 25 



When the rods employed are antimony and German silver the case is diffe- 

 rent : either rod heated at one end only, and applied to the other cold, will cause 

 western deflection. But when the antimony is heated throughout, it gives with 

 cold German silver a feeble eastern deflection. 



It is a singular circumstance that if two rods of bismuth be used, one on 

 each side of the galvanometer, either rod heated throughout and rubbed to the 

 other, will cause an extensive swing of the needle, but if the same rod be heated 

 at one end only, it will send the needle in the opposite direction. If two rods 

 of antimony be similarly treated, the heated rod, whether it is so partially or 

 entirely, will cause eastern deflection by attrition to the other. 



In order to try the effect of bending a rod into a circular form and uniting 

 its ends, I caused a thick ring of bismuth to be cast: it weighed twelve ounces; 

 its diameter was nearly five inches. This being connected with the zinc side, 

 and a mass of antimony with the silver side, it was found that by heating about 

 an inch of the bismuth, and rubbing it against the cold antimony, western de- 

 flection resulted ; but when the whole ring was equally heated, the same con- 

 tact caused the needle to pass to the east. Thus the ring acted in the same 

 manner as a rod. 



By comparison of the foregoing experiments it will appear that the deflec- 

 tions produced by rods are frequently different from those caused by hemisphe- 

 rical masses. The experiments require much circumspection ; many attempts 

 may be necessary to obtain true results. The least difference in the heat of the 

 two ends of a rod, when they are intended to be equal, will sometimes give a 

 false deflection. 



It thus appears to be true, with regard to this law, that metals divide them- 

 selves into two classes:— 1st. Those which, when associated with certain other 

 metals, give deflections on opposite sides of the magnetic meridian, according as 

 they are heated partially or totally. 2nd. Those which do not so comport 

 themselves when associated with the former class, but give a deflection on the 

 same side of the magnetic meridian, whether they be heated partially or totally. 

 The second class of metals only exist as such, so long as they are associated 

 with the first: for when associated with each other, they change their charac- 

 ter and act similarly to the first class. The existence of the two classes is, 

 therefore, not absolute: it is a relative condition subject to the contingent asso- 



VOL. SXIII. E 



