AS A MEASURING INSTRUMENT. 397 



same source through the same wire ; but currents produced by one 

 or more sources may also be made to act upon the galvanometer by 

 several wires. This process requires a galvanometer with two or 

 more wires. In other respects, the manipulation is the same as with 

 the former process, but since the number of the wires, and, conse- 

 quently, that of the currents which can be employed at once is limited, 

 it requires several series of the latter to embrace the whole quadrant. 

 It should be so arranged that the currents of one series shall produce 

 the same effect as a certain number of currents of the next previous 

 series, so that the several partial series may be combined into a gene- 

 ral one. Both Becquerel and Nobili* employed this process. 



Melloni based his admirable experiments on the method by differ- 

 ences. After satisfying himself that the deflection of his galvano- 

 meter was proportional to the force of the current within the first 

 twenty degrees, he determined the relation between these elements 

 beyond this limit in the following way :f He attached to the galvano- 

 meter a thermo-pile. and warmed one of its sides by bringing a spirit 

 lamp near it until a deflection of 20° was produced. He then placed 

 a screen before the lamp, and waited until the needle returned to 0°. 

 He next allowed the radiant heat from a second lamp to fall upon the 

 other side of the thermo-pile, and regulated the distance of the latter 

 so that a deflection of 24° in the contrary direction was produced. He 

 finally allowed both lamps to act simultaneously on the pile. lie 

 now obtained by the difference of the currents a deflection not of 4°, 

 but of 5.1°; he therefore concluded that the current which, from 

 producing a deflection of 20°, he made equal to 20, must have in- 

 creased 5.1 units to produce a deflection of 24°, and must therefore 

 be 25.1. Increasing the activity of the pile by advancing first one 

 lamp and then the other toward the pile, he determined in the same 

 way the increase of the force of current for the intervals 28° — 24°, 

 32 — 28°, <fec. After thus fixing a sufficient number of.points in his 

 scale of intensity, he filled up the intervals by hand. 



Nobili had already used the same process with a galvanometer with 

 two wires, through which be passed currents from two batteries in 

 opposite directions — first separately, and then together. 



All of these methods are liable to considerable objections. Although 

 not so tedious as a theoretical determination of the scale of intensity 

 would be, they are still troublesome. They are, moreover, based on 

 conditions which are difficult to fulfil, and for whose fulfilment we 

 b ve no control. It is theoretically true that the strength of the 

 rent in a thermo-pile is proportional to the number of the pairs of 

 p : set in action by equal differences ; but we have no certainty 

 ; the difference of temperature is the same in all the pairs, or that 

 the aie difference would produce the same effect in all. 



loni's method is probably the best of those named. But if the 

 ob ' ation of Becquerel,:): that equal variations of temperature pro- 

 die ■• hermo-currents of different intensity, according to their position 



e Pugg Annal. Bd. ix, S 346 and xx S. 226. 

 f Idem, Bd. xxxv, S 132. 

 | Idem, Bd. ix,S. 350. 



