] 32 Professor Forbes on the Refraction 



which may be carried off by wires, and caused to act upon a de- 

 licate galvanometer or multiplier, the needle of which serves as 

 an index ; the galvanometer consisting, of course, of a magnetic 

 needle, nearly freed from the influence of the earth's magnetism, 

 and so connected with the wire which transmits the electricity, 

 that the mutual influence of the magnetism and the electricity 

 shall (by the law of Orsted) be a maximum. 



3. It will readily be conceived, that, if a series of alternating 

 bars of bismuth and antimony be placed parallel to each other, 

 and the extremities alternately soldered together, when all the 

 extremities facing one way are heated (as by the radiant influ- 

 ence of a lamp), whilst the others remain at the temperature of 

 the apartment, the effects produced in a single pair, such as we 

 first supposed, will be produced at each junction, and that the 

 intensity of the whole effect will be greater, just as in the Voltaic 

 pile. At one time it appeared doubtful how far electricity, of 

 such small tension as is thus produced, could be so reinforced ; 

 but the instrument in question seems to prove the practicability 

 of it. About thirty pairs are employed, and so delicately are they 

 made, that the ends which exhibit one set of junctions are con- 

 tained within a superficial area of four-tenths of a square inch. 



4. The wires, from the extremity of the first and last element 

 (just as in the Voltaic battery), convey the electricity to the mul- 

 tiplier, which consists of a flattened coil of silver- wire, covered 

 with silk, the coils of the wire being parallel to the quiescent po- 

 sition of an astatic magnetic needle, which is perpendicular to 

 the magnetic meridian. The deviations are measured in the 

 usual manner, on a divided circle ; upon which, with practice, a quar- 

 ter of a degree may always be observed, and even minuter quan- 

 tities occasionally estimated. These divisions are not necessari- 

 ly proportional to the intensities of the currents which produce 

 the corresponding deviations. The coils of wire, extending a 

 long way on each side of zero, prevent the effect from diminish- 

 ing so rapidly as if they were concentrated there ; and M. Mel- 



