by ThermO'electric Currents, 411 



varies somewhat considerably, owing to circumstances which 

 it is impossible to foresee and guard against, so that it is ne- 

 cessary to make this table very frequently. Frequently in the 

 midst of a series of experiments a perceptible alteration may 

 occur, and the experimenter is always uneasy on this point. 

 Moreover, the sensitiveness of the galvanometer decreases 

 rapidly with the amplitude of the deviations; and it must not 

 be used to measure deviations of more than 60°, because 

 beyond that limit the indications of the instrument become 

 very uncertain. 



A part of these inconveniences is obviated in the apparatus 

 to which M. Pouillet has given the name of sine compass. 

 In this instrument the needle is simple, and its magnetic axis 

 is always placed so that it shall coincide with the direction of 

 the current; the intensities of the current are then propor- 

 tional to the sines of the angles which the magnetic meridian 

 forms with the direction of the axis of the needle. To be 

 convinced at any time of the identity of the measuring ap- 

 paratus, it suffices to pass through the wire a current of con- 

 stant intensity, and easily producible always of the same 

 strength; if the needle indicates the same deflexion, we are 

 certain that the apparatus has remained comparable. 



The sine compass should not be employed to measure 

 currents which produce deflexions of more than 50° to 60°; 

 because beyond these limits the sines increase but very slowly 

 for considerable variations of the arc, and the compass becomes 

 useless. Thus all the measures on the sine compass should 

 be comprised between 0° and 60°, and correspond to the in- 

 tensities of the thermo-electric current between the limits of 

 temperature which it is desired to measure. It thence results^ 

 that if we wish to determine elevated temperatures, it is requi- 

 site to be content with somewhat weak deviations for a differ- 

 ence of 100°, and the apparatus becomes not very sensitive. 

 Thus, in the magnetic pyrometer of M. Pouillet, the compass 

 indicated a deviation of 4° to 5° for a difference of temperature 

 of 100° of the two solderings of the thermo-electric element. 

 It is true, that, by giving a sufficiently large diameter to the 

 divided circle, and observing the deviation with the assistance 

 of a vernier, the subdivision of the degree may be carried as 

 far as wished ; and it is easy, if the divided circle has a dia- 

 meter of 10 to 15 centimetres, to apjjreciate angles of 1'. The 

 deviation produced by a difference of temperature of 100° 

 would consequently be measured to within ^^^ ; that is to 

 say, with more than sufficient accuracy. Unhappily, the mag- 

 net needle is far from exhibiting any such sensitiveness. In 

 the ordinary compasses the needle is supported on a pivot ; and 



2E2 



