442 M. Oersted on [June, 



altered, when the length of the circumference increases in the 

 same proportion as the number of the elements ; or, in other 

 words, elements of equal length form circuits which produce 

 equal deviations, whatever may be the number of the elements. 

 We confirmed these results by comparing the effects of two, 

 three, four, six, thirteen, and twenty-two elements. 



In order to form complex circuits capable of producing a 

 great effect upon the magnetic needle, it will be necessary to 

 employ very short elementary bars ; and to avoid the inconve- 

 nience which follows from the restoration of the equilibrium of 

 temperature which happens too rapidly in such small circuits, 

 the solderings must be placed alternately in contact with conti- 

 nued sources of heat and cold. There is still another increase 

 of effect in the complex thermoelectric circuit, which is not thus 

 limited by the length of the circumference ; but before it is men- 

 tioned we shall show the relation which exists between the 

 different elements of the complex circuit. 



Exper.5. — We examined the effects of the circuits by cooling 

 first one, then two, afterwards three, &c. of the solderings 

 which were rendered active ; and after several experiments, we 

 found the mean numbers to be as follow : In a circuit of three 

 elements, the first gave a deviation of 1 &§-° ; the first two 25i° ; 

 the three together 31°. In a circuit of four elements, the ice 

 placed upon one soldering gave a deviation of 13£° ; upon two 

 19°; three 25°; four 31f°. In a circuit of six elements, the 

 first gave a deviation of 9° ; the first two 13^-°; the first three 

 18i°; the first four 22° ; the first five 25^°; the six together 

 28f°. 



It will be observed that the deviation produced by the first 

 cooled soldering, is nearly represented by double the quotient 

 obtained by dividing the total deviations produced by the cir- 

 cuit, when all its elements are put in activity, by the number of 

 elements plus one. It is also evident that the other numbers 

 nearly approach the value of the simple quotient ; but still they 

 appear to form a decreasing series. We are now alluding to the 

 deviations measured by the angles, and not of the real extent of 

 the effects. If it were not necessary to regard the different dis- 

 tances of all the points which act upon each other in the differ- 

 ent positions of the needle, and even to consider what may be 

 the reciprocal situation more or less oblique of the edges of the 

 conductor and of the needle, the effects might be represented 

 by the tangents of the deviations. It is, however, remarkable, 

 that the experiments which we have made indicate a constant 

 relation between the deviations. If such experiments as we 

 have hitherto had an opportunity of performing were susceptible 

 of greater exactness, consequences interesting to the theory 

 would undoubtedly arise from them. 



Expe?\ 6. — Thermoelectric action may be rendered sensible 

 by means of the electromagnetic multiplier. In order to produce 



