APPENDIX TO MEMOIR OF PELTIER. 



187 



a l)attcry of six pairs of tlie same dimensions^ tlie same results were obtained, 

 and not a degree more. 



If, on the contrary, the interposed arc conducts feebly, the angular deviation 

 is no longer proportional to the surface immersed In another series of experi- 

 ments, Peltier caused a cuiTcnt to pass into a trough full of water, in which he 

 could interpose, at pleasure, diaphragms of platina, and he reached the follow- 

 in o- results : 



The galvanometer employed in this experiment was an instrument of 430 coils. 



The inspection of this table suffices to remove all doubts : as long as there 

 was no diaphragm, two pairs were sufficient to give 40 degrees of galvanometric 

 deviation, equivalent to 102 of force ; when there were two diaphragms, five 

 pairs were needed to arrive at the same angular deviation. When there A\as 

 one diaphragm, three pairs gave 32° = 48.5 of forces ; with two diaphragms 

 there was but 24° = 25 ; to regain the 32° it was necessary to employ four pairs. 

 With three diaphragms there resulted for three pairs only 20° =20 ; for four 

 pairs only 26° = 28. To regain or nearly regain the 32° = 48.5, it was neces- 

 sary to employ live pairs. In effect, by taking three pairs, there resulted, with 

 one diaphragm, 32°=48.5, with two diaphragms 24° = 25, with three dia- 

 phragms 20°=20. Thus the quantity of the current continued diminishing in 

 proportion as the resistance of the conductor augmented. Further, to regain that 

 quantity, it sufficed to increase the number of pairs ; then, indeed, the resistance 

 of the conductor was overcome and the same quantit\^ of electricity })assed anew. 



The inspection of this table proves, therefore, that to have the same number 

 of degrees after a different number of alternatives, it is necessary to modify the 

 electric source, and that the same deviation can never be reproduced after the 

 addition of a diaphragm, if the number of pairs be not augmented. The table 

 shows, also, that the loss of the current is so much less as the current has 

 already traversed a greater number of diaphragms. Thus, we find in the second 

 line for two pairs 102, 21.2, 14, and 12. The first diaphragm, therefore, has 

 caused the current to lose |- of its quantity ; the second, § ; the third, i. It is 

 not, as has been said, that the electricity, better sifted, passes more easily 

 through the new obstacles opposed to it ; the electricity has not changed its 

 nature, but it is that after having traversed, say two diaphragms, if a third be 

 presented to it, it has, in order to retrograde, to surmount anew the resistance 

 of the first two diaphragms ; it is no longer simply the obstacle of the battery 

 which opposes itself to its equilibration in returning, there are besides the two 

 diaphragms which it has already passed. From this it results that the more 

 diaphragms the current has traversed, the more resistance it finds in its return, 

 and the less loss it sustains consequently by the interposition of another dia- 

 phragm. 



From Avhat precedes we shall readily comprehend the gravity of the error 

 committed by physicists, and especially by the German physicists, who, in their 

 experiments on currents, in general only consider the current itself, and take 

 little or no account of the electro-motor. A current, however, is not an ideal 

 existence Avhich can be divorced from the -source which gives rise to it. 



Ohm and Gauss have, in their formulas, recognized as a principle that metallic 

 threads oppose to the passage of electric currents a resistance always directly 



