is«> ,<?. In a similar manner, the variations in length of one of 

 iitfe© circuits act upon the i7itensity of the current which ema- 

 nates from a constant source. Now it is possible that two 

 currents of unequal intensity may be incapable of interfering 

 in conditions in which they would be mutually destroyed if 

 their intensities had been equal. 



'.73. It must be remarked, that the two first objections do 

 not apply to the method of induced currents. But, in order 

 to deprive them of all value, I have imagined a third, which 

 consists of employing only a single thermo-electric pair, the 

 current of which passes at the same time in part into the wire 

 of the rheometer, and in part into a wire of derivation. This 

 wire was the copper one No. 4, and the pair that of the pl"#» 

 jUduiii wire No. 2 and the copper one No. 3. ? 



i^-Wmii ■■^■' ' '- ■'^■'■■-■'- ■^•■' d^--' •■• — 



^mAfir^- ^ i^^ Msthjlipj Derivations. 



74. When the circuit of a pair is closed with the rheome- 

 ter, three derivations may be effected : — from the platinum 

 wire to the copper one, by thus compelling a part of the cur- 

 rent not to traverse the measuring apparatus ; from the pla- 

 tinum wire to itself, by establishing at certain points of its 

 length junctions with the extremities of the wire of deviation, 

 and causing a variation both in the distance of these points 

 and in the length of this wire ; lastly, from the copper wire to 

 itself, by proceeding in an analogous manner. 



75. The wire of derivation was 0™'003 in diameter, and 

 consequently could not be wound round the rheostat. To en- 

 sure a constant contact on its surface (an extremely important 

 condition), and to cause its length to vary so slowly that this 

 was necessary, I employed a tube of pure copper, 0"^*05 long, 

 with a bore of 0™*0034, and fixed in one of the holes of a 

 Poggendorff's clamp ; its two extremities were split like a 

 porte-crayon, and the lips might be partially closed at will by 

 compressing rings. 



76. H^one of the three derivations occasioned jphanomena of 



interference. .r —i: .i.'r-,.;-^ •*.: .rf,.-<ir ;cav;r -l-.n-'j-d - 



77. The second derivation proved the remarkable action 

 of the conductibility on the intensity of the current, for the 

 copper wire No. 4 being a much better conductor than the 

 platinum wire of the pair, increased the deviation of the nee- 

 dle sensibly. In one of the experiments, by making thejunci 

 tions on the one hand on the platinum wire immediately after 

 its exit from the test-tubes, and on the other with the rheo- 

 meter, the index was propelled 15° from its first position. 

 This increase in the deviation became weaker, but without 

 any intermittence, when the first point of contact was brought 

 near the second. 



