84 Dr. Tyiulall on the Progress of the Physical Sciences : 



fectly homogeneous, something similar may he expected to take 

 place. 



To decide this question, the following experiments were carried 

 out : — A hrass wire 6 feet long was encircled by a number of 

 pieces, each 1 foot in length, cut from the same piece as the 

 6-feet wire; the short pieces were tied round the latter by a 

 non-conducting thread. When the portion of the wire adjacent 

 to the surrounding bundle was heated, no current teas observed; 

 the experiment was repeated with a second wire, but with the 

 same result. 



A brass wire 3 feet long and 3 lines in thickness was so re- 

 duced, that a length of 6 inches of its central portion had a 

 diameter of only half a line (see fig. 3). Both of the points g 

 were heated in succession, but in neither case was a current 

 exhibited. Eighteen inches of another brass wire of 3 lines dia- 

 meter were reduced to - 7 of a line, and each end of the portion 

 thus reduced was screwed into a piece of the thick wire from 

 which it was taken ; on heating the place of junction of thick 

 and thin there was no current. Again, part of a piece of brass 

 wire 3 lines in thickness was drawn out until a diameter of half 

 a line was obtained; both thick and thin portions were then 

 heated to redness, and the oxide carefully removed from the sur- 

 face ; their ends were laid one upon the other and thus heated, 

 but. no current was observed. In all these cases care has been 

 taken to have the thick and thin portions homogeneous ; and 

 we see that when this point is secured, a current never exhibits 

 itself. The mere difference in point of thickness is therefore 

 not sufficient to originate a current, as heretofore believed. 

 M. Magnus explains the knot experiments of M. Becquerel by 

 reference to the fact, that the structure of the wire was first 

 altered by heating it to redness. If the temperature applied do 

 not exceed 100° C, no current is ever observed. 



It has also been affirmed, that the production of a thermo- 

 current is in some measure dependent on the radiative power of 

 the metals employed. A German silver wire was covered by 

 galvanic precipitation with a coating of copper throughout a por- 

 tion of its length ; the wire was heated at the place where the 

 coating ceased, and a tolerably strong current was the conse- 

 quence. Was this result due to a contact of chemically different 

 metals, or to a difference in the radiative power of both ? The 

 wire was next covered with various non-conducting substances, 

 such as soot, gutta pcrcha, wood, &c. ; but when the point where 

 the coating ceased was heated, no current was observed. In like 

 manner, when one portion of a wire was finely polished, and an 

 adjacent portion rendered rough by sand-paper or by the file, on 

 heating the junction between rough and smooth there was no 



