On Tkevmo-Electro-magnetism. 127 



of the hexagon the direction of the magnetic needle. A compass 

 needle was then placed below this side, and as near to it as possible. 

 On heating one of the solderings with the flame of a lamp, they 

 produced a very sensible effect on the needle. On heating two 

 solderings, not contiguous, the deviation became, considerably 

 greater. When, lastly, the temperature of the three alternate 

 solderings was heated, a still greater effect was produced. They 

 likewise made use of an inverse process, that is to say, they re- 

 duced to zero, by melting ice, the temperature of one, or more 

 solderings of the circuit. It is readily conceived that, in this case, 

 the solderings which are not cooled must be regarded as heated 

 in reference to the others. This manner of operating allows 

 the different experiments to become comparable; otherwise the 

 laws of this class of phenomena could not be discovered. By 

 combining the action of the ice with that of the flame, namely, by 

 heating the three solderings that are not refrigerating, they arrived 

 at a very considerable effect indeed ; the deviation of the needle 

 amounted then to 60°. 



They afterwards continued these experiments with an apparatus 

 composed of 22 bars of bismuth, and 22 of antimony, much 

 thicker than those of the hexagon ; and became convinced that 

 each element contributes to the total effect. Having opened the 

 circuit in one point, they soldered to the separated bars, small brass 

 cups, which were subsequently filled with mercury, in order to 

 establish at pleasure a sure communication between their extre- 

 mities by means of metallic wires. A copper wire, a decimetre in 

 length, and a millimetre in thickness, was nearly adequate to 

 restore the complete communication. With two similar wires 

 placed alongside of each other, the communication was perfect. 

 A wire of the same diameter, but more than a metre long, trans- 

 mitted the current pretty well ; while a wire of platinum, half a 

 millimetre in diameter (about ■£$ of an inch), and 4 decimetres 

 long, established the communication so imperfectly, that the 

 deviation of the compass-needle did not amount to 1°. When 

 the interposed body was a slip of paper moistened with a satu- 

 rated solution of soda, no appreciable effect was observed. It is 

 worthy of remark, that an apparatus capable of affording electro- 

 magnetic effects of such magnitude, produced no sensible chemical 

 action or ignition. They further add, that the effect of the com- 

 plex electro-magnetic circuit, is much inferior to the sum of the 

 insulated effects, which the same elements could produce when 

 employed in the formation of simple circuits. 



Details of the Experiments of the preceding note, and ulterior 

 Observations. — The bars made use of in the following experiments 

 were parallelopipeds, having for their transverse section, a square 

 15 millimetres iu each side (about 0.6 of* an inch square.) They 



