or, since the wire in such case is merely necessary to conduct the electricity, 

 the phenomena of terrestrial magnetism only require the admission that a 

 series of electric currents continually circulate round the globe, according to 

 lines which intersect the magnetic meridians perpendicularly. 



To present an experimental verification of this theory, M. Ampere construct- 

 ed a plane geometrical figure a circle, for example of wire, and suspended it 

 in such a manner that, while the current circulated upon it, the figure was ca- 

 pable of moving on a vertical axis through its centre of gravity. It was ob- 

 served to throw its plane into a position at right angles to the magnetic me- 

 ridian. When the current was reversed, it turned round through one hundred 

 and eighty degrees, and reversed its plane. When a helix was suspended on 

 its centre of gravity, and a current was transmitted through the wire, it exhib- 

 ited all the properties of a magnet ; when suspended on a vertical axis, it as- 

 sumed the direction of the magnetic meridian ; and when suspended on a hori- 

 zontal axis at right angles to the magnetic meridian, it threw itself parallel to 

 the dipping needle. 



The hypothesis of Davy, that the nucleus of the globe consisted of the me- 

 tallic bases of the alkalies and earths, and that its surface was oxydated, sup- 

 plied Ampere with strong grounds of probability in support of these theoretical 

 ideas of terrestrial magnetism. It was easy to imagine that, at the surface of 

 contact of the metallic nucleus and the surrounding shell of oxydated matter, 

 there were constant chemical actions in progress, which might produce a se- 

 ries of electric currents at some distance below the surface of the earth, and 

 that these currents, acting through the shell of oxides, would produce the phe- 

 nomena of terrestrial magnetism. 



In the same year, M. de la Rive, of Geneva, published a memoir, in which 

 he showed that when a current is transmitted through a closed circuit of a rec- 

 tangular form, for example, it affected only the sides which have a vertical po- 

 sition. He established, as a general law, that a vertical current, capable of 

 revolving round a fixed vertical line as an axis, will place itself so that the 

 plane passing through its own direction, and the axis round which it revolves, 

 shall be at right angles to the magnetic meridian, the side on which the cur- 

 rent descends being on the east of the axis, and the side on which it ascends 

 being on the west. 



He also showed that a horizontal current, though not susceptible of being 

 influenced by the magnetism of the earth, is not therefore free from all action ; 

 on the contrary, he proved that when it is free to move parallel to itself, it will 

 move in this manner in the one direction or the other, according to its own di- 

 rection : and that this motion will equally ensue in all positions in which it 

 may be placed, whether it be directed north and south, east and west, or in any 

 intermediate azimuth. 



These laws, proved experimentally by M. de la Rive, were immediately 

 shown by M. Ampere to be direct consequences of his theoretical principles. 



In the year 1827, M. Savary directed his labors to follow out the researches 

 on the power of the Voltaic current to impart magnetism to iron, which had 

 been demonstrated by the experiments of Davy and Arago. M. Savary dis- 

 charged a Leyden jar through a metallic wire, needles placed near which were 

 found to be magnetized, and the strength of the magnetism imparted to them 

 was observed to vary with their distance from the wire. Being placed at va- 

 rious distances from it, the magnetizing power of the current was not found 

 either continually augmented, or continually decreased ; but, as the needle re- 

 ceded, it first increased, and then diminished, attaining a maximum at a certain 

 position. He also showed that as the distance varied, not only the intensity 

 of the magnetic force passed thus successively through maxima and minima, 



