326 Prof. J. W. Mallet on the Weight of a Wire placed 



of wires formed the fourth side ; these sides were laid down 

 on the floor north and south, east and west, respectively, by 

 compass. In the course of the experiments the frame was 

 shifted round so as to make the movable wires occupy in 

 succession the position of each of the four sides. The gal- 

 vanic battery (Grove's) was placed in an adjoining room; and 

 its polar wires communicated with those of the frame by tw^o 

 single mercury-cups. Ten cylindrical cells were used, in each 

 of which the immersed platinum surface was about 140 square 

 centimetres for each side of the sheet. The current came in 

 at the middle of the side of the square opposite to the sus- 

 pended wires, passed round to the first of the series of mercury- 

 cups by one of the half-length wdres, thence along the first of 

 the ten suspended wires, back around three sides of the square 

 to the second mercury-cup, thence over the second suspended 

 wire, and so on, until it returned to the battery by the second 

 half-length wire, leaving the frame close to the point at which 

 it came in. The whole arrangement of wires was, in fact, a 

 spiral coil of ten turns, distorted into a square, laid down in a 

 horizontal plane, and having one of its four sides free to 

 move up and down vertically while maintaining the current 

 unbroken. 



With this apparatus it was found that, when the movable 

 wires were made to run east and west, whether on the northern 

 or southern side of the square, the side of the balance to which 

 they w^ere attached sensibly preponderated if the current (as 

 usually defined) was passed from east to west, and the oppo- 

 site side w^ent down if the current was sent from west to east. 

 The motion w^as not simply a momentary impulse, but the 

 disturbance of equilibrium was permanent while the current 

 was maintained. The amount of motion was decidedly greater 

 in the direction of repulsion, or apparent loss of weight, than 

 in the opposite case. In both cases the amplitude of swing 

 of the balance-index was very small, for the obvious reason 

 that a very little rise or fall of the system of wires introduced 

 a serious difference of w^eight in the shape of disturbed buoy- 

 ancy of the copper depressed into or lifted out of the mercury. 

 The attempt was made to ascertain what weight would bring 

 back the balance-index to zero after each disturbance ; but the 

 measurements were not very satisfactory, in consequence of 

 the clogging of motion by the immersion of the ends of amal- 

 gamated copper wire in mercury and the ease with which 

 mercury was taken up or lost in varying amount on their 

 surfaces. The average of the best determinations obtained 

 when the galvanic battery was in full activity was about 



