four inches in length, balanced on a very fine pivot, at the centre of a circle divided into semi-degrees, is placed in a box, furnished with three adjusting-screws, so as to secure a horizontal position. Instead of the usual multiplying coil, which would too much weaken the current, there is only a single brass wire, of 1¼ line in thickness, which passes exactly through the centre below the needle. The extremities of this wire are furnished with two small cups containing a few drops of mercury, and which allows the circuit to be easily completed. The bottom of the box consists of a glass plate, not a metallic one, as it is usually made; a precaution which I have thought necessary for magneto electric experiments. This galvanometer is nevertheless so sensitive as to be affected by very weak electric forces. Although the deviations of the needle cannot serve as an exact measure of the force of the current, they however indicate the slightest variation the battery has undergone during the progress of the experiments. By employing a single voltaic pair, of small dimensions, as Professor Parrot has done, there are many means of restoring, at any time, the original state, and of having a constant force. Lastly, with respect to the importance of these experiments, it would be no great loss to reject a pair of plates which have become worn, and to substitute fresh ones in their place. Moreover, we may convince ourselves that it is far more necessary to keep account of the temperature of the conducting liquid than to clean the pair after each experiment, &c. But whatever pains we may take to operate under similar circumstances, we must recollect that the experiments cannot but be incorrect, unless we employ the galvanometer.
The most sure and exact method for these experiments, is to try the magnetic power of two or three bars at a time. The helices are united into a single conjunctive wire so as to submit these bars constantly to the influence of the same current. By interposing a galvanometer, which is indispensable, any variation in the action of the battery may even be turned to account; for by this means, we learn what relation exists between the relative power of the different portions of the circuit and the intensity of the current. Trifling differences in the construction of the helices may be removed by some method of elimination or by reciprocal combinations. The details into which I have entered will, I trust, be excused, as it is often impossible to
