astatic. Portions of a brass tube, about an inch long, are to be soldered to the ends of the wires forming the rectangle, for the purpose of holding a small quantity of mercury, to render the metallic contact complete. The annexed cuts exhibit a vertical section of the rectangle, and a horizontal one of its upper side.

The wires, forming the rectangle, are pressed close together, and secured by a waxed sewing-thread, rolled tightly round them. The rectangle is then fixed in a rectangular box, having the upper side formed of two sliding panes of window glass, for the purpose of shutting up the needle from the agitation of the air. Each pane has a small semicircle cut out of the middle of the edge, by means of a round file, so as to leave a circular opening directly above that in the rectangle. Various contrivances for suspending the magnetic needle might be adopted. The following is perhaps the most convenient :-Into a strong wooden sole, or base, fix two upright supports about three feet long. A small stage at the top, having a divided circle on its upper side, and which may be elevated or depressed at pleasure, completes the frame of the instrument. The stage has two holes of the same size as the supports, and at the same distance, with two small screws passing through its sides, opposite the centres of the openings, for the purpose of fixing the stage securely at the proper height. A small cylindrical wooden key or peg, having a small bore in the axis for the purpose of receiving the end of the glass thread, passes through the centre of the divided circle, and is made to turn easily, without much friction.

After numerous trials, the following appears to me the best mode of preparing the threads of glass, so as to have their extremities somewhat thick and tapering, for the purpose of securing them in the torsion key, and in the axis which carries

the magnetic needles. Take a solid rod of glass, or a piece of a clean thermometer tube, having a very fine bore, and draw out one of its ends, as in the annexed cut.

Α

Direct the very point of the flame on the thick portion at a, and pull it out, between the two hands, to the proper length. As it is hardly possible to get a thread of glass of the proper length and fineness, at the first trial, it will be found necessary to draw several, and select the one best adapted to the purpose.

Two slender darning needles, of the best steel, are then to be selected, the eyes to be broken off, and the ends filed to a point similar to the other ends, and then strongly magnetised in the usual way. The needles are then to be fixed transversely in a piece of straw, or other light substance, about an inch long, and at the distance of about half an inch from each other, with their corresponding poles in opposite directions-the one needle intended to be above the upper side of the rectangle and the other below it. One end of the glass thread is then to be securely fixed in the end of the straw, or light axis, by means of strong cement or sealing-wax, whilst the other extremity is fixed, in like manner, in the centre of the torsion key. A single fibre of silk, having a small weight attached to it, is fixed to the lower end of the axis, and made to pass through a small hole near the lower side of the rectangle, for the purpose of keeping the axis carrying the needles, in the centre of the circular opening in the coil. The upper needle has two pieces of fine straw, several inches long, fixed on its ends, so that the slightest deflection may be readily observed. The extremity of one of the straws is made to oscillate between two upright pieces of glass, to prevent the needle moving over an extensive arc, and thus lengthen the time necessary to complete an observation. The whole will be obvious from the simple inspection of the annexed vertical section of the instrument, in which A B is the rectangular coil of wire, NS, S' N', the magnetic needles; C, the stage with the divided circle and torsion key, and G the glass thread. If, instead of the glass thread, the needle be

suspended by a single fibre of silk, the instrument becomes a galvanoscope of extreme delicacy. The following experiment affords a striking illustration of the extreme sensibility of the instrument with this modification.

EXPERIMENT I.

File off a few grains from a piece of zinc and copper by means of a coarse file; place two of these near each other in the bottom of a clean watch glass; bring the clean ends of two fine copper wires, connected with the cups of the galvanometer, in contact with them, and then drop over them a small quantity of dilute acid, and the compound needle will be deflected several degrees.

The instrument by which I ascertained the existence of a Voltaic current from this elementary battery, consisted of a greater number of coils in the rectangle, and the needles were light and strongly magnetised.

Having thus minutely described the torsion galvanometer, I will now shew some of its applications; but before doing this it may be thought necessary to establish its accuracy, not by reasoning (which is already done in the Philos. Trans.), but by direct experiment. The following experiments will shew in a striking manner the perfection of this instrument above those formerly employed.

EXPERIMENT II.

Take two equal rectangular slips of copper and zinc, an inch broad and eight or ten inches long, and divide them into square inches by narrow bands of wax or cement. Solder copper wires to their extremities, and fix them in a small frame, so that they may always be placed at the same distance from each other. Immerse them in a vessel of water, containing a small quantity of sulphuric acid, to the first horizontal division; turn round the torsion key till the untwisting force of the glass thread balances the deflecting power of the electric current, and note the number of degrees of torsion. Immerse them to the second division, turn round the torsion key as before, and the degrees of torsion necessary to balance the deflecting force of the current, from two square inches, will be found double of those for one square inch. Repeat the experiment with three, four, &c., square inches, and the degrees of torsion will be found to be proportional to the surface of the plates immersed.

Having thus shewn experimentally the accuracy of the instrument, I shall now apply it to determine the power gained by Dr. Wollaston's contrivance of a Galvanic battery above those formerly in use.

EXPERIMENT III.

Having provided a clean slip of copper, two inches broad and about four inches long, I formed it into a rectangle, open

Ꭰ

at the top, and then covered the inner surface of the bottom with cement. A plate of zinc, of the same size with the rectangle of copper, was placed exactly in the middle, having a face of clean copper opposite each of the sides of zinc. Copper wires being soldered to the rectangle of copper and to the plate of zinc, and their ends dipped into the small metallic cups of the galvanometer, the elementary battery was then immersed in very dilute acid, and the torsion key turned till the deflecting force of the battery was vanquished, the number of degrees being about a thousand. Having removed the battery, I covered one side of the plate of zinc and the opposite surface of copper with cement, and repeated the experiment as before; when, as might naturally be expected, the number of degrees of torsion were found to be very nearly five hundred. We may, therefore, safely conclude that the double plate of copper doubles the quantity of electricity without, of course, altering its tension.

Immediately after Ersted's beautiful discovery of the mutual action of magnets and Voltaic conductors, it was known that an immense increase of electro-magnetic power is gained by diminishing the distance between the copper and zinc plates; but, for want of a proper galvanometer, the law does not seem to have been determined with that rigorous accuracy which places its truth beyond the possibility of doubt. To accomplish this was the object of the following experiment.

EXPERIMENT IV.

In order to avoid every source of inaccuracy, I procured a rectangular wooden box, about a foot long, two inches broad, and two and a half inches deep, into which plates of zinc and copper two inches square might be fixed at any distance from each other. Having filled the box with dilute acid, I placed the copper plate at one extremity and the zinc plate at the distance of nine inches, and observed the degree of torsion, as in the preceding experiments. I then untwisted the thread, placed the zinc at the distance of one inch from the copper, and observed the degrees of torsion, which were now nearly three times as great as before. This was next repeated with the