in a small piece of soft iron, as in the case of the corresponding ammeters, and this in turn may be made to displace an indicating needle over a scale so that corresponding to every given potential difference between the terminals of the instrument there is a corresponding fixed position of the needle on the scale. One of the most useful forms of electromagnetic voltmeter is that generally known as a movable coil voltmeter (fig. 3). In this instrument there is a

Fig. 3.—Round Dial Voltmeter of Kelvin Siphon Recorder, dead beat moving coil type, with front removed. fixed permanent magnet, producing a constant magnetic field, and in the inter space between the poles is fixed a delicately pivoted coil of wire carried in jewelled bearings. The normal position of this coil is with its plane parallel to the lines of force of the field. The current is got in and out of the movable coil by means of fine flexible wires. The movable coil has attached to it an index needle moving over a scale, and a fixed coil of high-resistance wire is included in series with the movable coil between the terminals of the instrument. When a difference of potential is made between the terminals, a current passes through the movable coil, which then tends to place itself with its plane more at right angles to the lines of force of the field. This motion is resisted by the torsion of a spiral spring resembling the hair-spring of a watch having one end fixed to the coil axis, and there is therefore a definite position of the needle on the scale corresponding to each potential difference between the terminals, provided it is within the range of the control. These instruments are only adapted for the measurement of continuous potential difference, that is to say, unidirectional potential difference, but not for alternating voltages. Like the corresponding ammeters, they have the great advantage that the scales are equidivisional and that there is no dead part in the scale, whereas both the electrostatic and electrothermal voltmeters, above described, labour under the disadvantage that the scale divisions are not equal but increase with rise of voltages, hence there is generally a portion of the scale near the zero point where the divisions are so close as to be useless for reading purposes and are therefore omitted. For the measurement of voltages in continuous current generating stations,

Fig. 4.—Edgewise Voltmeter, Stanley D'Arsonval type. movable coil voltmeters are much employed, generally constructed then in the "edgewise" pattern (fig. 4).

Electrodynamic Voltmeters.—A high-resistance electrodynamometer may be employed as a voltmeter. In this case both the fixed and movable circuits consist of fine wires, and the instrument is constructed and used in a manner similar to the Siemens dynamometer employed for measuring continuous alternating current (see Amperemeter). Another much-used method of measuring continuous current voltages of unidirectional potential difference employs the principle of potentiometer (q.v.). In this case a high-resistance wire is connected between the points of which the potential difference is required, and from some known fraction of this resistance wires are brought to an electrostatic voltmeter, or to a movable coil electromagnetic voltmeter, according as the voltage to be measured is alternating or continuous. This measurement is applicable to the measurement of high potentials, either alternating or continuous, provided that in the case of alternating currents the high resistance employed is wound non-inductively and an electrostatic voltmeter is used. The high-resistance wire should, moreover, be one having a negligible change of resistance with temperature. For this purpose it must be an alloy such as manganin or constantan. It is always an advantage, if possible, to employ an electrostatic voltmeter for measuring potential difference if it is necessary to keep the voltmeter permanently connected to the two points. Any form of electrokinetic voltmeter which involves the passage of a current through the wire necessitates the expenditure of energy to maintain this current and therefore involves cost of production. This amount may not by any means be an insignificant quantity. Consider, for instance, a hot-wire instrument, such as a Cardew's voltmeter. If the wire has a resistance of 300 ohms and is connected to two points differing in potential by 100 volts, the instrument passes a current of one-third of an ampere and takes up 33 watts in power. Since there arc 8760 hours in a year, if such an instrument were connected continuously to the circuit it would take up energy equal to 263,000 watt-hours, or 260 Board of Trade units per annum. If the cost of production of this energy was only one penny per unit, the working expenses of keeping such a voltmeter in connexion with a circuit would therefore be more than £l per annum, representing a capitalized value of, say, £10. Electrostatic instruments, however, take up no power and hence cost nothing for maintenance other than wear and tear of the instrument.

The qualities required in a good voltmeter are:—(i.) It should be quick in action, that is to say, the needle should come quickly to a position giving immediately the P.D. of the terminals of the instrument. (ii.) The instrument should give the same reading for the same P.D. whether this has been arrived at by increasing from a lower value or decreasing from a larger value; in other words, there should be no instrumental hysteresis. (iii.) The instrument should have no temperature correction; this is a good quality of electrostatic instruments, but in all voltmeters of the electrokinetic type which are wound with copper wire an increase of one degree centigrade in the average temperature of that wire alters the resistance by 0.4%, and therefore to the same extent alters the correctness of the indications. (iv.) It should, if possible, be available both for alternating and continuous currents. (v.) It should be portable and work in any position. (vi.) It should not be disturbed easily by external electric or magnetic fields. This last point is important in connexion with voltmeters used on the switchboards of electric generating stations, where relatively strong electric or magnetic fields may be present, due to strong currents passing through conductors near or on the board. It is therefore always necessary to check the readings of such an instrument in situ. Electrostatic voltmeters are also liable to have their indications disturbed by electrification of the glass cover of the instrument; this can be avoided by varnishing the glass with a semi-conducting varnish so as to prevent the location of electrostatic charges on the glass.

See J. A. Fleming, Handbook or the Electrical Laboratory and Testing-Room (London, 1903); G. Aspinall Parr, Electrical Engineering Measuring Instruments (London, 1903); K. Edgecumbe and F. Punga, "On Direct Reading Measuring Instruments for Switchboard Use," Journ. Inst. Elec. Eng. (London, 1904), 33, 620.  (J. A. F.)