standard condenser of known capacity, either directly or indirectly. Before proceeding further, however, I would deal in brief with a special form of galvanometer, adapted for the measurement of such transient currents as those involved in the electrostatic discharge of a condenser of large capacity, such as a long section of submarine cable, for example. Such a discharge current lasts for a perceptible period of time, whereas the momentary discharges from condensers of low electrostatic capacity are comparatively brief, and may readily be compared by the transient swing of an ordinary Thomson reflecting galvanometer. Nevertheless, as the reader may have occasion to deal with large electrostatic capacities in one or another connection, a knowledge of the construction and principle of what is known as the "ballistic galvanometer" will not come amiss.

The ballistic galvanometer, as commonly constructed, is shown in the accompanying illustration, and is exactly similar to an ordinary Thomson reflecting galvanometer in all respects except the suspended system, the detailed construction of which is represented in diagram in Fig. 62, where a b represents the ordinary aluminium axis, pass

ing through the centre of a split cylindrical magnet A, shaped somewhat like a thimble; a transverse section along the line c d is shown at B. These magnets, with their semi-cylindric poles n s, replace the little watchspring magnets in the Thomson instrument, and may be mounted, two with their like poles opposite, in the centre of each coil, and two others immediately outside the coil above and below, as represented in the complete illustration above, or in any other similarly suitable manner, the object of the cylindrical form given to the magnets being that they may offer as little resistance to motion as possible in their passage through the surrounding air. As a matter of fact, in practice, the motion of the needle does not commence until the current causing it has ceased.

When, as is often the case, the number of oscillations made by the needle of a ballistic galvanometer in a given time, is required, it can be obtained by fixing the eye 2

d

FIG. 62.

upon a certain point on the galvanometer scale within the limits of the deflection, and counting the number of times the reflected spot of light passes that point, whilst travelling in the same direction, during the time stated.

For the mathematical and mechanical proofs underlying the action of the ballistic galvanometer, the reader is referred to Kemple's "Handbook of Electrical Testing." To proceed, however, with the more immediate subject

matter of this section, viz., the determination of electrostatic capacity. For the majority of the tests about to be described in this connection, the ordinary Thomson reflecting galvanometer is admirably suited, more especially if fitted with the usual damping device to check the more or less irresponsible swings due to the sudden condenser discharges, etc.

The simplest method for the determination of electrostatic capacity is known as the direct deflection method, and is represented diagrammatically in Fig. 63, where G represents a Thomson galvanometer, which may or may not be provided with a shunt across its terminals, according to whether the discharges to be compared are small or large; it is therefore to a great extent dependent upon the battery power E used, but, in any case, a shunt will be found useful, and should be included among the apparatus for the test, although it has been omitted from the figure for the sake of clearness. K represents Lambert's discharge key, previously described and illustrated, whilst C is the standard condenser of known capacity, for which is ultimately substituted the electrostatic capacity or condenser, whose value it is required to determine, such as a length of cable, for example.

The modus operandi is as follows:-K2 is first depressed for a definite period, such as 30 seconds or more, in order to charge the standard condenser C to saturation.

It is then released, and K1 depressed in its turn; this action serves to discharge C through the galvanometer, the magnitude of the initial swing of the needle of which should be noted; we will call it d. The standard condenser C is then disconnected and replaced by the condenser under test, which in the case of our length of cable, for example, would consist of one end (the other being free), and the earth surrounding it, such as the armouring or lead sheathing, if it possessed any, or the water in a tank if the cable be submerged therein. The operations with K1 and 2 are then repeated, and a second deflection or swing dl is obtained.

Then the capacity under test: C: d1: d

or, the required capacity =

C dl
d

If a shunt be used in either of the above operations, the necessary multiplication of the observed deflections G+ S, by the now well-known formula S

be effected.

must, of course,

Gott's method is illustrated in Fig. 64, where C and C1 represent the standard and the condenser under test respectively, E the battery, K a Webb's discharge key previously described, K1 an ordinary circuit key, G the galvanometer, and r, rl two resistances giving a fairly large

range of proportional adjustment, with attendant accuracy of comparison. r and rl are first adjusted as nearly as possible in the proportion of C to C1; K is then depressed, and held in that position by means of its detent, thus charging both C and C1 simultaneously in series, from the battery E. After a definite period as before, K1 is closed, and, if a deflection be obtained on the galvanometer G, both keys are released, and the respective condensers short-circuited, or put to earth as the case may be, in order to dissipate their respective charges.

r and

rl are then readjusted, and the operation repeated until
no deflection results upon the galvanometer when Kl is
depressed.

This being the case, C1 = C C 21

r

The proportional resistances r and r1 may conveniently be arranged in the form of a slide resistance of some 10,000 ohms, the galvanometer being connected to the slider.

Thomson's method of capacity measurement is very similar to the foregoing, the connections being represented in Fig. 65, where C and C1 represent the standard and condenser under test respectively, E the battery, G the galvanometer, and 1, 2, 3, 4, and 5 simple circuit