painful or even fatal. He should not take shocks from any but small coils, say those giving 4-inch sparks as their maximum.

As regards the winding of the secondary coil, we may say that this wire is wound in sections, separated from one another by ebonite partitions; these sections being such that any two portions of the wire which are at a great difference of potential at the make or break, are separated from each other by one of these partitions. In fact, great care is taken to prevent the coil being ruined by internal discharge.

The primary may be made and broken in many ways; the 'hammer' arrangement is only one of many forms. Thus the 'makeand-break' is sometimes effected between a platinum point and an alloy of mercury and platinum, under the surface of alcohol; alteration of the surfaces due to the extra-current spark being thus obviated. The speed of the make-and-break may be regulated by hand, by clockwork, or in other ways. Thus the primary currents are more under control, and the inverse and direct induced currents may be observed separately.

§ 8. The Part Played by the Condenser.-Connected with the metal piece Lg are a series of sheets of tin-foil; and connected with Ns are another series. These two sets lie as shown in the fig. ii. of the last section, being separated by sheets of waxed paper or of other insulating material. The whole thus forms a condenser of very large surface, the set connected with L forming one plate, and that connected with N forming the other. Now when the primary current is broken by the rise of b, an extra current is self-induced in the primary. Were there no condenser, this would leap across the space between 6 and M in disruptive discharge; thus both prolonging the time t1 of break, and injuring the surfaces of contact. But with the condenser the case is different. Before the extra current can give a spark across the gap between 6 and M, it must raise the two large condenser plates connected with b and M, to the necessary difference of potential. These plates have a very large capacity, however. Hence the extra current is employed in charging the condenser, and does not give a spark across M › at all.

This is the main use of the condenser; it prevents sparking,' and thus permits of an abrupt break to the primary.

Another use is as follows. While b is still in mid air between M and A, the extra current not only charges the condenser, but again rebounds (as it were); and, traversing the primary in the

contrary direction, helps to reduce A more abruptly to the neutral

state.

A further effect of the condenser is to lower the E.M.F. of the inverse current that is induced in B on 'making' the primary. For the current has to charge the condenser; and hence its rise to a maximum is delayed. We have said that it is on the direct induced current that we must depend for a high E.M.F.; and we shall show in § 11 how it is an advantage to reduce as far as possible the E. M.F. of the neglected inverse current. The condenser, therefore, will have been of service in a third way.

§ 9. Condition of the Secondary Circuit when Closed.Referring to the formula of § 5, we see that in the coil just described the number N is the same for both induced currents, but the time t2 is greater than t1.

From this it follows that the inverse induced E.M.F. e2 is less than the direct E.M.F. e,, but exists for a longer time. Therefore the inverse current is weaker than the direct, but lasts longer.

Now suppose the secondary circuit to be closed; no air-space left between the terminals and p' of § 7. We should predict that for each make-and-break the total quantities of electricity flowing in the two directions respectively will be equal. Various experiments may be tried with the closed secondary circuit which will illustrate the conditions of things that then obtains.

Experiments.—(i.) When we include in the closed secondary circuit a cell consisting of a copper solution and two copper plates sufficiently large to obviate polarisation and to give good conduction, there is zero resultant action. Here the action depends upon quantity, not E. M. F.; and the result indicates that the quantities in the two currents are equal, and flow in opposite directions.

If the plates are small, so that there is high resistance and polarisation, there will be some slight action showing a predominance of the direct current which has a higher E.M.F.

(ii.) If we use a cell of acidulated water and platinum electrodes, we have results which vary with the size of these electrodes and with the conductivity of the dilute acid. In general we have liberated at each electrode a mixture of the two gases; these will be in proportions which are nearly equivalent at each pole, if the electrodes are large and the liquid very conducting, but will show a predominance of the direct current if the converse conditions hold. This is again due to the higher E.M.F. of the direct current.

(iii.) If we include in the circuit a galvanometer of sufficiently long coil, we have a feeble action in favour of the direct current; but this action is very

faint compared with that which obtains when (as will be explained in § 10) we cut off the inverse current altogether.

(iv.) A person included in the circuit experiences shocks from both currents, but mainly from the direct with its higher E. M.F. (see § 7, caution).

§ 10. Secondary Circuit with Air-Break.—If now we separate the terminals of the secondary circuit, gradually increasing the distance between them, we find that less and less of the inverse current breaks across the air-space; until at last, by interposing a sufficient interval, we have left only the direct current, which from its high E. M.F. is able still to bridge over the gap with a disruptive discharge.

Experiments.—(i.) Interposing such an air-space in the circuit, we find that electrolytic cells in the circuit exhibit decompositions that indicate a powerful direct current.

(ii.) A galvanometer included in the broken circuit now shows a steady deflexion due to the direct currents that rapidly succeed one another.

§ 11. Electrostatic Condition of the Open Secondary Terminals. The Charging of Leyden Jars.-Now let us suppose the interval between the secondary terminals to be so great that not even the direct induced current can strike across.

When the current in the primary is 'made,' the secondary terminals pp become of positive and negative potentials respectively; we may, e.g., suppose to become + and p' to become But, as the E.M.F. of the inverse current is not great, so the. electrostatic difference of potentials observed at p and p' will not be great.

When the primary current is 'broken,' the direct induced current will cause to become —, and p' to become + ; and the electrostatic difference of potential between p and p' will now be far greater than it was.

When, as with the vibrator or other high speed interruptor, the direct and inverse currents follow one another rapidly, the terminals and will show a permanent potential difference, owing to the difference in the E.M.F.s of the two currents. We shall find to be, and p' to be +. This V is of course not really constant, but is continually rising and falling. When, however, the currents follow one another with sufficient rapidity, the needle of a quadrant electrometer will give a deflexion indicative of the average AV between p and p'.

We saw in § 8 that the condenser acted not only to increase the E.M.F. of the direct, but also to decrease the E.M.F. of the inverse, induced current. Hence the condenser acts so as to increase the permanent V between the secondary terminals.

Coil used with Leyden jar.-If the two coatings of a Leyden jar are connected with m and n in the manner shown in the figure,

the discharge is modified in form. We do not now, as is the case when no jar is used, have a discharge each time that the 'break' is abrupt enough to give the E.M.F. required to overcome the airspace m n. On the contrary, the jar is charged only in virtue of the difference in E.M.F.s between the two currents; and the discharge across m n takes place only when the jar is sufficiently charged; so that the period between two discharges depends upon the capacity of the jar. The striking distance is very much smaller than when no jar is used; and, for some reason which the present writer believes has not yet been made quite clear, it depends further upon the capacity of the jar.

Charging a Leyden jar battery.-The above given method may be employed to charge a very large jar or battery. But where it is desired to charge such a battery and to leave it charged, another arrangement is adopted. Here the battery is charged by means of the direct current alone, an air-space being interposed. When the inner coating of the jars rises to a certain potential, sparks will cease to strike across the air-space. The battery may then be removed and its charge utilised.

The use of a 'secondary condenser,' as it is called, causes the

spark to be much denser; the noise also becomes very much greater.

§ 12. Various Phenomena of the Secondary Discharge.

(I.) Luminous effects.-When the discharge takes place under ordinary atmospheric pressures, and without the use of a 'secondary condenser' (i.e. a Leyden jar connected with the secondary terminals), it generally takes the form of a zigzag bluish-white spark, a peculiar sharp cracking sound accompanying it. Under certain conditions this may be seen to be surrounded by a luminous haze. Indeed, there appear to be two modes of discharge that occur at the same time; viz. the true spark, and a quieter form resembling the 'brush discharge' spoken of in Chapter VI. § 15. When the space is too great to admit of a spark passing, there may be observed a phenomenon resembling the brush, or silent, discharge that occurs under similar conditions with electric machines.

The luminous effects occurring in high, and in low, vacua will be discussed in § 14, &c.

(II.) Chemical effects.-When, by means of an air-break, the inverse current is cut off and the direct only passes, we may obtain all the electrolytical phenomena described in Chapter XII. But there is another effect produced by the spark of the induction coil, as also by the continued spark of the Holtz or other similar machine. If the spark be passed for a considerable time through mixtures of gases, there will in certain cases occur combinations that cannot be obtained directly by any other means. Thus we may in this way cause nitrogen to combine directly with oxygen.

(III.) Heating effects.-The temperature of the spark, especially when a secondary condenser is used, is very high. It is, however, rather difficult to distinguish clearly between effects due to the high temperature of the spark, and those due to the violent mechanical disturbance that accompanies the disruptive electric discharge. If, for example, we find the material of the terminal carried off and vaporised, we cannot say that this is wholly due to heat; it may be partly, at least, what we may more fairly call a mechanical action.

(IV.) Use with the spectroscope.-It is found that the discharge has the effect of carrying off and vaporising part of the material of the terminals; this being especially the case when the secon