LIGHTNING1

J. J. Simpson

THE ELECTRICAL CONDUCTIVITY OF AIR AND CLOUDS

Undoubtedly, ninety percent of physicists, without hesitation, would give an explanation of the phenomenon of a thunderstorm discharge. This explanation would run approximately as follows: “During a thunderstorm a cloud receives an electric charge, which increases until a spark leaps from it to another cloud or to the earth, in the same way as an isolated copper sphere connected to an electrostatic machine is discharged.” This means that clouds are regarded as ordinary conductors that can be charged and then discharged at a sufficiently high potential. I hope that I shall be able to demonstrate the complete erroneousness of such a conception.

Almost at the very beginning of the science of electricity it was known that air is not an absolute nonconductor of electricity. On the basis of observations it was established that charged bodies lose their charge under conditions excluding defects of insulation. At first it was assumed that the loss of electric charge is greatest in damp weather, when the air is saturated with moisture, but in 1887 Linss established the opposite phenomenon, i.e. a greater loss of charge in fine weather, with dry air, thereby proving that in the question of the electrical conductivity

air, the humidity or dryness of the air are not decisive factors. However, the question of how this leakage of charge occurs seemed extremely unclear, since at that time the assumption that gases, like metals, could possess conductivity appeared incredible. The generally accepted view was that air molecules, colliding with a charged body, carry away with them part of its charge when they rebound. The well-known experiment with a pendulum, in which an insulated pellet, oscillating between charged and uncharged bells, causes the latter to ring, probably served as the basis for such a view. In any case, it seemed just, and it was impossible to refute it.

The discovery of the rays of Röntgen and of the rays of Becquerel led to a rapid development of our ideas about the electrical conductivity of gases. It was established that these rays can ionize gas molecules, tearing electrons out of them and thereby imparting a positive charge to the molecule; on the other hand, a neutral molecule that captures a free electron acquires a negative charge. A gas molecule charged in this way is an ion (in this sense I shall use the word “ion” everywhere). Consequently, a gas molecule that has lost an electron is a positive ion, while a molecule that has captured an electron is a negative one.

In 1900 and 1901, Elster and Geitel in Germany and Wilson in England proved that air free of dust contains positive and negative ions, and that the electrical conductivity of air is determined by their number and mobility in an electric field. It was further established that in the lower layers of the atmosphere each cubic centimeter of air contains about 500 ions of each sign, moving with a speed of 1.5 cm/sec in an electric field with a gradient of 1 V/cm. A simple calculation shows that this accounts for a small but quite definite electrical conductivity of air, the magnitude of which can be estimated on the basis that the resistance

of the lower layers of the atmosphere in clear weather is equal to \(4.5 \cdot 10^{15}\ \Omega\). As Swann pointed out (to illustrate this figure), a column of air several centimeters long offers to an electric current a resistance equal to that of a copper wire of the same cross-section, 30,000 billion miles long, i.e. 40 times greater than the distance from the Earth to Arcturus. The electrical conductivity of the atmosphere determined in this way may seem exceedingly small, but nevertheless it is of essential significance. It is sufficiently great that an isolated charged conductor placed in it would lose, within a minute, 3% of its charge, and practically all its charge would escape into the air within half an hour.

If, however, we turn to the consideration of the electrical conditions inside a cloud, we shall see that here they are entirely different. Inside a thundercloud ions are practically completely absent, since, even if they are formed as rapidly as in clean air, they are immediately absorbed by the particles of the cloud and lose their mobility. Thus the electrical conductivity inside a cloud is small, and a thundercloud is one of the best insulators. The author has experimentally convinced himself that charged bodies placed inside a cloud, over the course of an hour, and even over a longer interval of time, did not lose their charge at all. We must evidently change our ideas about thunderstorm phenomena: instead of electrically conducting clouds situated in a non-conducting atmosphere, we must picture non-conducting clouds in an electrically conducting atmosphere. Our conception of the electrical conductivity of clouds corresponded to the explanations of thunderstorm phenomena usually given in textbooks; according to these explanations, a spark jumps between the nearest points of two different clouds, and a charged cloud induces opposite charges on the upper and lower surfaces of the cloud lying below. If, however, we regard the cloud as a perfect non-conductor, then such an explanation must be discarded, since charge cannot accumulate on the surface of an insulator,

just as it usually accumulates on the surface of a conductor.

Below I shall describe the process of separation of large quantities of electricity in different parts of a cloud; for the present, however, I shall dwell on the consideration of the electricity present in the cloud, regardless of how it has appeared. The electric charge of a cloud must obviously be associated with water, or with the particles of the cloud, with rain, hail, or snow. This charge may be very large,—so large that the electric force acting on raindrops during a thunderstorm may exceed the force of gravity. Since the charge gradually accumulates in some part of the cloud, it must eventually discharge, either owing to an opposite charge accumulated in another part of the cloud, or as a result of the induction of a like charge on the surface of the earth. Physicists have paid very little attention to the question of how this discharge begins and how it propagates. We know quite a lot about the discharge of electricity through air from one conducting electrode to another; but in thunderstorm phenomena we have no electrodes, or, at best, only a single electrode, when the charge passes into the earth or out of the earth.

Mechanism of Thunderstorm Discharges

I have already said that inside the cloud the charge is concentrated on particles of water; therefore, in moving, the charge must carry water along with it. It is known that the mobility of water particles, even very strongly charged ones, is so small that only an electrical breakdown of the air can produce a noticeable current. In other words, a discharge can occur only in the case of such a breakdown. I must therefore say a few words about the process of ionization and about the electrical breakdown of air.

At normal pressure and temperature, electrical breakdown of air occurs when the field strength reaches 30,000 volts per centimeter. Under these condi-

in which, over a short interval of its free motion, an electron acquires such a velocity that, upon colliding at the end of its path with a neutral molecule, it proves capable of tearing one or several electrons from it. These liberated electrons, entering a field of the same intensity, reproduce the very same process, as a result of which, in a short time, a large number of free electrons is formed, moving at high speed along the electric lines of force. The speed of motion of electrons differs greatly from the speed of motion of ions in the same field. This difference is so great that, even in the strongest field, negative and positive ions, in comparison with electrons, may be regarded as practically immobile. The carriers of positive electricity are always only ions; the elementary particles of positive electricity—protons—being atoms of matter, are ions. Therefore, when air is punctured, only large quantities of negative electricity move, in the form of electrons, while the positive electricity remains bound to the practically immobile positive ions.

Since during a thunderstorm an electric charge of one sign accumulates on one part of a cloud, and a charge of the opposite sign on another part, the field intensity between them increases. In Fig. 1 (a) these charges are represented by the signs \(+\) and \(-\); the lines of force are indicated there as well. Since the charge is distributed nonuniformly, the lines of force are not parallel; they approach one another at the place where the two charges join, where the field intensity is maximal.

With a sufficient increase in the field intensity, puncture of the air will occur first of all along the line \(AB\). As soon as this happens, this small region will become highly electrically conducting; the effect obtained is analogous to that which we have when introducing a piece of wire into an electrostatic field. The lines of force shown in diagram 1 (b) indicate the new distribution; the field intensity

increases considerably toward the edges of the conducting region, where the lines converge toward it.

In Fig. 1 (c) the region of electrical breakdown of air is shown on an enlarged scale. The electrons moved rapidly to the left, i.e. in the direction toward the positive charge; the right half of the conducting region was filled by positive ions, which, as has already been indicated, may be regarded as immobile. Such is the picture immediately after the breakdown of the air; of particular interest, however, are the subsequent changes. In the diagram, the arrows on the lines of force indicate the direction of motion of the electrons. To the left, the electrons are drawn out of the conducting region into the surrounding non-ionized air; from there they pass into the more intense field and are captured by neutral molecules, forming negative ions. These ions have low mobility and remain in the region of the dispersed field, at the end of the conducting channel.

Fig. 1. Diagram of the formation of a conducting channel.

The action of this cloud of negative ions on the electri-

The electric field is shown in Fig. 1 (d); from it one sees that along the edges of the conducting region the original intense field no longer exists, since this region has expanded considerably and many of the lines of force terminate in the region where the negative electricity associated with the new negative ions is located.

Turning again to Fig. 1 (c) and considering the conditions at the other end of the conducting region, we find in them a great difference. The channel is filled with positive ions, too massive to move appreciably, despite the tendency of the field to push them out; as a result the shape of the channel remains unchanged. However, owing to the concentration of lines of force that has occurred after the breakdown of the air, the field strength at the edge of the conducting region is very great, and the air in the space surrounding the end of the channel [indicated by the dotted line in Fig. 1 (c)] is unable to withstand the electric stress. As a result a new breakdown occurs, with a new liberation of a large number of electrons. These electrons immediately rush into the conducting channel, where they strongly ionize the air by colliding with its molecules as they pass through the channel. Ultimately they penetrate into the cloud of negative ions through the other end of the channel. The transfer of electrons from the newly ionized region into the other end of the channel is accompanied by the filling of this region with positive ions; in other words, this channel, as shown in Fig. 1 (d), is simply prolonged into the region of negative electricity. However, the process does not end there, since the end of the lengthening channel remains sharp, as a result of which the field at its end is still sufficiently intense to ionize the air. Thus the channel itself rapidly lengthens in the direction opposite to the motion of the electrons and penetrates far into the region where, before the discharge, the field strength was too weak to cause breakdown of the air.

Often accidental causes bring about branching as the channel grows, as a result of which the picture shown in Fig. 1 (e) is obtained. This figure illustrates

two extremely important properties of the electric discharge in air at normal pressure. 1. The occurrence of a discharge causes the channel to be pierced in only one direction, namely toward the place where the negative electricity is situated; in the other direction no channel is formed; there is found only a diffuse cloud of negative ions. 2. All the branches have one and the same direction, namely: toward the location of the negative electricity. This circumstance is of great importance, since it gives us the possibility of determining the positive end of lightning, if its branches are visible.

Fig. 2. Discharge between positive and negative electrodes.

Until now my description of the thunderstorm discharge has been purely theoretical, but a series of laboratory experiments carried out by me has fully confirmed this theory. These experiments are described by me in a special article on lightning, and I shall mention here only one of them. Two copper disks were placed on a photographic plate, representing charged regions of a cloud, one positive \((A)\), the other negative \((B)\). For concentration of the field, a small wire protruded from each disk. The disks were connected to a Wimshurst electrostatic machine, as a result of which a discharge occurred, the results of which are indicated on

Fig. 2. Long, thin channels, very pointed at the ends, emerge from the positively charged electrode; from the negatively charged one, however, no channels emerge; there is only a small cloud around the end of the wire.

Types of thunderstorm discharges

Wishing to apply the results of this experiment to thunderstorms and discharges, we must subdivide these discharges into three possible types: a) discharges between oppositely charged regions of the atmosphere, b) discharges between positive electricity enclosed in a cloud and the earth, and c) discharges between negative electricity enclosed in a cloud and the earth.

a) Discharges between oppositely charged regions of the atmosphere. In Fig. 1 a discharge is shown between two regions, respectively positively and negatively charged. As a result of the discharge, we have, inside the positively charged region, a large diffuse cloud of negative ions, and inside the negatively charged region—a branching channel with positive ions enclosed in it. The discharge has destroyed the field, but has not produced a significant displacement of electric charges; the latter depends on the normal motion of atmospheric air, owing to which the transferred charge passes into the surrounding parts of the cloud, where it combines with the original charge, as a result of which complete neutralization occurs. We see that the discharge begins where the field intensity is maximal, and the channel increases and branches in the direction of the negative charge.

Fig. 3. Diagram of a discharge from a positively charged region.

b) Discharges between positive electri-

electricity contained in the cloud, and the earth. The shaded surface in Fig. 3 depicts a cloud in which positive electricity, indicated by the plus sign, is concentrated. The discharge begins in the lower part of the charged region, where at first the field intensity was at its maximum. The discharge proceeds in the direction of the induced negative charge of the earth and,

Fig. 4. Photograph of a discharge taken with a stationary camera. (Photograph by B. Walter.)

as it develops, the conducting channel branches. The degree of branching varies greatly; we can observe all gradations of discharge, from an unbranched straight flash up to very highly branched discharges accompanying severe thunderstorms.

The formation of a cloud of negative ions near the end of the discharge undoubtedly plays a major role in determining its character, since it reduces the field at the base of the channel and

tends to hinder the passage of electricity along the latter. A cloud of negative ions may actually block the channel even before an amount of negative electricity sufficient to neutralize the positive charge in the cloud has passed through it. In such a case the discharge stops until either the blockage is removed, or a lateral discharge opens a new path between the end of the channel and another part of the cloud. Thus the discharge along the main channel may be intermittent, since the channel remains ionized for an appreciable interval of time after the cessation of the partial discharge.

The process described is illustrated very well by two photographic pictures by Prof. Walter, made by him in Hamburg with the aid of two photographic cameras, one of which was fixed immovably, while the other was mounted on a rotating stand.

Fig. 5. Photograph of the discharge shown in Fig. 4, taken with a rotating camera. (Photograph by B. Walter.)

In Fig. 4 a picture taken with the stationary camera is shown. The fact that the branchings of the channel inside the cloud converge to a single point clearly indicates that the discharge begins not on the surface of the cloud, as would be the case if the cloud were a conductor. In Fig. 5 a picture taken with the aid of the rotating camera is shown. The camera

moved in such a direction that the first flash appeared on the photograph at the right. It should be noted that this flash was the only vertical discharge. After \(1/10\) sec there followed a second discharge, which passed along the channel of the first to the point indicated by the arrow, and then formed a second channel; this applies to the two main channels indicated in the photograph taken by the fixed camera. Then, after short intervals of time, two more discharges occurred, after which the discharge, although very weakened, nevertheless did not cease for more than \(1/10\) sec; in the end it concluded with two more flashes, following one another after a short interval of time. Between the first and the last discharges a little more than half a second elapsed. In this example the charge of the cloud that caused the lightning was undoubtedly positive, since the branches were directed in the direction opposite to the cloud; as for the intermittent character of the discharge, it can be satisfactorily explained as the result of the channel being clogged by negative ions.

With a large accumulation of positive electricity inside the cloud, channels ending in it cannot reach the earth. Unfortunately, in view of the fact that such shortened flashes occur inside the cloud, it is not possible to verify this directly. If one of the branches reached the lower part of the cloud and became visible, then it would undoubtedly reach the earth, i.e., as it approached the latter the intensity of the electric field at the end of the channel would increase greatly. Although by means of photography we cannot prove the existence of lightning flashes ending in the air, there are nevertheless many indirect indications that flashes of this kind occur in the tropics, where thunderstorms take place in higher layers of the atmosphere than in the temperate zone.

c) Discharges between negative electricity enclosed in a cloud and the earth. Before the passage of the discharge, the electric field beneath a negatively charged cloud is similar to the field beneath a positively

charged cloud, only the direction of these fields is opposite. Breakdown of the air in this case also occurs at a point lying near the lower edge of the charge; but the discharge then propagates not in the direction toward the earth, but along the lines of force into the interior of the cloud. This is shown schematically in Fig. 6. To obtain a photographic image of such a discharge is, unfortunately, not possible, since it occurs in the cloud. There is reason to suppose that such discharges occur very rarely, since the mechanism of concentration of negative electricity within a small part of the cloud is unknown, in contrast to the case with positive electricity.

Fig. 6. Diagram of a discharge within a negatively charged region.
Fig. 7. Diagram of a discharge between a negatively charged region and the earth.

For reasons that will be indicated below, negative electricity is usually distributed through a large mass of the cloud. The electric field beneath such an extensive cloud is relatively homogeneous and is characterized by vertical lines of force. The region of maximum intensity will now be located not near the cloud, but near the earth, where there are mountains, towers, trees, and other elevations, on which the lines of force will concentrate. If the disruptive voltage is then reached, a conducting channel is formed, rapidly rising upward toward the cloud, where it then branches, as shown in Fig. 7. To make photographic images of such

flashes is extremely difficult, since branching occurs almost always inside the cloud. The best photograph of flashes of this type (belonging to J. C. Clark and taken in June 1892) is reproduced in Fig. 8.

Of the three types of discharges that we have considered—the discharge inside a cloud, the discharge to earth from a positive cloud, and the discharge to earth from a negative cloud—the last two are of greatest importance for the electrical engineer, since they cause the destruction of buildings, overhead wires, etc. But the characteristic features of the two types of discharges are very different.

Fig. 8. Photograph of a discharge at Herne Bay.

A discharge from a positive cloud begins in the high layers of the atmosphere and branches on its way to the earth. Objects connected with the earth’s surface may therefore experience the strike either of the main stem or of one of its branches. On the other hand, a discharge in the direction of a negative cloud begins from a rising object, through which the entire discharge passes. Thus, with a positive discharge, the probability of a strike is much greater than with a negative one.

For a discharge of the same quantity of electricity from a negative and from a positive cloud, the quantity of electricity flowing through the object struck by lightning

object, is not the same; with a positive discharge, the greater part of the electricity remains in the channel and its branches, so that only a small part of the entire charge falls upon the object under consideration; with a negative discharge, however, the latter passes wholly through the object struck by the lightning. Therefore the intensity of the stroke is greater in a negative discharge than in a positive one. Positive discharges, as I have shown elsewhere, usually occur more often than negative ones.

On the basis of the theoretical considerations set forth, one may conclude that discharges issuing from positively charged clouds, being comparatively rare, possess an extraordinarily great intensity. In order to verify these conclusions, I studied more than 400 photographic images of lightning, determined the sign of the discharge from the direction of the branches, and came to the conclusion that the number of discharges between a positively charged cloud and the earth exceeds, at least fourfold, the number of discharges between a negatively charged cloud and the earth; there is even reason to suppose that the ratio between these numbers is closer to 10 than to 4.

Until recently physicists were much occupied with the question of whether a thunderstorm discharge is unidirectional or oscillatory. At the present time this question has been definitively resolved, thanks to the observations of Watson, Norinder, and Matthias, who, with the aid of a cathode oscillograph, proved that the principal discharge in the bright flash consists of a unidirectional current beginning from zero, increasing to a maximum, and again decreasing more or less rapidly to zero.

On the basis of the theory expounded above this fact is very easily explained. In an oscillatory discharge the capacitor has an essential significance; as it discharges, it passes through a state of equilibrium, thereby causing a reverse current which recharges the capacitor. So long as a cloud was regarded as a conductor that could be discharged like a Leyden jar, it was natural to consider lightning an oscillatory discharge. But at the present time we

we know that, in practice, a cloud is a perfect non-conductor, as a result of which it has no capacitance and cannot act like a Leyden jar. Therefore we have every reason to regard the main discharge as unidirectional.

LIGHTNING AND ATMOSPHERIC DISCHARGES

Here, however, one more factor must be borne in mind. Although the cloud as a whole has no capacitance, the conducting channel with its branches is a conducting system and therefore possesses capacitance and self-induction. We may imagine the channel of a thunderstorm flash, directed toward the earth or away from the earth, as a huge wireless antenna; if, moreover, the resistance of the channel is not too great, it will produce electrical oscillations, like any other radio antenna. Applying the usual formula, we find that a vertical antenna, grounded at its lower end, 2 km high and 5 cm in diameter, can give an oscillatory discharge when the resistance is not more than one ohm per meter. There is no doubt that the resistance of a fully developed thunderstorm channel, maximally ionized, may be still smaller, as a result of which the channel acquires the ability to produce electrical oscillations.

These oscillations will be superposed on the main current, but will not thereby change its direction. In other words, the intensity of the unidirectional current of the main discharge will pulsate with a frequency corresponding to the period of the free oscillations of the channel; this effect is somewhat reminiscent of the action of the “singing arc,” in which the conducting path in the air between the electrodes is maintained by a unidirectional current, while the oscillations that produce the musical tone are superposed on the current in accordance with the natural frequency of the circuit.

Oscillations of this kind, superposed on a thunderstorm discharge, were in fact discovered and studied by Norinder, and Watson-Watt established their extensive development in atmo-

atmospheric discharges observed by him in Khartoum; their oscillograms are shown in Fig. 9. If, as Appleton supposes, atmospheric discharges owe their origin chiefly to thunderstorm discharges, then this “ripple,” these small waves, find their explanation in the oscillations inside the thunderstorm channel that follow from our theory.

Fig. 9. Wave form of atmospheric discharges observed by Watson-Watt.

The wavelength of the radiation caused by the oscillations is approximately twice the length of the channels whose two ends are in the air, and four times the length of the channels that reach the earth. If we assume that the average length of a thunderstorm channel is 2 or 3 km, as Watson-Watt does, then the radiation wavelength will be 10 km, whereas Watson-Watt found that the wavelength of the waves of the “ripple” he investigated is close to 30 km, i.e. it is of the same order of magnitude.

Incidentally, it should be noted that the wavelength of these oscillations lies within the range of wavelengths used in wireless telegraphy, and therefore the oscillations arising in this way are apparently the principal cause of disturbances connected with atmospheric discharges.

Quantitative Aspect of the Question

It is very important, in all practical questions relating to the lightning discharge, to know the magnitude of the electrical factors associated with it. Almost everything that we know on this subject we owe to C. T. R. Wilson; here I shall give only the results of his investigations, without attempting to characterize the methods he used; I shall merely mention that they also included measurement of those changes in the strength of the electric field which are produced by a lightning discharge at a known distance from it.

Wilson found that the quantity of electricity associated with a lightning flash of average intensity varies between 10 and 50 coulombs; he regards 20 coulombs as a typical quantity, an astonishingly small value, equal to only 20 A/sec; this quantity of electricity may seem too insignificant from the point of view of technology. In 1920 Wilson was the first to determine, by means of observations, the quantity of electricity associated with a lightning flash. However, even the great Faraday, in his early researches on electricity, arrived at analogous results. Namely, Faraday wrote that the quantity of electricity necessary for the decomposition of a drop of water (700 coulombs) is “equal to the quantity of electricity discharged in the strongest thunderstorm.”

Wilson further calculated that in a thundercloud, before the passage of a discharge of 20 coulombs, the electric potential is equal to a billion volts. Of course, owing to the resistance of the air in the channel, by the moment when lightning strikes an object situated on the surface of the earth the magnitude of the potential is reduced; it may be thought, however, that when lightning strikes such conductors as, for example, transmission lines not directly connected with the earth, potentials of tens of millions of volts are attained. The energy of the discharge can be directly calculated from the quantity of electricity and the voltage. In the case of a flash producing a discharge of 20 coulombs, it is determined as 10 ergs. Being expressed

in ordinary units, the energy of an average thunderstorm discharge is equal to 3000 kW/sec. Here, too, as in the case of potential, this energy does not entirely reach the earth’s surface, since a large part of it is absorbed in the channel; however, many of the strongest discharges must transmit a large part of their energy to the earth.

A very important element of a thunderstorm discharge is its duration. The latter was not determined by Wilson, but at present we have the results of investigations recently carried out by Norinder and Matthias; if, in addition, we assume that atmospheric discharges are due to lightning, then we shall have a larger quantity of data collected by J. J. Thomson. According to all these determinations, the average duration of a thunderstorm discharge is somewhat more than 0.001 sec. A discharge of 20 coulombs in the course of 0.001 sec gives an average current of 20,000 A—a value corresponding to direct observations of the strength of thunderstorm current made by Matthias. In the central part of the discharge period this value may be considerably greater, reaching 100,000 A.

Mechanism of a Thunderstorm

To understand the various thunderstorm phenomena it is necessary to have an idea of the processes causing the formation of the intense electric field associated with a thunderstorm. Unfortunately, the opinions of meteorologists on this question differ, and there is no single theory accepted by all. In 1909 I began to develop a theory on this question, which in 1927 I elaborated and supplemented definitively, and which I feel entitled to propose, at least, as an attempt to explain the mechanism of a storm.

A theory of the electrical processes during a thunderstorm must first of all give an explanation of the phenomenon of the initial separation of electricity into positive and negative (or, expressed otherwise, the “generation of electricity”) and then of the transition of both kinds of electricity into entirely different parts of the cloud. In the theory I propose, the ...

Through this region water droplets cannot pass, since the relative velocity between the air and a raindrop having a diameter of 0.5 cm is equal to 8 m/sec, while drops of larger sizes disintegrate owing to their instability.

The broken lines in Fig. 10 show the paths of falling raindrops. On the left they are vertical, while on the right they are deflected leftward by the air current. The magnitude of the deflection from the vertical evidently depends on the sizes of the drops. Large drops are deflected only slightly, whereas small droplets—the cloud particles—follow practically the streamlines of the air.

Fig. 10. Thunderstorm diagram. Meteorological conditions. Along the ordinate axis height is plotted in kilometers.

It is clear from the diagram, without further explanation, that water must accumulate above the region of maximum vertical velocity. Only large drops can penetrate into the lower part of this region, nearer to the place where the vertical velocity is equal to 8 m/sec. These drops will disintegrate, and their particles will be carried upward. The small droplets will again combine and again fall downward, and so on. The region where this process of destruction and re-formation of drops proceeds with particular intensity is indicated in the diagram by a curved dotted line.

The electrical conditions, associated with the meteorological ones,

depicted in diagram 10 are also shown graphically in Fig. 11.

In that region in which the vertical velocity of the air exceeds \(8\ \text{m/sec}\), electricity cannot accumulate. Above this region, in the area denoted by the letter \(B\), the process of splashing and re-formation of raindrops takes place—here every destruction of a raindrop is accompanied by positive electrification of the water composing this drop. In the air there is produced a corresponding negative charge, immediately absorbed by the particles of the cloud, i.e., by the finest droplets carried away by the strong

Fig. 11. Thunderstorm diagram. Electrical conditions.

air current. However, water that has received a strong charge cannot so easily leave region \(B\), since small droplets, rapidly coalescing, fall again in order to receive a new additional positive charge. Thus the water that has accumulated in \(B\) acquires a large charge of positive electricity, which is shown by the \(+\) sign in the figure. The negative charge of the air passes from \(B\) into the main part of the cloud, which as a result acquires a negative charge. Region \(B\) may be characterized as the place where electricity is separated into positive and negative.

Raindrops falling from this region will be very

apparently have a positive charge. Thus, the heaviest rain in the central part of the storm must be positively charged. On the other hand, outside the region of the ascending current, the raindrops must have a negative charge, i.e., they fall exclusively from the negatively charged part of the cloud. Observations confirm this distribution of positive and negative electricity in falling rain.

Figure 11 shows the types of thunderstorm discharges that may occur in connection with the described distribution of electricity. The principal place of origin of a thunderstorm discharge will obviously be the region of separation of electricity, since it is precisely here that the positive charge of the accumulating water is concentrated in unlimited quantity. From this positive charge the discharge may proceed toward the negatively charged part of the cloud; more often, however, discharges will occur in the direction of the earth, some of them reaching the earth and others ending in the air; the most frequent discharges in thunderstorms in the tropics belong to the latter type. Sometimes there may also be discharges originating from the earth and going toward the main negatively charged part of the cloud. These discharges will have branches directed upward, in contrast to discharges originating from the region of separation of electricity, whose branches are directed downward. Such, according to the theory of the breaking up of drops, is the mechanism of thunderstorms. Of course, in nature everything is much more complicated than in the simplified scheme outlined here. Therefore it should not be surprising that not all observations of lightning fit into this scheme. Nevertheless, this scheme can undoubtedly explain all the most important observed facts, in particular the distribution of positively and negatively charged raindrops and the frequency of the various kinds of thunderstorm discharges.

1. The twentieth Kelvin lecture, delivered at the Institution of Electrical Engineers. Printed in the Journal of the Institute of Electrical Engineers, Vol. 67, No. 395, Nov. 1929. Translation with some abridgments. ↩