Lightning in its various forms is the greatest danger to which transmission systems are exposed, and it attacks their most vulnerable point, that is, insulation. The lesser danger as to lightning is that it will puncture the line insulators and shatter or set fire to the poles. The greater danger is that the lightning discharge will pass along the transmission wires to stations and sub-stations and will there break down the insulation of generators, motors, or transformers. Damage by lightning may be prevented in either of two ways, that is, by shielding the transmission line so completely that no form of lightning charge or discharge can reach it, or by providing so easy a path from line conductors to earth that lightning reaching these conductors will follow the intended path instead of any other. In practice the shielding effect is sought by grounded guard wires, and the easy path for discharge takes the form of lightning arresters, but neither of these devices is entirely effective.

Aerial transmission lines are exposed to direct discharges of lightning, to electromagnetic charges due to lightning discharges near by, and to electrostatic charges that are brought about by contact with or induction from electrically charged bodies of air. It is evidently impracticable to provide a shield that will free overhead lines from all these influences. To cut off both electrostatic and electromagnetic induction from a wire and also to free it from a possible direct discharge of lightning, it seems that it would at least be necessary completely to incase the wire with a thick body of conducting material. This condition is approximated when an electric circuit is entirely beneath the surface of the ground, but would be hard to maintain with bare overhead wires. It seems, however, that grounded guard wires near to and parallel with long aerial circuits should tend to discharge any high electrostatic pressures existing in the surrounding air, and materially to reduce the probability that a direct discharge of lightning will choose the highly insulated circuits for its path to earth. Lightning arresters may conduct induced and direct lightning discharges to earth, without damage to transmission lines, so that both arresters and guard wires may logically be used in the same system.

Wide differences of opinion exist as to the general desirability of grounded guard wires on transmission lines, both because of their undoubted disadvantages and because the degree of protection that they afford is uncertain. It seems, however, that the defects of guard wires depend in large degree on the kind of wire used for the purpose, and the method of its erection. Galvanized iron wire with barbs every few inches has been more generally used for guard wires along transmission lines than any other sort. Sometimes a single guard wire of this sort has been run on a pole line carrying transmission circuits, and the more common location of this single wire is on the tops of the poles. In other cases two guard wires have been used on the same pole line, one of these wires being located at each end of the highest cross-arm and outside of the power wires. Besides these guard wires at the ends of the top cross-arms of a pole line, a third wire has in some systems been added to the tops of the poles. These guard wires have sometimes been secured to the poles and cross-arms by iron staples driven over the wire and into the wood, and in other cases the guard wires are mounted on small glass insulators. Much variation in practice also exists as to the ground connections of guard wires, such connections being made at every pole in some systems, and much less frequently in some others.

With all these differences in the practical application of guard wires it is not strange that opinions as to their utility do not agree. Further reason for differences of opinion as to the practical value of guard wires exists in the fact that in some parts of the country the dangers from lightning are largely those of the static and inductive sort, that are most effectively provided for by lightning arresters, while in other parts of the country direct lightning strokes are the greatest menace to transmission systems. At the present time, knowledge of the laws governing the various manifestations of energy that are known under the general head of lightning is imperfect, and the most reliable rules for the use of guard wires along transmission lines are those derived from practical experience.

A case where a guard wire did not prove effective as a protection against lightning is that of the San Miguel Consolidated Gold Mining Company, of Telluride, Col., whose three transmission lines ran from the water-power plant to points from three to ten miles distant, as described in A. I. E. E., vol. xi., p. 337, and following pages. This transmission operated at 3,000 volts, single-phase, alternating, and the pole lines ran over the mountains at elevations of 8,800 to 12,000 feet above sea-level, passing across bare ridges and tracts of magnetic material. It was stated that the country over which the circuits ran is so dry and rocky that it was practically impossible to secure good ground connections along the line, and no mention was made of the way in which the ground wire was grounded, or of the number of its ground connections. Furthermore, it does not appear that there was more than one guard wire on each pole line. Under these circumstances, and with a certain make of lightning arresters in use at the station, lightning was a frequent cause of damage to the connected apparatus. The insulation of some of the machinery is described as honeycombed with perforations which led to continual leakage, grounds, and short-circuits, which seems to indicate that the damage was being done by static and inductive discharges rather than by direct lightning strokes, one of which would have disabled a machine at once. The type of lightning arrester in use on this system was changed, and thorough ground connections were provided for the new arresters, after which the damage by lightning came to an end. It is not stated, however, that the guard wires were removed. This case has been referred to as one in which guard wires failed to give protection, but, as may be seen from the above facts, such a statement is hardly fair. In the first place, it does not appear that the single guard wire on each pole line was effectively grounded anywhere. Again, a large part of the damage to apparatus appears to have been the result of static or inductive discharges that could not in the nature of things have been prevented by a guard wire. Finally, as the guard wire was not removed after the new lightning arresters were erected, it is possible that this wire prevented some direct discharges over the transmission wires that would have been destructive.

On page 381 of the volume of A. I. E. E. above cited, it is stated that the frequency and violence of lightning discharges that entered a certain electric station on Staten Island were much less after guard wires had been erected along the connected circuits than they were before the guard wires were put up.

It is also stated on page 385 of the same volume that examination of statistics of a number of stations in this country and Europe had shown that in every case where an overhead guard wire had been erected over power circuits, or where these circuits ran for their entire distance beneath telegraph wires, lightning had given no trouble on the circuits so protected. Unfortunately, the speaker who made this statement did not tell us where the interesting statistics mentioned could be consulted.

On the first pole line erected for power transmission from Niagara Falls to Buffalo, two guard wires were strung at opposite ends of the top cross-arm on guard irons there located. This cross-arm also carried parts of two power circuits, and the nearest wires of these circuits were distant about thirteen inches from the guard irons. These guard wires were barbed, and grounded at every fifth pole, according to an account given in A. I. E. E., vol. xviii., at 514 and following pages. The character of the ground connections is not stated. Much trouble in the way of grounds and short circuits on the transmission lines was caused by these guard wires at times when they were broken by the weight of ice coatings and wind pressure. As a result of these troubles the guard wires were removed in 1898. Since that date it appears that the transmission lines between Niagara Falls and Buffalo have been without guard wires. Up to 1901, according to page 537 in the volume just cited, twenty per cent of the interruptions in operation at the Niagara plant were caused by lightning, and it seems probable that this record applies to the period after 1898, when the guard wires were removed. It is also stated that during a single storm the line was struck five times, and that five poles with their cross-arms were destroyed. If these direct lightning strokes occurred while there were no guard wires along the line, as seems to have been the fact, it is a fair question whether such wires well grounded would not have carried off the discharges without damage. In California, the country of long transmissions, the use of guard wires along the pole lines is quite general. Many of these lines run east and west across the State, and a single line may thus have elevations in its different parts all the way from that of tide-water up to several thousand feet above sea-level. Unless guard wires are strung with these lines there is much manifestation of induced or static electricity, according to an account at page 538, in vol. xviii., A. I. E. E., where it is said that in the absence of guard wires a person will be knocked off his feet every time he touches a transmission wire that is entirely disconnected from the source of power. It is also said that this static charge on idle power lines is sufficient, in time, to puncture the insulation of the connected apparatus. On the other hand, where the grounded and barbed guard wires are strung over the entire lengths of these long power lines, these lines may be handled with impunity when they are idle. Ground connections to the guard wire are said to be made at about every fourth pole, and to consist of a wire stapled down the face of the pole and joined to an iron plate beneath its butt. The barbed guard wire itself, of which each pole line appears to have but one, is regularly stapled to the tops of the poles.

At the reference just named it is related that on a certain transmission line running east and west across the State for a distance of forty-six miles, and protected by a guard wire, no trouble was experienced during a severe storm that swept north and south over the line. Meantime the damage on other lines in the same neighborhood, and presumably not protected by guard wires, was large.

Between the electric plant at Chambly, on the Richelieu River, and Montreal, Quebec, a distance of 16.6 miles, a transmission line of three circuits on two pole lines, with guard wires, was operated from some time in 1898 to December 1st, 1902, or somewhat more than four years. On the date last named the dam that maintained the head of water at the Chambly station gave way, and the plant was shut down during nearly a year for repairs. For as much as three years this line was operated at 12,000 volts, sixty-six cycles per second, two-phase. During the remainder of the period up to the failure of the dam the line was operated at 25,000 volts, sixty-three cycles, three-phase. In each transmission two pole lines were employed with two cross-arms per pole. One two-phase, four-wire circuit was carried on each of three of these cross-arms. At each end of the upper cross-arm on each pole, and at a distance of fifteen inches from the nearest power wire, a guard wire was mounted on a glass insulator. A third guard wire was mounted on a glass insulator at the top of each pole, and this third guard wire was about twenty inches from the nearest power wire. Each of these guard wires was made up of two No. 12 B. W. G. galvanized iron wires twisted together, with a four-point barb every five inches of length. Poles carrying these lines were ninety feet apart, and at each pole all three of the guard wires were connected by soldered joints to a ground wire that was stapled down the side of the pole, passed through an iron pipe eight feet long, and was then twisted several times about the butt of the pole before it was set in the ground. At three points along the line the conductors consisted of single-conductor underground or submarine cables that had an aggregate length of about twenty-five miles. No lightning arresters were employed at the points where the overhead transmission wires joined the underground cables.

These two-phase, 12,000-volt circuits were operated from some time in 1898 to some time in 1902, and during that time there was no damage done by lightning either at the Chambly plant, on the overhead line or the underground cable, or at the Montreal sub-station. This record is not due to lack of thunder-storms, for in the territory where the line is located these storms are frequent and severe. One very severe storm during the period in question resulted in serious damage on distribution lines at Chambly and Montreal, where the guard wires were not in use, but the transmission line and its connected apparatus escaped unharmed. The path of this storm was in the direction of the transmission line from Montreal to Chambly, and several trees were struck on the way. At the time of this storm and during an entire summer there were no lightning arresters in the power-house at Chambly.

In 1902, when the transmission line just considered was changed from two-phase to three-phase, and its voltage raised from 12,000 to 25,000, the method of protection by grounded, barbed guard wires, as above described, was retained. Two three-phase circuits were arranged on each of the two pole lines, with one wire of each circuit on an upper cross-arm and two wires of each circuit on a lower cross-arm, so that the nearest power wire on the upper cross-arm is thirty-two inches from the guard wire, and the nearest power wire on the lower cross-arm is about thirty inches from the guard wire at each end of the upper cross-arm. The guard wire at the tops of the poles is about thirty-three inches from each of the power wires on the upper cross-arm. In this three-phase line there is about 1,440 feet of three-conductor underground cable, and this cable lies between the end of the overhead line and the sub-station in Montreal. At the juncture of the overhead line and the cables there is a terminal house containing lightning arresters, and there are also arresters at the Chambly plant and the Montreal sub-station. No lightning arresters are connected to this line save those at the generating plant, the terminal house and the sub-station.

During that part of the year 1902 in which the new 25,000-volt line was in operation—that is, after the change and up to the time of the failure of the dam—this line and its connected apparatus were not damaged in any way by lightning, and the same is true for the period in which the line was idle pending repairs on the dam. The experience on this Montreal and Chambly transmission is probably among the best evidence to be found anywhere as to the degree of protection from lightning that may be had by the use of guard wires. In spite of cases like that just considered, where guard wires appear to have given a large degree of protection to transmission systems, many important transmissions are operated without them.

An example of this sort may be seen in the transmission line between the 10,000-horse-power plant at Electra, in the Sierra Nevada Mountains of California, and San Francisco, a distance of 154 miles, where it seems that no guard wires are in use. Another important transmission line that appears to get along without guard wires is that between the 10,000-horse-power plant at CaÑon Ferry, on the Missouri River, and Butte, Mont., sixty-five miles away. On the transmission line between the power-station on Apple River, in Wisconsin, and the sub-station at St. Paul, Minn., about twenty-seven miles long, there are no guard wires for lightning protection. Further east, on the large, new transmission system that stretches from Spier Falls and Glens Falls on the north to Albany on the south, a distance in a direct line of forty miles, no guard wires are employed. On its way the transmission system just named touches Saratoga, Schenectady, Mechanicsville, Troy, and a number of smaller places, thus forming a network with several hundred miles of overhead wire. Examples of this sort might be multiplied, but those already named are sufficient to show that it is entirely practicable to operate long transmission systems without guard wires as a protection against lightning.

With these examples of transmission systems both with and without guard wires, the expediency of their use on any particular line should be determined by weighing their supposed advantages against their known disadvantages, under the existing conditions. It seems fairly certain from all the evidence at hand that if guard wires are to offer any large degree of protection to transmission systems such wires must be frequently and effectively grounded. There is certainly some reason to think that the failures of guard wires to protect transmission systems in some instances may have been due to the lack of numerous and effective ground connections. Such, for example, may have been the case above mentioned, at Telluride, Col. On the other hand, it seems reasonable to believe that the apparently high degree of protection afforded by the guard wires on the Chambly and Montreal line is due to the fact that these wires are connected through soldered joints at every pole with a ground wire that is wound about its base. The nearer the guard wires are located to the power wires on a line the greater is the danger that a guard wire will come into contact with a power wire by breaking or otherwise. It is probable that the protection given by a guard wire does not increase nearly as fast as the distance between it and a power wire is diminished. Even if one guard wire on a line is thought to be desirable, it does not follow that two or more such wires should be used, for the additional protection given by two or three guard wires beyond that given by one wire may be trifling, while the cost of erection and the danger of crosses with the power circuits increase directly with the number of guard wires. At one time it was thought very desirable to have barbs on guard wires, but now the better opinion seems to be that, as barbs tend to weaken the wire, they lead to breaks and cause more trouble than they are worth. The point where the barbs are located seems to rust more quickly than do other parts of the wire. In some cases barbed guard wires that have given trouble by breaking have been taken down and smooth wires put up instead. If a guard wire is well grounded at least as often as every other pole, its size may be determined largely on considerations of mechanical strength and lasting qualities. For ordinary spans a No. 4 B. & S. G. galvanized soft iron wire seems to be about right for guarding purposes. Iron seems to be the most desirable material for guard wires because it gives the required mechanical strength and sufficient conductivity at a less cost than copper, aluminum, or bronze, and is easier to handle and less liable to break than steel. It was formerly the practice to staple guard wires to the tops of poles or to the ends of cross-arms, but it was found that the wire was more apt to rust and break at the staple than elsewhere, and in the better class of work such wires are now mounted on small insulators. This practice, as stated above, was followed on the Montreal and Chambly line. In all cases the connection between the guard wire and each of its ground wires should be soldered, and the ground wire should have a large surface in contact with damp earth, either through a soldered joint with a ground plate, by winding a number of turns about the butt of the pole, or by other means.

It is thought by some telegraph engineers that the use of a separate ground wire running to the top of each pole is quite as effective as a protection against lightning as is a guard wire that runs to all of the poles and is frequently connected to the ground.

This practice is mentioned at page 26 of “Culley’s Handbook of Practical Telegraphy.” Such ground wires are free from most of the objections to the ordinary guard wires. It seems quite certain that a guard wire along an alternating-current line, and grounded at frequent intervals, must act as a secondary circuit of a transformer by reason of its ground connections, and thus absorb some energy from the power circuits. No experimental data are yet available, however, to show how large this loss may be in an ordinary case. It is fairly evident that there must be some electrostatic effects between the working conductors and a guard wire, but here again data are lacking as to the amount of any such effect. On most, if not all, transmission lines the guard wire or wires, if used at all, are placed either above or on a level with the highest power conductors. With one conductor of a three-phase circuit mounted on a pin set in the top of a pole, and the two remaining conductors on a two-pin cross-arm beneath, in the method most frequently adopted for transmission lines of very high voltage, it is obviously impracticable to put guard wires either above or on a level with the power circuits. In the latest transmissions there is a strong tendency to omit guard wires entirely and rely on lightning arresters for protection.

Lightning arresters are wrongfully named, for their true purpose is not to arrest or stop lightning, but to offer it so easy a path to the ground that it will not force its way through the insulation of the line or of machinery connected to the system. The requirements of a lightning arrester are in a degree conflicting, because the resistance of the path it offers must be so low as to allow discharges of atmospheric electricity to earth and so high as to prevent any flow of current between the transmission lines. In other words, the insulation of the line conductors must be maintained at a high standard in spite of the connection of lightning arresters between each conductor and the earth; but the resistance to the arrester must not be so high that lightning will pierce the insulation of the line or machinery at some other point. When a lightning discharge takes place through an arrester the resistance which the arrester offers to a flow of current is for the moment greatly reduced by the arcs which the lightning sets up in jumping the air-gaps of the arrester. Each wire of a transmission circuit must be connected alike to arresters, and the paths of low resistance through arcs in these arresters to the earth would obviously short-circuit the connected generators unless some construction were adopted to prevent this result. In some early types of lightning arresters magnetic or mechanical devices were resorted to in order to break arcs formed by the discharge of lightning.

The type of lightning arrester now in common use on transmission lines with alternating current includes a row of short, parallel, brass cylinders mounted on a porcelain block and with air-gaps of one-thirty-second to one-sixteenth of an inch between their parallel sides. The cylinder at one end of the row is connected to a line wire and the cylinder at the other end to the earth, when a 2,000 or 2,500-volt circuit is to be protected. For higher voltages a number of these single arresters are connected in series with each other and with the free ends of the series to a line wire and to the earth, respectively. Thus, for a 10,000-volt line, four or, better, five single arresters are connected in series to form a composite arrester for each line conductor. For any given line voltage the number of single arresters going to make up the composite arrester should be so chosen that the regular working voltage will not jump the series of air-gaps between the little brass cylinders, but yet so that any large rise of voltage will be sufficient to force sparks across these gaps. A variation of this practice by one large manufacturing company is to mount the group of single arresters on a marble board in series with each other and with an adjustable air-gap. This gap is intended to be so adjusted that any large increase of voltage on the lines will be relieved by a spark discharge. An arrester made up entirely of the brass cylinders and air-gaps has the disadvantage that an arc once started between all the cylinders by a lightning discharge so lowers the resistance between each line wire and the earth that the generating equipment is short-circuited and the arcs may not cease with the escape of atmospheric electricity. To avoid this difficulty it is the practice to connect a conductor of rather large ohmic resistance such as a rod of carborundum in series with the brass cylinders and air-gaps of lightning arresters. This resistance should be non-inductive so as not to offer a serious obstacle to lightning discharge, and its resistance should be great enough to prevent a flow of current from the generators that will be large enough to maintain the arcs started in the arrester by the lightning discharge. Accurate data are lacking as to the amount of this resistance that should be employed with arresters for any given voltage. As a rough, approximate rule it may be said that in some cases good results will be obtained with a resistance in ohms in series with a group of lightning arresters that represents one per cent of the numerical value of the line voltage. That is, for a 10,000-volt line the group of arresters for each wire may be connected to earth through a resistance of, say, 100 ohms, so that if the generator current follows the arc of a lightning discharge through the arresters it must pass through a fixed resistance of 200 ohms in going from one line wire to another. This rule is given merely as an illustration of the resistance that will work well in some cases, and should not be taken to have a general application. If the resistance connected in series with lightning arresters is high, the tendency is a little greater for lightning to go to earth at some point in the apparatus where the insulation is low. If only a small resistance is employed to connect lightning arresters with the earth, the danger is that arcs formed by lightning discharges will be followed and maintained by the dynamo currents. In one make of lightning arrester the row of little brass cylinders is connected at the ends to carbon rods which form a resistance for the purpose just mentioned. Two of these carbon rods are contained in each arrester for 2,000 or 2,500 volts, and the resistance of each rod may be anywhere from several score to several hundred ohms as desired. This form of arrester may be connected directly from line to earth without the intervention of any outside resistance, since the carbon rods may easily be given all the resistance that is desirable.

One of the most important features in the erection of a lightning arrester is its connection to earth. If this connection is poor it may render the arrester useless so far as protection from lightning is concerned. It need hardly be said that ground connections formed by driving long iron spikes into the walls of buildings or into dry earth are of slight value as far as protection from lightning is concerned. A good ground connection for lightning arresters may be formed with a copper or galvanized iron plate, which need not be over one-sixteenth of an inch thick, but should have an area of, say, ten to twenty square feet. This plate may be conveniently made up into the form of a cylinder and should have a number of half-inch holes punched in a row down one side into which one or more copper wires with an aggregate area equal to that of a No. 4 or No. 2 wire, B. & S. gauge, should be threaded and then soldered. This plate or cylinder should be placed deep enough in the ground to insure that the earth about it will be constantly moist, and the connected copper wire should extend to the lightning arresters. It is a good plan to surround this cylinder with a layer of coke or charcoal.

A good earth connection for lightning arresters may be made through large water-pipes, but to do this it is not enough simply to wrap the wire from the lightning arresters about the pipe. A suitable contact with such a pipe may be made by tapping one or two large bolts into it and then soldering the wires from lightning arresters into holes drilled in the heads of these bolts. A metal plate laid in the bed of a stream makes a good ground.

With some of the older types of lightning arresters it was the custom to insert a fuse between the line wire and the ground, but this practice defeats the purpose for which the arrester is erected because the fuse melts and leaves the arrester disconnected and the circuit unprotected with the first lightning discharge. The modern arresters for alternating-current circuits are made up of a series of metal cylinders and short air-gaps and are connected solidly without fuse between line and earth.

It was once the practice to locate lightning arresters almost entirely in the stations, but this has been modified by experience and consideration of the fact that as the line acts as a collector of atmospheric electricity, paths for its escape should be provided along the line. Consideration fails to reveal any good reason why lightning that reaches a transmission line some miles from a station should be forced to travel to the station, where it may do great damage before it finds an easy path to earth. It is, therefore, present practice to connect lightning arresters to each wire at intervals along some lines as well as at stations and sub-stations. The main purpose of arresters is to offer so easy a path to earth that lightning discharges along the lines will not flow to points of low insulation in generators, transformers, or even the line itself. Practice is far from uniform as to the distance between lightning arresters on transmission lines, the distances varying from less than one to a large number of miles apart. In general the lines should be provided with lightning arresters at least where they run over hilltops and at any points where lightning strokes are unusually frequent. Where a long overhead line joins an underground cable arresters should always be connected, and the same is true as to transformers located on the transmission line. The multiplication of arresters along pole lines should be avoided as far as is consistent with suitable protection, because every bank of arresters may develop a permanent ground or short-circuit, unless frequently inspected and kept clean and in good condition.

Arresters, besides those connected along the lines, should be located either in or just outside of stations and sub-stations. If the buildings are of wood, the arresters had better be outside in weather-proof cases, but in brick or stone buildings the arresters may be properly located near an interior wall and well removed from all other station equipment. Transmission lines, on entering a station or sub-station, should pass to the arresters at once and before connecting with any of the operating machinery.

To increase the degree of protection afforded by lightning arresters choke-coils are frequently used with them. A choke-coil for this purpose usually consists of a flat coil of copper wire or strip containing twenty to thirty or more turns and mounted with terminals in a wooden frame. This coil is connected in series with the line wire between the point where the tap for the lightning arrester is made and the station apparatus. Lightning discharges are known to be of a highly oscillatory character, their frequency being much greater than that of the alternating currents developed in transmission systems. The self-induction of a lightning discharge in passing through one of these choke-coils is great, and the consequent tendency is to keep the discharge from passing through the choke-coil and into the station apparatus and thus to force the discharge to pass to earth through the lightning arrester. The alternating current employed in transmission has such a comparatively low frequency that its self-induction in a choke-coil is small. Increased protection against lightning is given by the connection of several groups of lightning arresters one after another on the same line wire at an electric station. This gives any lightning discharge that may come along the wire several paths to earth through the different groups of arresters, and a discharge that passes the first group will probably go to earth over the second or third group. In some cases a choke-coil is connected into a line wire between each two groups of lightning arresters as well as between the station apparatus and the group of arresters nearest thereto.

An electric transmission plant at Telluride, Col., where thunder-storms are very frequent and severe, was equipped with arresters and choke-coils of the type described, and the results were carefully noted (vol. xi., A. I. E. E., p. 346). A small house for arresters and choke-coils was built close to the generating station of this system and they were mounted therein on wooden frames. Four choke-coils were connected in series with each line wire, and between these choke-coils three lightning arresters were connected, while a fourth arrester was connected to the line before it reached any of the choke-coils. These arresters were watched during an entire lightning season to see which bank of arresters on each wire discharged the most lightning to earth. It was found that, beginning on the side that the line came to the series of arresters, the first bank of arresters was traversed by only a few discharges of lightning, the second bank by more discharges than any other, the third bank by quite a large number of discharges, and the fourth bank seldom showed any sign of lightning discharge. Over the second bank of arresters the lightning discharges would often follow each other with great rapidity and loud noise. The obvious conclusion from these observations seems to be that three or four banks of lightning arresters connected in succession on a line at a station together with choke-coils form a much better protection from lightning than a single bank. At the plant in question, that of the San Miguel Consolidated Gold Mining Company, the entire lightning season after the erection of the arresters in question was passed without damage by lightning to any of the equipment. During the two lightning seasons previous to that just named the damage by lightning to the generating machinery at the plant had been frequent and extensive.

A good illustration of the high degree of security against lightning discharges that may be attained with lightning arresters and choke-coils exists at the Niagara Falls plants and the terminal house in Buffalo, where the step-up and step-down transformers have never been damaged by lightning though the transmission line has been struck repeatedly and poles and cross-arms shattered (vol. xviii., A. I. E. E., p. 527). This example bears out the general experience that lightning arresters, though not an absolute protection, afford a high degree of security to the apparatus at electric stations.

Lightning arresters are in some cases connected across high-voltage circuits from wire to wire so that the full line pressure tends to force a current across the air-gaps. The object of this practice is to guard against excessive voltages on the circuit such as might be due to resonance. In such a case, as in that where arresters are connected from line wire to earth as a protection against lightning, the number of air-gaps should be such that the normal line voltage will not force sparks across the air-gaps and thus start arcs between the cylinders.

The number and total length of air-gaps in a bank of arresters necessary to prevent the formation of arcs by the regular line voltage depends on a number of factors besides the amount of that voltage.

According to the report of the Committee on Standardization of the American Institute of Electrical Engineers, the sparking distances in air between opposed sharp needle points for various effective sinusoidal voltages are as follows (vol. xix., A. I. E. E., p. 1091):

It may be noted at once from this table that the sparking distance between the needle points increases much faster than the voltage between them. Thus, 20,000 volts will jump an air-gap of only an inch between the points, but seven times this pressure, or 140,000 volts, will force a spark across an air-gap of 13.95 inches. Two cylinders or other blunt bodies show smaller sparking distances between them at a given voltage than do two needle points, but when a number of cylinders are placed in a row with short air-gaps between them the aggregate length of these gaps that will just prevent the passage of sparks at a given voltage may be materially greater or less than the sparking distance of that voltage between needle points. It has been found by experiment that the numbers one-thirty-second-inch spark-gaps between cylinders of non-arcing alloy necessary to prevent the passage of sparks with the voltages named and a sine wave of electromotive force are as follows (vol. xix., A. I. E. E., p. 1026):

According to these data, only ten air-gaps of one-thirty-second of an inch each and 0.3125 inch combined length are required between cylinders to prevent a discharge at 10,000 volts, though opposed needle points may be 0.47 inch apart when a spark is obtained with this voltage. On the other hand, eighty air-gaps of one-thirty-second of an inch each between cylinders of non-arcing metal, or a total gap of 2.5 inches, are necessary to prevent a discharge at 29,000 volts, though 30,000 volts can force a spark across a single gap of only 1.625 inches between opposed needle points.

Under the conditions that existed in the test just recorded the pressure at which the aggregate length of one-thirty-second of an inch air-gaps that just prevents a discharge equals the single sparking distance between needle points seems to be about 18,000 volts.

The object of dividing the total air-gap in a lightning arrester for lines that carry alternating current up into a number of short gaps is to prevent the continuance of an arc by the regular generator or line current after the arc has been started by a lightning discharge. As soon as an electric spark leaps through air from metal to metal, a path of low electrical resistance is formed by the intensely heated air and metallic vapor. If the arc thus formed is, say, two inches long it will cool a certain amount as the passing current grows small and drops to zero. If, however, this total arc of two inches is divided into sixty-four parts by pieces of metal, the process of cooling as the current decreases will go on much more rapidly than with the single arc of two inches because of the great conducting power of the pieces of metal. As an alternating current comes to zero twice in each period, the many short arcs formed in an arrester by a lightning discharge are so far cooled during the small values of the following line current that the resistance quickly rises to a point where the regular line voltage cannot continue to maintain them, if the arrester is properly designed for the system to which it is connected. In this way the many-gap arrester destroys the many small arcs started by lightning discharges that would continue and short-circuit the line if they were combined into a single long arc.

When an electric arc passes between certain metals like iron and copper a small bead is raised on their surfaces. If these metals were used for the cylinders of arresters the beads on their surface would quickly bridge the short air-gaps. Certain other metals, like zinc, bismuth, and antimony, are pitted by the passage of arcs between their surfaces. By suitable mixture of metals from these two classes, an alloy is obtained for the cylinders of lightning arresters that pits only slightly and is thus but little injured by lightning discharges. After long use and many discharges an arrester of the class here considered gradually loses its power to destroy electric arcs. This may be due to the burning out of the zinc and leaving a surface of copper on the cylinders.

Aside from the structure of an arrester and the normal voltage of the circuit to which it is connected, its power to destroy arcs set up by lightning discharges depends on the capacity of the connected generators to deliver current on a short-circuit through the gaps, and upon the inductance of the circuit. The greater the capacity of the generators connected to a system the more trying are the conditions under which arresters must break an arc because the current to be broken is greater. So, too, an increase of inductance in a circuit adds to the work of an arrester in breaking an arc.

An arc started by lightning discharge at that period of a voltage phase when it is at or near zero is easily destroyed by the arrester, but an arc started at the instant when the regular line voltage has its maximum value is much harder to break because of the greater amount of heat generated by the greater current sent through the arrester. For this reason the arcs at arresters will hold on longer in some cases than in others, according to the portion of the voltage phase in which they are started by the lightning discharge. Lightning discharges, of course, may occur at any phase of the line voltage, and for this reason a number of discharges must take place before it can be certain from observation that a particular arrester will always break the resulting arc. Between twenty-five and sixty cycles per second there is a small difference in favor of the latter in the power of a given arrester to break an arc, due probably to the fact that more heat in the arcs is developed per phase with the lower than with the higher frequency.

It will now be seen that while increase of the regular line voltage requires a lengthening of the aggregate air-gap in lightning arresters to prevent the formation of arcs by this voltage alone, the increase of generating capacity requires more subdivisions of the total air-gap in order that the arcs maintained by the larger currents may be cooled with sufficient rapidity. These two requirements are to some extent conflicting, because the subdivision of the total air-gaps renders it less effective to prevent discharges due to the normal line voltage, as has already been shown. The result is that the more an air-gap is subdivided in order to cool and destroy arcs that have been started by lightning, the longer must be the aggregate air-gap in order to prevent the development of arcs directly by the normal line voltage.

Furthermore, the practical limit of subdivision of the air-gap is soon reached because of the difficulty of keeping very short gaps clean and of nearly constant length. As a resistance in series with an arrester cuts down the generator current that can follow a lightning discharge, such a resistance also decreases the number of air-gaps necessary to give an arrester power to destroy arcs on a particular circuit.

The increase of resistance in series with a lightning arrester as well as the increase in the aggregate length of its air-gaps subjects the insulation of connected apparatus to greater strains at times of lightning discharge. On systems of large capacity the number and aggregate length of air-gaps in arresters necessary to destroy arcs must be greater than the number or length of these air-gaps necessary to prevent the development of arcs by the normal line voltage, unless a relatively large resistance is connected in series with each arrester. To reduce the strains produced on the insulation of line and connected apparatus under these conditions by lightning discharges, a resistance is connected in shunt with a part of the air-gaps in one make of lightning arrester. The net advantage claimed for this type of arrester is that a lower resistance may be used in series with all the air-gaps than would otherwise be necessary. One-half of the total number of air-gaps in this arrester are shunted by the shunt resistance and the series and shunt resistance are in series with each other. Only the series air-gaps or those that are not shunted must be jumped in the first instance by the lightning discharge, which thus passes to earth through these air-gaps and the shunt and series resistance in series. An arc is next started in the shunted air-gaps, and this arc is in turn destroyed because the shunt weakens the current in these gaps. This throws the entire current of the arc through the series air-gaps and the shunt and series resistance all in series with each other. As the shunt resistance is comparatively large, the current maintaining the arc in the series air-gaps is next so reduced that this arc is broken. Taking the claims of its makers just as they stand, the advantages of the shunted air-gaps are not very clear. The series air-gaps alone must evidently be such that the normal line voltage will not start an arc over them, and these same series gaps must be able to break the arcs of line current flowing through them and the shunt and series resistance all in series. Evidently the greatest strain on the insulation of the line and apparatus occurs at the instant when the lightning discharge takes place through the series gaps and the shunt and series resistances all in series with each other.

Why develop subsequent arcs in the shunted air-gaps? Why not throw the shunted air-gaps away and combine the shunt and series resistances?