(Continued from page 187)

The current survey consists of measurements of current flow on pipes at selected locations in the piping system. To determine the current flowing on a pipe, the voltage drop between two points on a continuous length of pipe is measured by means of a millivoltmeter. From this voltage drop and from the resistance of the included pipe, which may be obtained from published tables, the strength of the current is computed. To measure this drop it is necessary to expose the pipe and to make good electrical contact between the millivoltmeter leads and the pipe. The best contact for these current measurements is obtained by soldering the connecting wires directly to the pipe or to two brass plugs screwed into the pipe. This method is particularly advantageous when readings are to be taken over a considerable time, and it is customary to use rubber-covered wires and bring them to the surface of the street, leaving the ends in service boxes which then form permanent stations for electrical measurements. This is exceedingly convenient because it is then possible to make current measurements on the pipe without again making an excavation. Such permanent contact wires are illustrated in Fig. 3. It should be noted that small potential differences, such as 0.1 millivolt or less, may be caused by local galvanic or thermal action. Where such small values are found in a test for drop on a pipe a careful investigation should be made to ascertain whether the observed potential difference is actually drop due to current flow.

Since current destroys metal only where it leaves for soil, it is important to know where the current leaves a pipe. Current measurements on pipes are, therefore, frequently made at two or more stations simultaneously in order to determine the change, in current on the pipe between the stations. For these current measurements indicating or recording millivoltmeters are employed. By comparing the characteristic variations of currents and voltages measured for twenty-four-hour periods at selected locations with the characteristics of the neighboring electric railways, it is often possible to identify the source of the stray current flowing on a pipe. From a study of the results of the potential and current surveys it can be determined where current is leaving the piping, and at a number of such points excavations should then be made and the exposed pipe examined with a test hammer for corrosion of the ion. Where corrosion and pitting is found at points where current is leaving the pipe, it may be taken as evidence that at least part if not all of the corrosion has been caused by electrolysis.

Remedial Measures Applied to Pipes.

Attempts have been made to protect underground pipes from electrolysis by insulating them from earth by paints or dips. Practical experience as well as a large number of tests have, however, shown that no dip or paint will permanently protect a pipe against electrolysis in wet soil. It has been found that in most cases the applied coatings have either been completely destroyed by the effects of the wet soil and the electric currents, or defects in the coating have developed, causing concentrated corrosion at such defective spots. Where it is attempted to apply a heated material like pitch or asphaltum to a cold pipe, it is impossible to completely cover the pipe. Pitch and similar compounds have been applied to pipes with wrappings of jute or of some similar material. A number of layers can be applied in this way so as to build up any desired thickness of insulating covering. Such covering if sufficiently thick will afford protection against electrolysis, provided that it is mechanically perfect. The great difficulty in practice is to apply such a covering without leaving defective spots through which moisture will have access to the metal of the pipe.

Pipes which are covered with imperfect insulating coatings, or coverings exposing bare spots of metal, are in much greater danger from electrolysis where positive to earth than are bare pipes, for the reason that the stray currents will leave only from these bare spots and here produce concentrated corrosion. One form of insulating covering which appears to afford complete protection is a layer of from one to two inches of a material like coal tar pitch, parolite or asphaltum, of such a grade that it is not brittle, and so will not crack, but yet is hard enough to remain in place. Embedding a pipe in cement or concrete, even if this is several inches in thickness, will not protect it from electrolysis, because damp cement or concrete is an electrolytic conductor. Insulating covering, even if imperfect, is useful on pipes in negative districts where current flows from earth to the pipes, because this covering will increase the resistance from earth to the pipes and thereby correspondingly reduce the amount of current reaching the pipes. Current flow on metallic pipe lines can be practically prevented by using a sufficient number of insulating joints. Insulating joints in pipe lines should not be confined to the positive areas, but should be installed in all places along the pipe line where there is any considerable potential gradients in the earth parallel to the pipe. The frequency with which insulating joints must be installed in a pipe line in order to assure reasonable protection from electrolysis depends upon the potential gradient through earth and upon the electrical resistivity of the earth. Where service pipes are endangered by current which flows to them cither from the main or from house piping, such current flow can be prevented and the service pipes protected by placing insulating joints in them at the main or in the building, or at both locations. Insulating covering and insulating joints can be applied in special cases to individual pipes, but cannot ordinarily be applied in an extensive manner to a piping network.

A method of mitigating electrolysis which has frequently been employed in this country is pipe drainage. This consists of connecting the pipes to the railway return circuit at a sufficient number of points to render the pipes at all points negative to the electric railway tracks. Once electrical drainage is applied, it is usually found that thereafter everything is left to take care of itself, and no attention is paid to the railway return circuit. When electrical drainage is applied to a single system of underground pipes, without making a complete investigation of the effects of possible high resistance joints, etc., the installation may be made at relatively small cost, and when so applied it usually relieves the acute danger from electrolysis in the immediate neighborhood where the drainage connections are made. Both of these considerations have served to favor the electrical drainage system. However, a single drained underground piping system becomes a source of serious danger to other systems. If electrical drainage is applied comprehensively to all underground metallic systems, it will not only be found very expensive to install, but, likewise, expensive to maintain. A further and perhaps the most serious objection to the drainage system is that sufficiently complete tests cannot be. practically made to determine whether the drained system is safe or is still in danger from electrolysis.

In future installations of underground piping systems in the neighborhood of electric railways, precautions should be taken to minimize the flow of stray current to the pipes. To this end the pipes should be laid as far from the electric railway tracks as practicable. Metallic contacts with the tracks, .such as may exist at the iron gate boxes used in water piping systems, must be carefully avoided. Where the pipes cross steel bridges carrying electric railway tracks in metallic contact with the bridge structure, the pipes should be supported on wooden blocks or otherwise insulated from the metal of the bridge structure. Insulating joints should be installed at the entrance of pipes to car barns.

Remedial Measures Applied to Electric Railways.

The only way to entirely prevent electrolysis from stray railway currents is to prevent leakage of currents to earth from electric railway systems by the use of a separate and completely insulated return conductor, instead of using the running tracks as part of the return circuit. This is accomplished with the double underground trolley system, and with the double overhead trolley system. These systems, while entirely effective in preventing electrolysis, have, however, not been generally adopted, probably because of the added expense and of the added complication involved over the ordinary single-trolley system, While leakage of current from singletrolley electric railways cannot be entirely Drevented by any methods that can be applied to these railways, the amount of stray current produced by a single-trolley railway can by adequate measures be reduced to any desired minimum values. The direct cause of stray currents front electric railways is voltage drop in the running tracks. It is, therefore, clear that by reducing this voltage drop the stray currents leaking from such tracks will be correspondingly reduced. The reduction of stray currents through earth can best be accomplished by the following means, given in the order of their importance:

  1. By increasing the number of directcurrent supply stations in systems extending over large areas, so as to reduce the radius to which any one station supplies current, and also by supplying all of the railways in any locality from one supply station in this locality.
  2. By increasing the electrical conductance of the tracks, through the use of heavy rails, through the use of low-resistance rail joint bonds and cross bonds, and through the interconnection of the electric railway tracks of all systems, where these come close together.
  3. By removing current from the tracks by insulated return feeders, and by maintaining the negative bus-bar insulated from ground at the supply station, in all cases where the voltage drop in the tracks would otherwise be excessive. This arrangement is known as the insulated track feeder system, or the insulated return feeder system.
  4. By increasing the resistance between tracks and earth as much as practicable, through draining the roadbed, and, on private right-of-way, through maintaining the rails out of contact with earth.

Where a road operates on a private rightof-way, the rails can often be practically insulated from ground and the escape of current from the tracks prevented. For surface roads this can be accomplished by placing the rails on wooden ties above ground and using broken stone for ballast and keeping the rails out of contact with ground. In the case of railway lines operating on elevated structures, the rails can be fastened to wooden ties and kept out of contact with the structures. These rails, supplemented where necessary with negative feeder cables, also insulated from the structure, can then be used for the return conductor.

It has been proposed to employ a 3-wire system for distributing current to an electric railway. The practical effect of the 3-wire system is to very greatly reduce the track voltage drop and correspondingly reduce the total amount of corrosion from electrolysis. In addition to this the remaining corrosion from electrolysis is distributed over widely scattered areas.

The insulated track feeder system, in conjunction with proper track bonding, usually affords the most feasible means for reducing track voltage drop and thereby reducing stray currents through earth in an existing electric railway. In this system feeders insulated from earth are connected from the negative bus-bar to selected points on the track network. The stray currents which leak from the rails to earth concentrate in earth and on the underground piping in the neighboring of the railway power station, where they must return to the rails to get back to the negative bus-bar. If connection between the negative bus-bar and the rails at the power station is removed, and the currents collected from the rails at points near the center of each railway line by means of insulated track feeders, the concentration of current in the neighborhood of the power station is entirely removed. With this arrangement the current used by each individual line tends to flow away from the rails at both ends of this line, and toward the rails near the center of the line. The insulated track feeder system is frequently confused with the system of paralleling the tracks with return feeders, which has been most commonly used in American electric railways. From the standpoint of reducing track voltage drop the two systems are, however, totally different. With copper feeders paralleling the tracks, the voltage drop in the tracks ⅛ reduced only in the proportion that the conductance of the track circuit is increased. With the insulated track feeder system the voltage drop in the insulated feeders does not occur in the tracks nor in the earth, and therefore may be made as high as economy dictates.


Experience shows that an increasing amount of damage by electrolysis is occurring on underground piping systems in many localities throughout the country where adequate measures have not been taken to reduce this damage. The principal and generally the sole sources of stray electric currents causing this damage are the single-trolley direct-current electric railways employing the running tracks in contact with earth as part of the return circuit. Experience extending over many years in foreign countries and over ten years in this country has shown that practicable and economical methods of construction can be applied to such electric railway systems which will remove acute dangers from stray currents to underground piping systems and which will greatly reduce the electrolysis danger in all cases, and in most cases will make this danger negligible. Mitigating methods applied to underground pipes fail to attack the source of the trouble and should be applied only in special cases, if at all, and then only after adequate methods of minimizing the production of stray currents have been applied to the railway system. Metallic connections from underground water pipes to the railway return circuit which cause these pipes to become a substantial part of this return circuit are inadequate for the protection of the pipes and are frequently dangerous. Such connections greatly increase current flow on pipes, and while they may afford local protection, they generally distribute electrolysis troubles to other localities where they are more difficult to find, and in this way frequently give a false impression of immunity. Metallic connections from water pipes to the railway return circuit should generally not be permitted and in no case unless a careful study of conditions has shown that no serious danger will be produced. Such connections should never be applied to an underground piping system as the principal means of electrolysis mitigation.

In view of the fact that the railway companies in common with the pipe-owning companies are public utilities operating under public franchises and utilizing city streets, it is the duty of both of these utilities to co-operate in order that the causes and extent of any danger from stray currents can be more readily ascertained. Further, the satisfactory solution of the electrolysis problem is one which requires the co-operation of all of the interests concerned. Electrolysis is an engineering problem and can be handled by engineering methods in such a manner that no hardship need be imposed nor should be imposed on any one. There is no reason why the negative feeder system should not be laid out along the same engineering lines as the positive feeder system. I think that if the electric railway companies would realize this and the owners of underground properties would co-operate in a practical way, we could obtain a satisfactory and practical solution of the electrolysis problem. For instance, it often happens that the judicious installation of a few insulating joints will save a lot of money in railway track feeders, and in such cases such joints should be installed.

A most important step towards securing the co-operation, which is absolutely necessary in order to obtain adequate and permanent relief from electrolysis, has been made by the formation of the American Committee on Electrolysis. This committee includes representatives of the electric railway, water, gas, electric light and telephone interests. This committee was organized in 1913 and has completed a preliminary report setting forth the facts regarding electrolysis, upon which the representatives of all of the varied interests have agreed. The committee has already accomplished a great deal towards producing a closer co-operation between the interests owning the electric railways and those owning the underground structures, and it is to be hoped that the future work of this committee will result in the unanimous adoption of recommendations which will reasonably safeguard underground piping systems against electrolysis.




Electrolysis is the process of decomposing a chemical compound by means of an electric current. Electrolysis, in the sense in which it will be discussed here, and in which you are particularly interested, refers to the corrosion of underground metallic structures, such as iron and lead pipes, by stray electric currents which reach these structures and flow to surrounding soil. Soil, when entirely dry, practically does not conduct electric current. Pure water likewise has such a high electrical resistance, compared with ron or lead, that it may be considered a a non-conductor. Water is, however, readily made conducting by the addition of even very small amounts of salts, and conduction through water is, therefore, always electrolytic. Soil in its natural state is always moist, and on account of dissolved salts, such as chlorides, nitrates, etc., which are always present, is an electrolytic conductor. Electric current may be conducted by metallic conduction or by electrolytic conduction. Metallic conduction occurs when an electric current passes through a metal, and is characterized by the fact that no chemical change is produced in the conductor, the only effect being the production of heat. Electrolytic conduction occurp when an electric current passes through an electrolyte. When an electric current flows from a pipe or other metallic structure to surrounding soil, chemical decomposition of the metal will take place resulting in corrosion of the pipe or structure. Concrete, when buried in earth, is moist, and it then becomes an electrolytic conductor, so that an electric current flowing from iron to surrounding concrete will corrode the non by electrolysis. The mass of a metal corroded by electrolysis in a given time depends only on the “current,” and, with the current dcnsilies and other conditions usually found in the case of underground pipes, is equal to that calculated by Faraday’s law. Iron is oxidized by electrolysis at the rate of approximately 20 pounds per year for every ampere of current flowing from the iron to surrounding soil. Under some conditions, particularly with every small current densities, this corrosion may be considerably greater, while with larger current densities than the above, this corrosion may be considerable less than the theoretical rate. The actual rale may vary’ in practice from one-half to one and one-half times the theoretical rate. Lead is oxidized by electrolysis under ordinary conditions in soil at a rate equal to approximately 74 pounds for every ampere of current leaving the lead in one year, and this theoretical rate may also vary somewhat in practice. The amount of corrosion produced bv electrolysis is independent of the voltage, except in so far as this determines the amount of current flowing, and the smallcst fraction of a volt can produce corrosion from electrolysis under suitable conditions.

Fig. 2.—Diagram Showing Stray Railway Currents With Assumed Distribution of Potentials Caused by These Currents.

The rapid corrosion by electrolysis from external currents is usually localized and results in pitting of the metal. Such pitting may, however, in some cases also result from ordinary soil corrosion, so that the appearance of a corroded metal structure does not by itself afford conclusive evidence as to whether or not the corrosion was produced by electrolysis from external electric current. Where the direction of current flow between an underground pipe and surrounding soil reverses more or less continually, it has been found that the corrosion occurring during the time that current flows from the pipe is largely offset by a reversed action which occurs during the time that current flows to the pipe. The resultant corrosion by electrolysis from periodically reversed direct current is for this reason much less than when the current always flows in the same direction, and this corrosion decreases w’ith increasing frequency of reversal. Investigations have shown that even if such reversals occur only once in twenty-four hours the actual amount of corrosion for iron is only about one-fourth of what would occur if the same amount of current always flowed from the pipe to surrounding soil. When an alternating current of commercial frequency flows between a pipe and surrounding soil, the amount of corrosion produced by electrolysis is of the order of 1 per cent, or less of the corrosion which would be produced by an equal direct current flowing continuously from the pipe to the soil.

Sources of Stray Currents Which May Produce Electrolysis.

Electrical distribution systems which are grounded at two or more points will, by the law of divided circuits, cause currents, called “stray currents,” to shunt through the earth between the grounded points, and these stray currents frequently reach underground metallic structures and Corrode them by electrolysis. In practice, it is found that the most important sources of stray electric currents, which so endanger underground structures, are direct-current electric railways, which use the running tracks in contact with ground for part of the electric circuit. For such railways, it is the common practice to supply current to the cars from an overhead trolley wire Or from a third rail, and to return this current to the power station through the running tracks, supplemented in large systems by return feeders. A single-trollev electric railway is illustrated in Fig. 1. In this figure the path of the electric current from the positive terminal of the generator through the circuit and back to the negative terminal is shown. In such a trolley system the running tracks consist of rail lengths mechanically fastened together by fish plates of steel bridging across the rail ends and bolted to both rads. Such fish plates while mechanically fastening the rail lengths together do not afford good electrical connections between the • successive rail lengths. For this reason, copper wires or straps, called rail bonds, are generally used to bridge across the abutting ends of the rail lengths for the purpose of affording a good electrically conducting path between successive rail lengths. The two rails of a single-track road, or the four rails of a double-track road, are also generally connected together at frequent intervals by cross bonds so that the two or the four rails may be available for the return of current*. Instead of using copper rail bonds, the rail ends are sometimes welded together, or soft steel plates are welded across each side of the abutting rail ends, thus forming both a strong mechanical and a good electrically conducting connection between the successive rail lenths.

In the simplest form of single-trolley railway, shown in Fig. 1, the rails are connected to the negative terminal of the generator at the power station, and the only path for current to return to the power station is by way of the running tracks. If the running tracks are laid upon wooden ties above ground with broken stone for road ballast, as is common on steam railroads, the rails do not come in direct contact with ground, and the return current will be practically confined to the running tracks. If, however, the running tracks are laid below ground so that the top of the rails is level with the surface of the street, as is common in cities, then the rails will be exposed for a considerable area to contact with ground. If the tracks are laid on a concrete base a considerable area of the rails will similarly be in contact with the concrete. Since both damp soil and damp concrete are under ordinary conditions conductors of electricity, part of the current returning through the rails will shunt from the rails through the neighboring earth, as is illustrated diagrammatically in Fig. 2. It will be seen that w’ith the usual connection of positive terminal of the generator to the trolley wire, and the negative terminal to the rails near the power station, the current will leave the rails for earth at points distant from the power station, and return to the rails, in the neighborhood of the pow’er station, in its path back to the negative terminal of the generator. Since every electric circuit must be completely closed, all current escaping through earth must again leave earth to return to the dynamo so as to complete the electric circuit. Where underground metallic structures, such as gas or water pipes, lie in earth in the path of these, stray currents, and where these pipes have electrically conducting joints, such as lead-calked joints or screw-coupling joints, current w’ill flow from earth to such pipes and flow on such pipes towards the power, station. In the neighborhood of the power station this current will leave the pipes to return through earth and the tracks to the negative terminal of the generator, as shown in Fig. 2.

If the negative terminal of the generator or negative bus-bar is connected to the rails at points some distance from the power station by means of insulated negative return feeders, then at such connection-points the rails will be rendered negative in potential to earth, and currents will tend to flow from underground pipes through earth to return to the rails in the neighborhood of these connections. Such stray railway currents on pipes will, therefore, tend to leave these pipes to return to the rails in all regions where these rails are connected to return feeders. As a matter of practice it is often found that where the rails in contact with earth are used alone for the return of current, a considerable portion of the total current leaks from rails through earth.

Alternating currents have been used for some years past in a number of electric railways employing the running tracks as a part of the electric circuit, and where these tracks are in contact with earth, stray alternating currents through earth are produced. Where an alternating current lions from iron or lead to surrounding soil, corrosion from electrolysis may also be produced, but this proceeds at a relatively very slow rate, as a ready explained. With alternating currents, electrolysis is, however, produced at the two electrodes, instead of at one electrode only, as with direct current. So far as the writer is aware, no damage from electrolysis due to such stray alternating railway currents has been reported to this date. This may be due to the slow rate at which corrosion is produced by alternating currents, together with the fact that most of these railways are of relatively recent installation. It may also be due to the fact that stray direct currents are nearly always present with the alternating currents, and the effects of these direct currents may have inhibited or masked the effect of the alternating currents. It is, therefore, not possible at this time to draw a positive conclusion as to the danger from stray alternating currents.

General Effects of Stray Electric Currents on Underground Piping.

The current flowing through the rails from the trolley cars back to the power station produces in these rails a drop in potential; that is to say, points in the rails away from the power station have a positive potential with reference to the rails at the power station. Since potentials are measured relatively it is convenient to consider the negative terminal of the dynamo, which is assumed connected to the rails at the power station, as at zero potential. The distribution of potentials in the rails of a simple electric railway system and in the underground piping is illustrated in Fig. 2, in which convenient values have teen assumed. It will be noted that the underground pipes are negative to the rails at points away from the power station, and positive to the rails near the power station. It is also seen that the negative potential of the pipe, plus the drop on the pipe, plus the positive potential of the pipe, equals the drop in the rails. In the case assumed there is a potential difference of 550 volts maintained” at the power station; of this, 10 volts is lost in the trolley wire, 520 volts is used by the motors of the car, and 20 volts is left to bring the current back to the power station. If the negative bus-bar and the rails at the power station are considered as at zero potential, the rails at the car in the assumed case will have a potential of 20 volts. Thus, for practical purposes, the earth with its underground pipes is subjected to a potential difference of 20 volts, and the amount of stray current produced is. that due to these 20 volts. If the rails are laid in the usual way. that is, in contact with ground, the 20 volts in the rails will send some shunting current through the earth and on the underground pipe as shown in the diagram. Under the assumed conditions, there is a drop of 8 volts from the rails to the pipe near the car, a drop of 4 volts in the pipe itself, and a drop of 8 volts from the pipe through earth to the rails at the power station. It is, therefore, seen that it is the voltage drop in grounded rails caused by the return current which is the cause of stray currents through earth.

Fig. 1.—Diagram of Single-Trolley Electric Railway, Showing Path of Current from Generator Through Positive Feeders, Trolley Wire, Car and Rails.

From the explanation of metallic and electrolytic conduction given in the first part of the paper, it will be understood that where stray currents flow on underground pipes they do no harm except where they leave the pipes to flow to the surrounding soil. At such points corrosion of the iron from electrolysis will take place. In the simplest case, illustrated in Fig. 2, current flows from rails-through earth to the pipes at points distant from the power station, flow s on the p pcs, and leaves the pipes to return through earth to the rails in the neighborhood of the power station. Where the current flows from the rails to earth, the rails will be corroded, and where the current flows from the pipes to earth, the pipes will be corroded. If the pipe line is a uniform electrical conductor, and the relative arrangements are as show n in Fig. 2, then the pipes will be corroded only in the neighborhood of the power station. If, however, the pipe line is not a uniform conductor, as for instance, if there are one or more high resistance joints in this pipe line, then the current on the pipe will shunt around such high-resistance joints and produce oxidation or corrosion on the positive sides of the joints. Where there are two or more underground piping systems it also frequently happens that current shunts from one system to another through the intervening soil, producing electrolytic corrosion where the current leaves the pipes. Where a direct-current trolley system passes through a town which has an independent piping network, and where the power station supplying the trolley line is in some other locality, then if stray electric currents are produced by the trolley line where it passes through the town, they will flow’ 6n the piping system making this piping system positive to earth and to rails at points nearest the railway power station, and negative at points farthest away from the pow’er station. In this case electrolysis of the piping will be produced at the ends of the piping system which are nearest to the power station.

When cast iron is corroded by electrolysis, the oxides of iron mixed with graphite usually remain in place, leaving the outside appearance of the pipe unchanged. This material resulting from the electrolysis of cast iron usually has the consistency of hard graphite, and can be cut with a pocket knife. There have been many cases in which a cast-iron main was carrying gas or water without any apparent leak, where a light blow with a hammer drove a hole right through the pipe. Here the electrolytic action had corroded the iron entirely through the pipe and the oxide of iron had remained in place, and, together with the surrounding soil, had prevented the pipe from leaking. Whether or not the mixture of iron oxide and graphite resulting from electrolysis remains in place so as to maintain a pipe gasor water-tight, depends upon the surrounding soil conditions. It is, therefore, seen that an underground piping system may be suffering severely from electrolysis without giving any outward sign of the damage. A physical examination with a test hammer is required in the case of cast-iron pipe to establish definitely whether or not it has been damaged by electrolysis. For a given current leaving a pipe, there is practically no difference between cast-iron, wrought-iron and steel in the amount of iron destroyed. The electrical resistivity of cast-iron is, however, about ten times as great as that of wrought-iron or steel, and the usual lead joints in cast-iron pipes also have a resistance which is many times greater than the screw-coupling joints used with wrought-iron and steel pipes. A given voltage drop through earth will, therefore, cause a much smaller current to flow on a cast-iron pipe than on a wrought-iron or steel pipe, thus making cast-iron pipes much less subject to electrolysis than wrought-iron or steel pipes. Electrolysis Surveys on Underground Piping Systems.

Fig. 3.—Permanent Electrical Test Wires Attached to Main and Brought to Surface of Street Through Service Box.

In order to determine the electrolysis conditions of an underground water piping system a potential survey and a current survey are generally made. Where it is also desired to determine w hat remedial measures shou d be applied for protection against electrolysis, tests of the current and voltage distribution in the grounded circuits of the electric railway system by which the stray currents are produced should also be made. In the potential survey, voltage measurements between the water pipes and trolley rails and between the water pipes and all other neighborhoring underground piping and cab c systems are made at a number of selected locations. They should preferably be made simultaneously between all of the structures tested at any given location, and either indicating or recording voltmeters may be employed; but the tests should be made for a sufficiently long period at each location to cover a complete cycle of car operation. Where the water pipe itself is not exposed, connections for these voltmeter measurements may be obtained by means of house service pipes, which are satisfactory for these tests, because the voltmeter has a high resistance and takes only a very small current.

(To be continued)