extent, is considered as a remote ground of negligible resistance. A lowreading ammeter and a voltmeter give the current flowing and the potential difference between the section of the track under test and the remote ground, and from these data the resistance to earth is easily and accurately calculated. By taking several hundred feet of track, the effect of the short leakage paths at the ends of the section is practically eliminated. The resistance so found is for a single-track roadbed, but in the open type of construction the resistance to earth is concentrated largely in the ties and therefore the resistance of double track can be taken as one-half that of single track. This is not true when the rails are imbedded in earth or concrete. In this case the resistance of a double track may be taken as about seventy per cent of that of a single track if only approximate results are required. This method of measuring roadbed resistance necessitates working at night, as does the differential method, since it usually requires several hours to remove and replace the bonds and joint plates on four joints. 3. Location and Testing of High-Resistance Joints in Pipes. In making electrolysis surveys it is often necessary to determine whether or not there are any considerable number of high resistance joints in a given portion of a pipe network. This has been a particularly important test in making investigations of joint electrolysis in pipe systems, and may often be useful in determining upon the method of protection to be used in particular cases. High resistance joints may be most conveniently located by means of potential drop measurements along the pipes. The method usually followed is to drive bars down until they come in contact with the pipe and measure the potential drop on the pipe at such points, the spacing of the points being usually about 100 feet. A series of such measurements is made throughout the entire length of the pipe and the relative magnitudes of the voltage drops on adjacent sections would indicate which, if any, is affected by high resistance in the pipe line. When it has been determined that any particular hundred-foot length includes one or more high resistance joints, this section can be further subdivided by exactly the same procedure until a relatively high drop is obtained between two points less than a pipe length apart, which must include a high resistance joint. By comparing the drop across the high resistance joint with the drop in a measured distance on continuous pipe, the resistance of the joint in terms of equivalent feet of pipe can be obtained. 4. Tracing the Source of Stray Currents. Conditions are often encountered in which stray currents on pipe networks may come from any one of two or more railway lines, and it is important to determine from which line the current is derived. This can be determined in either of two ways. One method is to connect a measuring instrument of the recording type to the pipe under test, which may be connected either to indicate the current flow along the pipe or the potential difference between the pipe and the earth. With the instrument thus connected, a record is obtained, while all railway systems are operating under normal conditions. Then one of the railway systems is shut down for as long a period as practicable. If the shutting down of the plant makes a marked difference in the record, it is an indication that a large part of the stray current at least comes from that particular point. By shutting down the different systems in rotation, a fairly definite knowledge of the source of the stray current may be obtained. Sometimes it will be found that the shutting down of one plant increases the current flow in one direction, while shutting down another plant may give rise to a large current flow in the opposite direction, both currents being larger than when both plants are running. This will indicate that the stray current from the two systems tend to neutralize each other, thus giving rise to better conditions in certain localities than those which prevail when either system is operating alone. The other method of tracing the source of stray current in any particular case consists in the use of two or more recorders, one of which makes a graphic record of the current or voltage, the source of which is to be determined, while the others are used to make simultaneous records of the loads on the various power supply stations which may possibly affect the area in question, or more particularly the loads on certain feeders from those stations. In most cases there will be sufficient similarity between the chart of the stray current and some one of the feeder or station load charts to establish quite definitely the source of the stray current. 5. Location of Unknown Metallic Structures or Connections. It is often desirable to locate metallic connections between pipes and various other structures, such as railway track returns which may often exist without the knowledge of either the pipe or railway company. Two methods are available for doing this: One consists in connecting an external electrical circuit between convenient points on the pipe system and the railway system, and sending between them either an alternating current of audible frequency or a direct current interrupted at audible frequency. An exploring coil is then carried along the pipe system in such position that the alternating or pulsating magnetic field produced by the current superposed on the pipe current will induce an electromotive force in the coil. This can be made audible by the use of a telephone receiver. By this instrument, the path of the current can be traced and in most cases the location of concealed connections to the pipe can be determined. The method works very satisfactorily on relatively simple pipe networks, but in very complicated systems where there are a great many pipes laid in the street it becomes relatively difficult to trace out any particular structure. A method that has been more recently developed and which is considerably more simple than the above, consists in doing away with the additional current superposed on the pipe network, and using the exploring coil and telephone to listen to the commutation not in the railway current carried by the pipe. This method is much to be preferred where the pipe currents are large enough to give sufficient sensitivity. In some cases, however, where the currents on the pipe are very small, the first method may have to be resorted to. III. Interpretation of Results of Electrolysis Surveys No definite rules of procedure can be laid down for the interpretation of the results of electrolysis surveys that can be used except by engineers thoroughly familiar with all the factors involved. The significance of any particular set of readings is so dependent upon other conditions that all factors must be taken into account or else the conclusions are likely to be in error. However, it is desirable to point out certain principles that must be kept in mind even by the experienced engineer in order to arrive at correct conclusions. A. INTERPRETATION OF POTENTIAL MEASUREMENTS 1. Maximum Voltages and Track Gradients. These measurements, when considered in the light of a full knowledge of all conditions, give valuable data on the condition of the railway track system and the concentration of return current on certain sections of track. They also are valuable when considered in the light of the load on the different lines, as they offer a fairly accurate indication of the track losses and the necessity for the use of additional track feeders. When such potential measurements are taken over relatively short lengths of track, such as 1,000 or 2,000 feet, the comparison of such measurements on adjacent sections of track will often reveal bad places in the track that are in need of rebonding. 2. Potential Difference Measurements. Potential difference readings between pipes and railway tracks and between various underground structures are not a quantitative measure of the danger to the affected structures. These readings are valuable in pointing out the general areas in which trouble may be expected to occur and in which more careful search may be made if desired. They are, however, of qualitative significance only. The current leaving a structure for the earth in any locality, which is the real cause of the electrolysis damage, is a function not only of the potential differences but of the resistance of the earth paths. This has been shown to vary throughout extremely wide limits, so that the measurement of potential difference gives no definite quantitative measurement of the extent of the hazard to the pipes. Such measurements are very valuable and have an accurate quantitative significance, however, when used to determine the relative electrolysis conditions under different systems of mitigation If, for example, under a given set of conditions a considerable number of potential difference measurements are made between the various underground structures and then a change is made in the mitigative system, and the same measurements repeated, the two sets of readings may be used to represent the comparative hazard in the two cases. This is true only if the mitigative measures under test are applied exclusively to the railway return system. High potential differences between gas and oil pipes and other metallic structures with which they may come in contact are objectionable especially in confined spaces, such as basements in which explosive mixtures may be encountered. This is because a transient contact between the two structures may cause an arc which may result in fire or explosion. B. INTERPRETATION OF CURRENT MEASUREMENTS ON UNDERGROUND STRUCTURES 1. Relation of Stray Current to Corrosion. The magnitude of the current on an underground structure does not alone afford a measure of the total injury to the structure. If all the current that flows on the pipe is discharged directly into the earth, then the total corrosion will be approximately proportional to the current flow. Even here, however, the rate of damage to the pipes is not only a function of the total weight of metal corroded away, but of the distribution of such corrosion as a result of pitting or of localized discharge from one system to another where they approach close to each other. Further, if there are metallic connections either known or unknown between portions of the pipe networks and the railway tracks, which carry off a large part of the current on the pipe through metallic paths, the total amount of corrosion cannot be determined by measurement of the current flow. For this reason current measurements on pipes should likewise be regarded as having only a qualitative significance in so far as any absolute hazard to the pipes is concerned. If, however, the pipes have no drainage connection and changes are made in the railway track network, the corresponding changes in the currents on the pipes may, if a sufficient number of readings have been taken, indicate the relative improvement in the electrolysis conditions. 2. Relation of Current to Fires and Explosions. In interpreting the significance of current measurements on gas or oil pipes, due account should be taken of the possibility of fires and explosions due to arcs formed either when pipes are disconnected, or when pipes make transient contact in confined places such as cellars. No definite information is at present available as to what limiting currents on such pipes may be considered safe, but it is generally recognized that the presence of currents on gas and oil lines is more objectionable than in the case of other pipes. C. INTERPRETATION OF MEASUREMENTS OF CURRENT The only accurate criterion of electrolysis damage is the intensity of current flow to earth at any point on the pipe or cable. If an accurate measure of this current flow from the pipe at any point could be made, it would come nearer giving a true indication of electrolysis conditions than any other measurement. At the present time there is no practical means available for making such measurements. The development of a simple, inexpensive and accurate means for measuring such currents locally, constitutes one of the chief needs in the field of electrolysis testing at the present time. D. USE OF REDUCTION FACTORS In many cases it is not practicable to take readings of current and potential at any point over a sufficiently long time to get all day average values of the readings at that point. Such readings should always be taken for as long a time as circumstances permit, but in making electrolysis surveys, it is usually necessary to take a large number of readings scattered over a wide area so that some of the readings can be continued only for a comparatively short time. Such short-time readings cannot, in general, be used directly as a basis for determining electrolysis conditions and in order to interpret properly the results of the survey, the readings must be reduced to some common basis, as, for example, either the twentyfour-hour average, the operating-day average, or the average for the hour of the peak load. Each of these bases has certain advantages and disadvantages depending partly on the individual conditions, and the method of procedure will often differ, depending on the method to be followed in interpreting the results. All are affected by such factors as rush or light days, unusual weather conditions, electric heaters in cold weather, morning and evening peak loads, and other causes, and these factors must be considered. The great unreliability of short-time readings for determining electrolysis conditions is especially noticeable when comparing the load curve of a line having a 5, 10 or 15 minate schedule with that of hourly interurban service, or when comparing that of a station having a 45 per cent. load factor with one having a load factor of 10 per cent. Because of this great variation and uncertainity in short time measurements and for the purposes of interpretation and comparison, it is desirable that long time readings be obtained, but if this is impossible, all short-time readings should be reduced to values for some representative period, preferably the twenty-four-hour average. Experience shows that in the majority of cases, short-time readings of from 15 minutes to an hour, taken on a city network between the hours of 10 a. m. and about 4 p. m., approach rather closely the twentyfour-hour average values, and it is found permissible to neglect the use of reduction factors in connection with readings taken during this period of the day. When, however, readings are taken at any time during the morning or evening peak, or after nine or ten o'clock at night, it is necessary to use a proper reduction factor if anything like reliable conclusions are to be reached. In general, it seems preferable to reduce such readings to the all-day average basis, rather than to the operating-day average since the operating day varies in length in different cities. E. EFFECT OF REVERSALS OF POLARITY Throughout a large portion of the territory served by a grounded railway system, it will be found that the potential differences between pipes and earth frequently reverse in direction, the pipes becoming alter |