two portions of the pipe network are separated, and also due to arcing between adjacent pipes in confined air spaces such as cellars, where there may be considerable potential differences due to such currents. In some cases, also, excessive heating has resulted, due to the presence of abnormally large currents on small pipes, and the presence of such heavy currents may make it very difficult to prevent local interchange of current between neighboring structures. Heavy currents on lead power cables are also objectionable because the heat generated in the lead sheath may limit considerably the carrying capacity of the conductors within the sheath. In view of these factors, it becomes important to measure currents on pipes and cables in many instances. Relative current measurements on pipe and cable systems made before and after the application of mitigative measures are also valuable as an index of the effectiveness of the mitigative system employed. This is true, however, only when there has been no installation of new drainage connections or changes in existing drainage connections on the affected structures.

(b) Selection of Points of Measurement. In general in selecting points for making current measurements, it is desirable to secure some points at which maximum current flow may be anticipated, and also a considerable number of points that may be regarded as representative of conditions generally. As a rule, the maximum current in an undrained pipe network may be expected in pipes extending approximately parallel to the tracks and near the neutral or slightly positive areas. Also numerous cases will usually be found in any network in which one or, at most, a few mains serve as connecting links between local networks, and such mains usually will be found to carry much larger currents than mains forming a portion of the network. On drained pipe systems, the maximum currents will as a rule be found in the pipes extending in all directions from the points at which drainage cables are connected. It is impossible to lay down rules more detailed than the above for the selection of points at which measurements should be made. Experienced judgment should be followed in all cases.

(c) Methods of Measuring Current Flow in Pipes. Four general classes of methods of measuring current flow in pipes and other metallic structures have been used. The one that is perhaps the most frequently used is the ordinary drop-in-potential method in which the voltage drop on a measured length of pipe, not including a joint, is taken and the current calculated from this voltage drop and the estimated resistance of the portion of the pipe across which the potential drop is measured. Complete tables for the resistance per unit length of the various sizes and kinds of pipe in common use are given in the appendix. Careful tests made on a great variety of specimens of pipe of different kinds indicate that measurements of this kind can be depended upon to give results accurate to within about 10 per cent. which is ample in most cases encountered in practice.

A second method, used in special cases where greater accuracy than is possible by the drop-in-potential method is necessary, is the method for calibrating the pipe either by sending a known current through it super

posed on the railway current already flowing in the pipe, or by shunting through an ammeter, certain portions of the current actually flowing in the pipe. These methods have taken various forms, one of the most important of which is described later.

A third method consists in the use of what is known as a direct current ratio relay in a manner somewhat analogous to the use of a current transformer on alternating current circuits. This is useful only where currents of fifty amperes or more flow on the pipe.

A fourth method consists in surrounding the pipe with an iron ring containing an airgap and providing means for measuring the magnetic flux set up across the airgap by the current in the pipe. Several different methods are available for making these measurements. The last two methods may also be used for calibrating the pipes, thus eliminating in some measure the uncertainty arising from the calculation of the pipe resistance. It is questionable, however, whether in most cases the greater accuracy thus achieved is sufficient to warrant the use of the more complicated methods. There are given below somewhat detailed descriptions of the first two of these methods.

Drop-in-Potential Method. This method consists in connecting potential terminals to a section of pipe a few feet apart and measuring the millivolt drop, and in calculating the current from this millivolt drop and the resistance of the section under test. It is very widely used, and its great simplicity adapts it to work of this kind. This method has the great advantage that it can not only be used with an indicating instrument, but also with a recording instrument unless the currents are very small, and thus not only a permanent graphic record be obtained, but also the average value for a given period can be determined. The tables appended to this report are based on careful measurements made by the Bureau of Standards on several hundred specimens of iron and lead pipes from various sources, and they are accurate enough for all practical purposes.

In using this method it is necessary to make an excavation at the point where the measurement is to be taken and attach two leads to the pipe, preferably as far apart as practicable without including a joint. This connection may be made in numerous ways, but perhaps the best way is to insert at each point a corporation cock in which a rubber-covered wire has been soldered. If the connections are to be permanent, the leads should be brought underground to a point inside the curb and there terminated in an ordinary service box or other suitable receptacle so that they will be protected from traffic but readily accessible for repeating the measurement at any time. One method of making such connections and protecting the leads is shown in Fig. 21, and another which has been very successfully used in paved streets is shown in Fig. 22. It is also important that the junction between wire and corporation cock be protected by painting with a heavy asphalt or similar paint. If the current on the pipe is large enough to be of practical significance it can be read with an ordinary sensitive millivoltmeter either indicating or recording. In special cases where the current is extremely small, only a high sensitivity indicating millivoltmeter or even a portable galvanometer can be used.

Calibration of Pipes. One of the methods most commonly used for the calibration of pipes involves superposing a current on that already in the pipe and measuring the change in millivolt drop due to this superposed current. This method, originally used by Professor B. F. Thomas, was first described by Dr. Carl Hering in the Transactions of the American Institute of Electrical Engineers for June, 1912. Theoretically, this method should give very high accuracy, but it should be borne in mind that the resistance of the pipe thus determined is correct only for the conditions under which the measurement was made. Iron pipes, especially wrought iron and steel pipes, have a high temperature coefficient of resistance and variations in this resistance due to temperature changes between winter and summer may introduce variations of five per cent or more in this resistance. For this reason it is very doubtful whether the complication involved in the use of this method is justified, but it has been used by some engineers. Further, owing to the presence of rapidly fluctuating railway currents on the pipes, the application of this method is often difficult.

Use of a Direct-Current Ratio Relay. An instrument known as the direct-current ratio relay for measuring current in conductors which cannot be opened for the insertion of ammeters or shunts has recently been devised. The ratio relay permits the measurement of variable unidirectional currents of relatively large magnitude only, on an ordinary direct current ammeter. This instrument gives very good results when very large currents are being measured, but in its present form it is not suitable for measuring currents of a few amperes, such as are most frequently encountered in electrolysis testing.

4. Comparing Currents Under Different Conditions.

In case the object in view is the determination of relative current in pipes under different systems of mitigation, this can be done simply by measuring potential drops between service or between adjacent fire hydrants. In general, the resistance may be regarded as sufficiently constant so that the currents under the two conditions of test will be proportional to the voltages at corresponding test stations.

5. Measurement of Current Flowing from Underground Structures to Earth.

It is extremely desirable to measure the amount of current flowing from a particular portion of a pipe or cable network directly into the earth. In fact, if such measurements could be made conveniently and with sufficient accuracy they would be by far the most important and valuable measurements that could be made in an electrolysis survey, since this measurement would afford the most accurate measure of the rate at which damage is progressing. Unfortunately, there has not been available up to the present time any very satisfactory method of measuring such current flow except in very special cases. Four different methods have been proposed under special conditions for making this measurement. These are: (a) differential current measurements; (b) the use of a Haber earth