EARTH CURRENTS. Our knowledge of earth currents dates from the introduction of telegraphy. Ordinary telegraph lines are traversed by natural electric currents which occasionally interfere seriously with their use. It is these currents which we shall mainly consider here. There are, however, electric currents arising from electro-chemical action in mineral deposits, electric waves due to natural causes such as thunderstorms or to wire less, and widely distributed artificial currents due to electric trac tion and lighting. The pioneers of our subject in England included W. H. Barlow and C. V. Walker, in charge, respectively, of the Midland and South-eastern telegraph systems. Barlow noticed the existence of a more or less regular diurnal variation, and the fact that earth currents proper occur in a line only when both ends are earthed. Walker collected statistics of large earth cur rents. His and Barlow's results indicated that the lines which suffer most from earth currents in England have the general direc tion north-east to south-west.
As Walker pointed out, it is the direction of the terminal plates relative to one another that is the essential thing. Thus in France, in 1883-84, E. E. Blavier, comparing continuous records from two lines, one aerial, the other underground, connecting Paris and Nancy by different routes, found an exact parallelism between the currents. The difference of electromotive force between the two terminal plates is the cause of the current, but the intensity varies inversely as the total resistance of the wire. The currents in the ground must depend on the resistance of the soil, which will vary with the depth and the geological conditions. Observations on experimental lines can convey only a general idea as to the cur rents in the earth itself, but they may be expected to show their variations.
Our present knowledge owes much to practical men, but the fact that telegraph systems are primarily commercial undertakings is a serious obstacle to their employment for research. Some valuable observations have, however, been made on long telegraph lines. In addition to the French observations already referred to, mention may be made of important observations made from 1 884 to 1888 in the case of two German lines, Berlin to Thorn and Berlin to Dresden, discussed by B. Weinstein, also of observa tions during 1921 and subsequent years on several Swedish lines, discussed by D. Stenquist.
Each galvanometer had a resistance of about 200 ohms, shunted by a resistance of only 2 ohms. The total effective resistances in the north-south and east-west lines were 225 and 348 ohms respec tively. What is really measured by such apparatus is the voltage between the two earth plates. This voltage between the two earth plates as measured is usually stated in mv./km.: millivolts per kilometre.
Industrial currents put an end to the usefulness of the earlier installations. Of those still functioning the best known is that of the Ebro observatory, Tortosa, dating from 191o. One line, which we may call the N.'-S.' line, has earth plates r • 28 km. apart, and the straight line joining them is inclined at 25 ° 16' to the geo graphical meridian, thus running roughly from north-north-west to south-south-east. The other, or W.'-E.' line, has plates 1.415 km. apart, and is nearly perpendicular to the N.'-S.' line, being inclined at 24°46' to the geographical west-east line. The method of recording is similar to that described at Parc St. Maur. Other more recent installations are those of the Carnegie institution at Watheroo (Western Australia) and Huancayo (Peru). The two lines at Watheroo run exactly north-south and east-west, each being about a mile long.
A source of uncertainty, especially in short lines, is the probable existence of polarization currents from the electrodes. These can be largely, if not entirely, avoided in temporary installations by the use of special non-polarizable electrodes. But difficulties still remain in permanent installations. For many purposes a constant polarization current is immaterial, but varying polarization effects might cause serious confusion. Provision for the detection of spurious electrode effects is thus desirable, and is included in the installation at the Carnegie institution which is situated at Watheroo (Western Australia).
Diurnal Variation.—The diurnal variation data in Table I. are all given as inequalities, i.e., as algebraic departures from the mean value for the day. The inequality may be regarded as super posed on a "constant" value, which, in general, was so large that the direction of the total current was invariable. In cases like the Ebro, where the absolute difference of potential between the earth plates was known, the entries in Table I. represent milli volts per kilometre. In cases where the unit was arbitrary, the units employed in Table I. were such as to make the A.D., i.e., the average numerical departure of the hourly values from the mean value for the day, equal r oo. A few explanatory remarks will be found useful. The Greenwich data are based on a discus sion by Sir G. B. Airy of results from 1865 to 1867. The Green wich observations were among the earliest experimental studies made of the subject. Airy resolved the observed currents from the two lines in two directions, one N.'-S.' in the magnetic meridian, the other W.'-E.' perpendicular to it. Magnetic declination at the time was about 2 r ° W. In every month the mean of the 24 hourly values represented a current from north to south in the magnetic meridian, and from east to west in the perpendicular direction. Col. ro represents the inequality in the former current, + signify ing increase in the current from north to south. Col. 13 rep resents the inequality in the latter current, but + signifies algebraic increase in the direction west to east, and so a numerical fall.
Col. 14 is based on a discussion by H. N. Dickson of hourly readings during a fortnight in Sept.-Oct. 1885 of the current in the telegraph cable between Ft. William and the Ben Nevis observa tory. The "constant" part of this current was directed up the hill, nearly west to east. Thus in this case + represents numerical as well as algebraic increase.
Cols. 9 and r 2 are based on Weinstein's discussion of the re sults from the German lines, Berlin to Thorn, 262 km. away, and Berlin to Dresden, r 20 km. away. Relative to Berlin the geographical co-ordinates of the two other terminals were Thorn o° 29' N. lat., 5° r 2' E. long., and Dresden r ° 28' S. lat., o° 2 E. long. The directions taken as -}- in Table I. are in each case the opposite of Weinstein's. His units being arbitrary, such (different) units were employed in cols. 9 and 12 as to make the A.D. in each case r oo. But if we used the same absolute unit for the two direc tions, we should reduce the figures in col. r 2 in the proportion roughly of 45 to roo.
Cols. r and 5, which refer to true geographical directions, are derived from the official Ebro publication for the year 1926, and are based on five days a month, mainly the international mag netically quiet (or Q) days selected at De Bilt ; when the Ebro record was incomplete another quiet day was substituted. Cols. 2 and 6 also refer to the Ebro; they were calculated by D. Sten quist from all the available days, quiet or disturbed, of the six years 191o, 1912, 1913, 1916, 1917 and 1918. Cols. 8 and II, applying to the Ebro, reproduce the results in cols. I and 5 respectively, multiplied by the factors necessary to make the A.D.=1oo in each case. In addition to the inequalities in cols. I and 5, obtained by combining results from the Ebro N.'–S.' and W.'–E.' lines, we have in cols. 3, 4 and 7 inequalities for the two lines separately, going only to the nearest mv./km. in the N.'–S.' line. Cols. 3 and 7 are based like cols. 1 and 5 on the quiet day results of 1926 in the official publication. Col. 4 refers also to 1926, but is derived from the international magnetic disturbed (or D) days. Of the 6o disturbed days eight were omitted, as the trace was incomplete. No substitutes were used, and equal weight was given to each month.
The agreement between the Ebro results in cols. 1 and 2, and again in cols. 5 and 6, is good. The closeness of the amplitudes in the two cases must, however, be regarded as largely accidental. The amplitude really varies with the amount of disturbance and with the sunspot frequency. The six years on which cols. • 2 and 6 depend had a mean sunspot frequency of only 44. 2, as against 63.9 for 1926. The effect of disturbance on the regular diurnal variation at the Ebro can best be inferred from the corresponding quiet and disturbed day results in cols. 3 and 4 of Table I. Dis turbance does not seem to have much influence on the type of the diurnal variation, but it largely increases the amplitude.
In view of the irregularities natural in inequalities from a single year, the A.D. is probably a better measure of the amplitude than is the range in most cases in Table I. If we take the A.D. as our measure, the ratio borne by the amplitude of the diurnal inequal ity in the west-east line to that in the north-south line is 0.43 at the Ebro (1926) and 0.45 at Berlin. According to observations by D. Stenquist, extending over several months of 1924 and 1925, the ratio was 0.54 at Lund (56° N.) and 1.84 at Lulea (66° N.). This tendency to a rise in the relative importance of the diurnal changes in the west-east direction as we go to higher latitudes in Europe was pointed out by Stenquist.
The fact that the amplitude of the regular diurnal inequality increases with sunspot frequency was first demonstrated by Wein stein for the German lines. Stenquist found a similarly large effect at the Ebro. L. A. Bauer also dealt with the matter in a discussion of Ebro data from 1910 to 192o. Employing all days with the exception of the highly disturbed, he found that the ab solute daily range, i.e., the excess of the absolute maximum over the absolute minimum of the day, in the N.'–S.' line was 53% larger in 1917, the year of sunspot maximum, than in 1913, the year of minimum.
The data in cols. II, 12 and 13 are expressed in terms of units such that the mean from the 12 months is in each case I oo. Col. II refers to Weinstein's value of the resultant current as derived from the diurnal inequalities in the German lines. Cols. 12 and 13 are derived from the all day absolute ranges in the Ebro N.'–S.' line. Col. 12 is based on data given by Bauer for the II years 1910 to 1920. Col. 13 merely repeats the data in col. 3, but in a shape comparable with col. 12. Col. 14 is based on a table given by Stenquist of the number of disturbances from 188 i to 1884 in telegraph lines at four Norwegian stations, varying from 6o° N. lat. to 70° N. The occurrences at the four stations, summed up as if independent, numbered 904. Each monthly sum was multiplied by to, so as to bring the monthly mean up to too.
Cols. I I and 14 may be regarded as representing respectively the regular and irregular daily changes. In col. 11 the maximum of activity occurs in the summer months, the minimum in the winter.
In col. 14, on the other hand, the maximum activity is in the equinoctial months, and the more conspicuous minimum occurs in summer. Cols. 12 and 13 agree with col. 14 in placing the maximum of activity in the equinoctial season. Cols. 3, 4, 8 and 9 agree with cols. 12 to 14 in making equinox the season of great est activity. Cols. 1, 2, 6 and 7 agree with col. 11 in showing a tendency for the regular diurnal changes to be larger in summer than in winter.
Cols. 5 and Io of Table II. represent the mean monthly values of the "constant" part of the current in the two Ebro lines. The difference between the results in the N.'–S.' line from January to June and from August to November is extraordinary. A rapid fall occurred in July and a rapid rise in December. The phenom enon is almost equally prominent in the results given by Bauer for the resultant currents at the Ebro from 1914 to 1918, the chief difference being that in Bauer's average year the decline did not set in until August. His July values varied between 329 and 61i, his September values between 13 and 33. The phenomenon is probably of local origin and might be a consequence of variable plate polarization. But there is nothing in the run of the figures in cols. I to 4 and 6 to 9 which suggests that the diurnal changes are affected.
Close parallelism with magnetic phenomena is also shown in an investigation by W. J. Peters and C. C. Ennis of the 27 day recurrence interval. This is a well known magnetic phenomenon, conditions, whether specially disturbed or specially quiet, tending to recur after 27 days. Peters and Ennis employed the absolute daily ranges in the Ebro N.'–S.' line from 191 o to 1924. Primary pulses were formed from selected disturbed and quiet days— usually five of each per month—and the adjacent days, in a similar way to that employed by Chree for terrestrial magnetism, and exactly as in the case of magnetism these were found to have associated pulses of disturbance or quietness, with their crests falling 27 days subsequent to those of the primary pulses.
All the phenomena mentioned here point to an intimate con nection between earth currents and terrestrial magnetism. Many authorities have supposed it a case of cause and effect, differing, however, as to which is the cause and which the effect. So far as the limited data available enable us to judge, neither hypothesis suffices to explain the facts. It is practically certain that at least a substantial part of the regular diurnal changes in terrestrial mag netism is due to overhead electric currents, and the association of earth current with magnetic disturbances, and of both with aurora, points to the upper atmosphere as the ultimate seat of at least disturbance phenomena. The 27 day recurrence phenomenon sup ports the view, now generally held, that aurora and magnetic dis turbance are due to electrical discharges from the sun. The exact nature of the discharge is still, however, a matter of speculation.
If the main cause of the daily magnetic variations is situated above the earth, this varying external field must necessarily induce electric currents within the earth. From consideration of the rec ords of the diurnal variation of magnetic force at numerous obser vatories, Schuster showed that about two-thirds of the diurnally varying field is of external origin, the remainder coming from within. This implies that the induced currents in the earth are less than the primary currents in the atmosphere. According to Chapman the earth must be much more conducting below a depth of about 25o km. than in the layer above this. The specific resist ance found for the core is 2,700 ohms per cm. cube. The core is about i,000 times as conducting as dry earth, but only about o o as conducting as the salt water of the oceans.
Chapman and Whitehead have compared the diurnal variation of the horizontal potential gradient at Ebro with calculations based on the computation of the magnetic field. The agreement in phase is close enough to indicate that the theory is substantially correct ; the observed variations are more than five times as great as the computed ones; but this is probably due to the local distribution of resistance in the rocks of the Ebro valley.
The movements of a magnetic needle being governed by the electric currents in the upper atmosphere as well as those under ground it is not to be supposed that magnetograms and records of the horizontal electric force will be very simply related. In general it is true that an increase in the flow of electricity under ground from north to south is accompanied by a movement of the com pass needle to the west and a flow from east to west is accom panied by an increase in horizontal force. The ratio of the scales on which the variations of electric and magnetic force occur is much greater for rapid oscillations than for the regular diurnal changes. This may be due to the fact that the earth currents taking part in the rapid oscillations do not extend to any great depth.
In the case of Ben Nevis, as already remarked, the "constant" part of the potential difference indicated a current flowing up the mountain. A similar tendency has been described in other cases, but more evidence is needed to justify the conclusion that there is a general tendency for earth currents to flow up mountains.
Certain mineral deposits, e.g., pyrites, especially when the sur rounding material is damp, give rise to local earth currents of perceptible amount. This property may be utilized by prospectors, as is described by J. Bartels in the Lehrbuch der Geopliysik, P. 575.
As a specific example, reference may be made to investigations carried out by W. J. Rooney and O. H. Gish in the immediate neighbourhood of the observatories of Watheroo and the Ebro. Suppose four terminals sunk in the earth along a straight line, the distance between adjacent pairs being in each case a. The extreme pair of terminals are in a circuit containing a source of current. If 1 is the measured current in the circuit and V the potential difference found with a potentiometer between the inner pair of terminals, the resistivity of the earth p, assumed uniform, is given by p = 2 lraV/I.
As regards the applications, in Rooney and Gish's words "The value of resistivity thus found must, however, in general be con sidered an average, in which the resistivity of the earth near the line of terminals is the more heavily weighted." They assume that material whose distance from the line of electrodes exceeds a may practically be left out of account.
The mean values calculated for p at Watheroo, from a series of sections and in different months, in ohms per varied from 58o,000 when a was 2.5 metres, to Too when a was 6o or Too metres, and then increased to 5,00o when a was 600 metres. It was inferred that the sandy soil at the surface had a very high resistivity when dry, but beneath it at a comparatively small depth was a layer of low resistivity. This was confirmed. From the increase in the values obtained for p when a exceeded ioo metres it was inferred that below Too metres there was a stratum of much higher resistivity.
At the Ebro the resistivity with a= 2.5 metre, i.e., near the surface, was much less than at Watheroo, in one case as low as 1,300; but with a as large as 6o or ioo metres it rose to io,o0o, being then and for greater values of a much larger than at Watheroo. Rooney and Gish point out that the higher resistance of the deeper strata in the Ebro valley is accompanied by higher values of the potential gradient, so that the currents in the earth in the two localities are comparable.
Disturbances Due to Artificial Earth Currents.—Prac tically every kind of artificial use of electricity, traction, lighting, telegraphy, telephony and wireless, may set up earth currents of one kind or another.
The most important case is that of the disturbance caused by the use of direct current for electric traction. The currents are large, and even if there were a perfectly insulated return there would be a considerable resultant magnetic force at distances from the track which were not largely in excess of the distance apart of the direct and return currents. At a distance of 2 m. or more from an ordinary electric railway the disturbance is usually much the largest in magnetographic records of changes in the vertical component of the earth's field. At Washington (D.C.), U.S.A., sensitive instruments have shown disturbance effects at a distance of 12 miles. Practically all the magnetic observatories once func tioning near large towns—including of late years Kew and Green wich—have been put out of action.
As regards damage from electrolysis caused by the so-called "vagabond" currents to underground gas and water pipes, nu merous observations have been made, especially in Germany and the United States. Owing, perhaps, to the conflicting interests involved, the extent of the damage has been variously estimated. However this may be, the prosecution of research into natural earth currents has difficulties to contend with which did not exist years ago.