XIII. A Direct-reading Conductivity-Bridge. WHEN it is required to check the conductivity of several tons of copper wire, of the same nominal diameter, the method employed generally resolves itself into making a careful test of the conductivity of the wire of a selected hank, and afterwards balancing a length of this on a slide-bridge against successive equal lengths of the remaining hanks of wire of that diameter. If an ordinary slide-bridge is employed, the scale-readings corresponding to the balanceposition of the sliding contact are not directly proportional to the successive conductivities. To arrive at the conductivity a certain amount of arithmetic is necessary. If, however, the slide-bridge is arranged somewhat differently, the readings of the slider can be made to be directly proportional to conductivities. Let a be the resistance of the selected wire of known conductivity k, and let b be the resistance of any of the remaining wires whose conductivity k is to be determined. The third arm of the bridge is a fixed resistance d, and the fourth arm is the slide-wire c, the resistance of which can be adjusted to balance d by moving the slider. * Communicated by the Physical Society: read 8th March, 1901. Or, replacing the reciprocals of b and a by their conductivities dks=ck ka d k = hac. Hence the required conductivity k is given in terms of the slide-wire reading, multiplied by a constant kald. It is necessary that the contact between c and d should be invariable. The writer finds that the best contact is attained by a device resembling a drill-bow: the slider carries a small brass bow strung with one or more platinum wires, each of which passes once round the slide-wire. Vertical stops fixed to the slider prevent the platinum wires from lagging behind the bow as it is moved to right or left. To graduate the slide-wire scale so that it shall read directly in conductivities, two values of b must be determined by any of the ordinary methods; these wires can then be put into position successively in the conductivity-bridge, and their balance positions marked on the scale. The whole length of the scale can then be divided in corresponding proportion, in equal divisions. The length of the equal divisions into which the scale is divided depends upon the diameter of the wire under test. It is convenient to be able to vary the length of these equal divisions, and also to be able to move the zero to right or left along the slide-wire. This can best be done by a parallel-ruler arrangement, the straight edges being in a vertical plane through the slide-wire. If the top straightedge is then divided into equal parts, plumb-lines from those divisions divide the slide-wire into equal parts of any required length; and the parallel-ruler can be moved, as a whole, along the wire. The angle at which the straight-edge is set depends. upon the diameter of the wire under test, the length being constant; the apparatus may therefore be calibrated in diameters as well as in conductivities. XIV. On the Magnetic Effect of Electrical Convection. FROM Historical Review. ROM a series of experiments extending over the last four years, M. V. Crémieu † has come to the conclusion that a moving electrified body produces no magnetic effect. * Communicated by Prof. J. S. Ames. C. R. t. cxxx. p. 1544 (1900); t. cxxxi. p. 578 & p. 797 (1900); t. cxxxii. p. 327 & p. 1108 (1901). The conclusion is in direct opposition to the classical experiment performed by Professor Rowland* in Berlin, in 1876, which was subsequently repeated in this laboratory by Professor Rowland and C. T. Hutchinson + in 1889, and confirmed also by experiments performed by Röntgen ‡ and Himstedt § in Germany. A further investigation of this very important question was therefore considered desirable. Accordingly in the fall of 1900, under the supervision of Professor Rowland, the following research was undertaken. As a result, it has been shown conclusively that electrical convection does produce magnetic action. The idea that a moving electrified body might produce a magnetic effect similar to that produced by an electric current first occurred to Faraday || in 1837. The first to perform an actual experiment on the subject, however, was Professor Rowland. The experiment was carried out in Germany in 1876, though the idea of it had occurred to him as early as 1868, and is recorded in a note-book of that date. His method was similar to that described by Maxwell in his Treatise on Electricity and Magnetism (§ 770), written presumably in 1872 or 1873. In the Berlin experiment, Professor Rowland used a horizontal hard-rubber disk (diameter 21 cm.), coated on both sides with gold, revolving between two glass plates with their inner surfaces also gilded. Each side of the disk formed a condenser with the gilded surface opposite it. The condensing plates were earthed; and the disk charged by means of a point brought up within mm. from the periphery. An astatic needle was suspended above the upper condensing plate, so that its lower magnet was nearly over the edge of the disk, and perpendicular to a radius of the same. The needle was inclosed in a metal case, so as to screen it from electrostatic action. The speed of the disk was 61 revolutions per second. On reversing the sign of electrification of the disk, a deflexion of from 5 to 7.5 mm. was obtained, depending on the conditions. The deflexion was the same whether the gilded surfaces of the disk and condensing plates were divided into sectors, or left continuous. The deflexion as observed and calculated agreed quite well, but as the needle was always very unsteady, it was impossible to make the readings with any great accuracy. In 1883 Pogg. Ann. clviii. p. 487; Am. Journ. Sci. 1878, p. 30. + Phil. Mag. xxvii. p. 445 (1889). Lecher repeated Rowland's experiment, but obtained negative results. Only a short account of this investigation was published, no details of the apparatus being given. It is impossible, therefore, to say why he obtained no effect. It may have been because his apparatus was not sufficiently sensitive. In 1885, in an attempt to detect the existence of displacement currents, Röntgen † had occasion to use an apparatus similar to that employed by Professor Rowland in his Berlin experiment. The ebonite disk in this case was not covered with gold. Incidentally, the effect on the suspended needle of charging the disk, when rapidly rotating, by a brush discharge from a series of points, was tried. The Rowland effect was readily observed. On reversing the sign of electrification, a deflexion of from 8 to 10 mm. was noted. In 1888 this apparatus was again used to investigate the action of a dielectric moving in a homogeneous electrostatic field. If the upper condenser-plate is charged negatively (say) and the lower positively, then there will be an apparent positive charge on the upper surface of the disk, and an apparent negative charge on the lower surface. If then the motion of an apparent charge due to the polarization of the dielectric produces magnetic action, the needle, which was nearer the upper than the lower surface, ought to show a deflexion on rotating the disk and charging the condenser formed by the two stationary plates. After all precautions had been taken to prevent the disk from assuming a real charge, a deflexion of from 2 to 3 mm. was observed on reversing the sign of electrification of the condenser-plates §. On increasing the difference of potential between the condenser-plates until a brush discharge took place between the upper plate and the disk (the upper plate was 14, the lower 25 cm. from the disk), the Rowland effect was again readily observed. In 1889 Professor Rowland and C. T. Hutchinson || repeated in this laboratory the original Berlin experiment, employing, however, a different form of apparatus. Instead Rep. d. Phys. xx. p. 151 (1884). t Sitzb. d. Berl. Akad. 1885, p. 95. Ibid. 1888, p. 23; Wied. Ann. xl. p. 93. § Röntgen gives no calculation of the effect which should be expected. On the assumption that the magnetic force due to any element of surface is proportional to the quantity of apparent charge passing a given point in unit time, I have calculated from the data given in Röntgen's paper that the deflexion should be 2.1 mm. I employed the formulæ used by Rowland in his Berlin experiment (loc. cit.). Phil. Mag. [5] xxvii. p. 445. of a single disk rotating in a horizontal plane, they used two vertical disks rotating about horizontal axes in the same line; the needle system, inclosed in a brass tube, was placed between the disks, opposite their centres. Each disk was surrounded by a guard-ring, both disks and guard-rings being gilded on the side facing the needle. Between the disks were placed two condensing-plates-glass plates gilded on the surfaces facing the disks. Between the condensing-plates was suspended the needle. This arrangement permitted of accurate calculation of the effect which should be expected. The condensing-plates were charged, from a Holtz machine and battery of leyden-jars, to a potential of about 5000 volts; the disks were earthed by means of metal brushes bearing on studs fixed in the periphery of the disks. The speed of the disks was about 125 revolutions per second. The deflexion of the needle on reversing the electrification was from 5 to 8 mm. This experiment offers a means of determining the ratio of the two systems of electrical units. The mean of all their determinations gave for this 3-19 x 101o; the determinations which differed from this the most being 3.74 x 101o and 2.26 x 101o. During the same winter of 1889 Himstedt * carried out a series of experiments on this subject in Giessen. In his first experiments he employed the same apparatus as had been previously used by Röntgen, with some modification in details. Later, an apparatus was constructed similar to that used by Rowland and Hutchinson, but considerably more sensitive. Two disks of ground-glass, 20 cms. in diameter, were mounted so as to revolve about horizontal axes in the same line. A strip on the edge of both sides of the disks was made conducting by rubbing into the ground surface a thin coat of graphite. On both sides of each disk were placed condensing-plates, and between the two inside plates was suspended an astatic needle, properly shielded, the upper needle being just above, the lower just below the conducting strip. The disks were charged by an induction-machine and a battery of leyden-jars by means of a sliding contact. With a speed of 117 revolutions per second and disks charged to 5000 volts. (condenser-plates earthed), a deflexion of 100 mm. on a scale 3 m. distant was obtained on reversing the sign of electrification of the disks. This apparatus did not permit of ready calculation of the deflexion to be expected, so no absolute measurements were made. Himstedt, however, showed that the deflexion was in the direction to be expected, reversed with the direction of rotation of the disks, was proportional *Wied. Ann. xxxviii. p. 560 (1889). |