The first galvanometer experimented upon was a fine Thomson instrument, whose coils were wound by While, and whose needle system was made by Very*. The magnetic system was built up of hollow cylindrical magnets, rolled up out of the thinnest sheet-steel, heated red-hot in fused ferrocyanide of potassium, and hardened in mercury in a powerful magnetic field. The system was then astaticized by the usual method of stroking with a weak bar-magnet. Every care was taken to secure the maximum intensity of magnetization. The constant of the galvanometer in its most sensitive state was C=1.5 × 10−9†. The best value of the constant of the Allegheny galvanometer already referred to was C=1.3 x 10-9. On October 23 the magnetic system was removed and remagnetized, and reastaticized by the method described above. No other change was made. The new constant after magnetization was C=8× 10-10, an increase in delicacy of nearly 100 per cent. The second galvanometer to whose magnetic system this method of treatment was applied was one constructed by Queen and Co. especially for the Observatory. The magnetic system had very nearly the same dimensions as the one already described, but the individual magnets were solid. The method of hardening and magnetizing was not described by the makers. The constant for this galvanometer as received from them was: C=1·6× 10−9 for coils in series, R=309 ohms, C=5·5 × 10-9 for coils in parallel, R=19′2 ohms. The needle was removed and remagnetized by the new method and the constant redetermined. The new constant was: C=1x 10-10 for coils in series, C=34× 10-11 for coils in parallel. Part of the immense improvement here is due to a more accurate centering of the coils, to the magnetic system, and to a reduction of the excessive damping; yet with a most *This instrument was a duplicate of the celebrated Thomson galvanometer, with Very hollow magnets, used by Langley at Allegheny in his bolometric work; and which was considered at that time to be the most sensitive (for its resistance) in existence. See paper by S. P. Langley, "On hitherto unrecognized Wave-lengths," Phil. Mag. [5] xxii. p. 149 (1886). +C current in amperes which produces a deflexion of 1 millim. on a scale at the distance of 1 metre, when the time of a single swing is ten seconds. liberal allowance for the influence of these changes, the sensitiveness was increased fully ten times by the remagnetization. This would seem to indicate either that the first magnetization had been very inefficient, or that the needle in transit from Philadelphia to Washington had been accidentally subjected to some very strong demagnetizing influence. It serves to show, however, the importance and necessity of some such ready and efficient means of remagnetizing the systems of galvanometers in which a high degree of sensitiveness and not absolute constancy is required. Steel magnetized to the degree of intensity here attained will of course gradually lose a part of its magnetism, but not more (in my experience) than 10 or 20 per cent. in many months, if carefully handled. This small loss is not of importance compared with the gain in sensitiveness secured, as the original strength may at any time be quickly and easily restored or even slightly increased by remagnetization. The advantage of the method which specially commends it to general laboratory use is the simplicity of the apparatus required. The whole arrangement may be made in any laboratory in a single afternoon. The same electromagnets may be used for systems of varying dimensions by adding adjustable pole-pieces. But if many systems are to be treated, a more convenient although more elaborate arrangement, like that shown in Plate XIII. figs. 3 & 4, will be desirable. In this the magnetic system is held lightly between two long jaws copper a, a, adjustable in width by means of the screw b. An adjustable fork or table c, which may be replaced by a clamp if desired, serves to carry the fibre support. of The two halves of the electromagnet, which may be wound as before, or with four coils as here shown, are carried on arms d, e, pivoted at f, so that they may be easily swung apart or brought together, the motion being made symmetrical with respect to the jaws a, a by means of the links and sliding block. The lower set of poles are adjustable on the yokes for systems of different lengths, and the whole is mounted on an L-shaped base, which may be placed so that the needle is either vertical as shown in the figure, or horizontal. This method has also been applied with much success to the initial magnetization of some new systems for the same galvanometer already described, and for the new very sensitive one which is described in a subsequent paper. A further *Up to a certain point repeated magnetization increases the permanent magnetism. See experiments of Frankenheim, Pogg. Ann. cxxiii. (1864); and Fromme, Pogg. Ann. vii. (1875), Wied. Ann. iv. (1878). advantage which this method possesses over the usual one in the case of a new system is that, in making it up, we have only unmagnetized needles to handle; an advantage which those who have had much to do with this kind of work can readily appreciate. Astro-Physical Observatory, Washington, D.C., February 1893. LVIII. On the Effects of Magnetic Fields on the Electric Conductivity of Bismuth. By JAMES B. HENDERSON, B.Sc.* [Plates XV. & XVI.] T HAT magnetic fields have an effect on the electric conductivity of metals was noticed first in the year 1856 by William Thomson† (Lord Kelvin), who was led to suspect it from the effects which he had discovered magnetization to have on the thermoelectric properties of metals. He experimented on iron and nickel, and found in both an increase of resistance along the lines of force, and a diminution perpendicular to them. Bismuth was first experimented on by Tomlinson‡, who found an increase of resistance due to longitudinal magnetization of Bi wire, and he found a similar increase in Fe, Ni, Co, and steel wires. Later investigations in this subject have been made by Righi§, Hurion, Leduc ¶, Ettingshausen and Nernst**, Ettingshausen ††, Goldhammer‡‡, Lennard and Howard §§, and Lennard ; but the last of these is the one which has the most important bearing on the present investigation. In this Lennard used spirals of Bi wire, in all fifteen spirals being tested, whose wires varied from 0.2 to 0.4 millim. in diameter, and from 50 to 150 centim. in length, the respective resistances varying from 6 to 25 Siemens units. The resistance was determined by the Wheatstonebridge direct-current method, and also by the method using *Communicated by the Author. † Math. and Phys. Papers, ii. p. 307. $ Journ. de Physik. iii. p. 355 (1884). ‡‡ Wied. Ann. xxxi. p. 360 (1887); xxxvi. p. 804. alternating current and telephone, and a remarkable difference was found in the resistances obtained by the different methods. Two kinds of Bi were used, one chemically pure and the other containing traces of Fe and Zn, but differences were not more than those due to observational errors. Deviations from the mean values amounted to 1 per cent. The temperature varied from 10° to 25° C., and once was 0° C. The present investigation was instituted to determine definitely the relation between the magnetic field and the resistance of Bi wire going to much higher field-intensities than had ever been experimented with, and also to determine the influence of temperature on that relation if any was found. The investigation was started purely from the scientific standpoint, but the importance of it to the practical application of Bi wire as a field-tester was not lost sight of. Owing to the purity with which Bi wire is now prepared for instruments for magnetic-field testing, and the convenient form for experimenting which these field-testers offer, it was determined to use them in this investigation. Two such were employed, the spiral of one having a diameter of about 18 millim. and a resistance of 24 ohms, and that of the other a diameter of 6 millim. and a resistance of about 9 ohms. The form of the instrument is shown in fig. 1. Fig. 1. The magnetic fields were obtained by means of a Ruhmkorff electromagnet, and the very highest field-intensities from the large ring-electromagnet lately designed by H. du Bois * for the production of very strong fields for experimental purposes. With the large spiral the ordinary domeshaped pole-pieces belonging to the Ruhmkorff magnet were used, the holes in them being first blocked up with pieces of soft iron to render the field as uniform as possible. For the small spiral special pole-pieces were prepared, which were designed to fit the ring-electromagnet, but by using a pair of flat poles with them they also fitted the Ruhmkorff magnet. They had an angle of 60°, and their faces, which were 7 millim. in diameter, were held at a distance of 1.5 millim. apart by means of a brass casting to which both pole-pieces were rigidly attached. This casting consisted of two thick rings held rigidly parallel and coaxial by means of two stout * Magnetische Kreise, p. 277; Wied. Ann. li. (1894); Phil. Mag. May 1894. Phil. Mag. S. 5. Vol. 38. No. 234. Nov. 1894. 2 L distance-pieces, and to each ring one of the pole-pieces was soldered, the inner surfaces of the rings being conical to receive them (see fig. 4). The distance-pieces were not at opposite ends of a diameter of the rings, but one was displaced a few degrees round the circumference of the ring to allow of the introduction of the spiral and the ballistic coil, which turned about the same horizontal axis and in the same vertical plane (see fig. 2). Fig. 2. The ballistic method was used to measure the field, the galvanometer being of the form designed by du Bois and Rubens*, used with the four 20-ohm coils in parallel. It was standardized by means of the induced current produced in a fine coil placed at the centre of a long straight solenoid of thick wire, when a current through the latter was made or brokent. The constants of both coils being known and the current in the solenoid measured, the induced current *Wied. Ann. xlviii. p. 234 (1893). † For particulars of apparatus see Lehmann, Wied. Ann. xlviii. (1893). |