were very small, and those for the later epoch were of opposite signs, according as they were calculated from the earlier or the later survey. The same calculation was made for two other circuits, one (y) in Great Britain and one (8) in Ireland; but instead of using the calculated terrestrial lines, the true values of the forces and declinations were employed, as deduced from the nearest stations. The great local disturbances in Antrim interfered with the value of the Irish circuit. The following table gives the results in amperes per square kilometre obtained from circuits a and ẞ by the terrestrial lines for 1886.0, and also by the mean terrestrial lines for 18910, which occupy the mean position between those given by the two independent surveys, and lastly the results for circuit (y). From these we may conclude that there is not in the United Kingdom a vertical current amounting on the average to 0.1 ampere per square kilometre. There may possibly be a current of about a tenth or a twentieth of that amount, and if so the signs show that it probably flows from air to earth, but on account both of the smallness of the results and of the discrepancy between the values obtained for 1891 by two independent calculations, we cannot assert that such a current actually exists. The calculations do not disprove Dr. Schmidt's hypothesis, as we cannot argue from the condition of a small portion of the earth's surface to that of the whole. The most that can be said is that no evidence in favour of the existence of vertical currents can be drawn from one district which has been very minutely surveyed. 3. On the Equation Connecting the Potential Difference, Current, 4. On the back E.M.F. and True Resistance of the Electric Arc. 5. Note on the Electrolysis of Iron Salts. By W. M. HICKS, F.R.S., and L. T. O'SHEA. To prepare iron free from sulphur and carbon by electrolysis we used a solution of the double ferrous ammonium chloride. Large masses of metal can only be obtained by paying particular attention to the following points. 1. The strict neutrality of the solution: its strength must be regulated so as to offer suitable resistance for the regulation of the potential difference between the terminals and of the current density. We used a five per cent. solution of ferrous chloride to which sufficient ammonium chloride was added to form the double salt. This strength must be maintained, for if the amount of iron salt falls too low the ammonium chloride is decomposed and ferrous hydroxide is precipitated. The solution should be free from ferric salt, as this causes the formation of ferric hydroxide, which settles on the cathode plate. The solution can be freed from ferric salts by shaking with reduced iron and filtering just before being used. 2. The current density. For electrolytic analysis Classen gives a current density of 0.05 to 1·0 ampère per 100 sq. cm., and S. P. Thompson 008 to 0.25 ampère per 100 sq. cm. for steel facing. We find that it is advisable to strike the deposit with a density of 0·2 ampère per 100 sq. cm., then reduce to 0.15 to 0.18 ampère per 100 sq. cm., and that this should not be exceeded, though densities as low as 0.08 can be used. The potential difference between the terminals we used was 07 volt, and this was obtained by placing a single storage cell, voltage 2, in series with a dilute sulphuric acid cell with lead electrodes and a small external resistance. 3. The electrodes. The cathode must present a perfectly clean surface. We used thin copper sheet and found the best method of cleaning was to flush it with nitric acid and then scrub it with excess of a strong solution of potassium cyanide. All parts of the cathode except that on which the deposit is to form must be insulated from the solution. This may be done by coating the necessary parts with Brunswick black, on which, if perfectly dry, the solution has no action. The anode was a sheet of rolled Swedish iron, and to prevent the impurities it contained from mixing with the electrolyte it was enclosed in a porous cell. The accumulation of sulphuric acid in the electrolyte was prevented as much as possible by changing the solution in this cell twice daily. The solution used contained one per cent. ferrous chloride. The surface of the cathode is covered with small conical cavities, due to the formation of microscopic gas bubbles; in the early stages of the deposition great care must be taken to remove these by periodically exposing the surface to the air and in addition rubbing the surface. By taking these precautions the process can be carried on without interruption, and a firm, coherent deposit obtained of great purity. 6. On a Magnetic Field Tester. By Professor W. E. AYRTON, F.R.S., and T. MATHER. 7. On inc Velocity of Light in Rarefied Gases through which an Electrical Discharge is passing. By EDWIN EDSER, A.R.C.S., and SYDNEY G. STARLING, A.R.C.S. Lord Rayleigh has published the result of a research relative to the velocity of light in an electrolyte through which an electric current is passing. Dilute sulphuric acid was used, and the conclusion reached was that the velocity was unaffected by the passage of the current. It seemed to us worth while to perform a similar experiment to the above, substituting a rarefied gas for the electrolyte. The conditions are somewhat different, for though it has been shown by Professor J. J. Thomson that the ordinary stratified discharge can be assimilated to a current passing through an electrolyte by considering each stratification to correspond to a Grotthus chain, yet there is left still outstanding the phenomenon known as the kathode rays. Professor J. J. Thomson considers these rays to consist of a number of atoms, each carrying its tubes of force, and moving with a velocity comparable with 2 x 105 cm. per second. If these tubes of force consist of or comprise a number of vortex filaments of the ether, we might reasonably hope to detect some difference in the velocity of the light, according as it moves with or against this ether current. The method used by us is essentially that employed by Professor Michelson in re-performing Fizeau's experiment, two vacuum tubes with plane glass ends being substituted for the tubes through which in his experiment water was caused to flow. A discharge was produced in these tubes by a Ruhmkorff coil, and the interference bands were carefully observed. No alteration could be detected on the discharge being reversed. The experiment was repeated at various pressures (determined by the McLeod gauge) with the same result. A valid objection which might be raised to the above experiments is, that on account of the extremely short duration of these discharges, which succeeded each other only a few times a second, any shift of the bands would not be capable of detection by the observer. The most satisfactory results could be obtained by using a constant current from a battery of a large number of cells; but as these were not at our command, we attempted to obtain a prolonged discharge from a battery of ten one gallon Leyden jars, a piece of wet string being included in the circuit. The jars were charged by means of the above-mentioned Ruhmkorff coil, suitable arrangements being made for the purpose; and the discharges occasionally succeeded each other so quickly that for some seconds the phenomena were apparently continuous. The duration of each discharge was determined by means of Mr. Boys's wheel of lenses, and was found to exceed second. Not the slightest flicker of the bands could, however, be detected. A check experiment was then performed by dropping a piece of plate glass, which would produce a shift of the bands when placed in the path of the light, from such a height that its effect would last second and second respectively. In each of these cases a flicker was observed. From this we conclude that the kathode rays and the positive column, either alone or conjointly, do not affect the velocity of light passing along their path. 8. On the Hysteresis of Iron in an Alternating Magnetic Field. The law connecting hysteresis and induction when the latter reaches a high value has not until now been ascertained. It has hitherto been assumed that the hysteresis would increase continuously with the induction without limit. It is, however, probable that in a slowly performed cycle of change of magnetisation a maximum value of the hysteresis will be reached when the iron is saturated, and that further increase of the number of lines of force induced will produce no further increase in the hysteresis. When a rapidly alternating field is considered the conditions are somewhat different. As the magnetising current is increased the iron reaches its saturation value at an earlier point in each period, thus increasing the rate of change of magnetisation. But it has been shown by different experimenters that the value of the hysteresis per cycle is practically unchanged through wide variations in speed, and hence it may be expected that the hysteresis of iron under all conditions will arrive at a definite maximum. To verify this experimentally, the hysteresis in a small sample of iron was measured, when it was placed between the poles of a powerful electromagnet excited by an alternating current. Both magnet and sample were laminated, the subdivision of the latter being especially fine, in order to eliminate as far as possible errors due to eddy currents. The sheets consisted of soft charcoal-iron of thickness 0085 cm., and between each was a layer of tissue paper. The maximum value of the eddy currents was less than 2 per cent. of the hysteresis. The hysteresis was measured by the rise in temperature of the iron after 90 seconds, the speed of alternation being constant at 103 cycles per second. The sides of the sample were coated with layers of sheet cork, the radiation being very small. At the end sufficient protection could not be allowed owing to the small size of the air-gap, and hence transference of heat was prevented by maintaining equality of temperature between the pole pieces and the sample by means of streams of hot or cold paraffin oil over the pole pieces. All temperatures were measured by thermoelectric couples of german silver and copper. The experiments show that the curve when plotted with the induction as abscissa, and the hysteresis as ordinate, exhibits a flexure at an induction of about 16,000, and becomes practically horizontal at 23,000. This corresponds to a value of intensity of magnetisation of 1,640, which is just the saturation value. The same characteristics are observed when the intensity of magnetisation is taken as the abscissa, the curve bending over until it is almost horizontal at the point of saturation. The experiments prove that the hysteresis of iron is a function, not of induc tion, but of intensity of magnetisation, since both values become constant together, and that the relation between them is not a logarithmic curve, but is a curve showing one flexure, and clearly indicating in its upper portion the condition of the iron represented by Professor Ewing's third stage of magnetisation. WEDNESDAY, SEPTEMBER 18. The following Papers and Report were read : 1. On the Change of Molecular Refraction in Salts or Acids dissolved in Water. By Dr. J. H. GLADSTONE, F.R.S., and WALTER HIBBERT, F.I.C. The authors had recently undertaken a research on the questions-Does the specific refractive energy of a salt or acid when deduced from its solution in water differ from that of the solid compound? and Does the specific refractive energy vary according to the amount of water? The outcome of the experiments of the authors and others is that the water does bring about in many cases a small alteration, especially in passing from the solid or liquid to the dissolved condition; that this alteration is sometimes an increase, at other times a decrease; and that it depends upon the chemical nature of the compound. Some points of physical interest were, however, noted, and were being more fully examined. One of these is the analogy of this refraction change in several instances with the change in the power to rotate the plane of polarised light as determined by Dr. Perkin. As this small change of refraction evidently indicates some rearrangement of the constituents of the salt or acid in the water, it may throw light upon present theories of solution. The experiments, even those of Kohlrausch and Hallwachs on extremely dilute solutions, do not support the view that the binary compound when greatly diffused tends to exhibit the properties of a gas. There is an evident relation between this change of refraction and the electric conductivity of the solution. Thus, in the acids the order of the two phenomena is the same, the hydrochloric acid showing the greatest effect, rapidly followed by hydrobromic and hydriodic acids; then nitric acid, afterwards sulphuric acid; and at a great distance acetic and other organic acids. In the case of nitric and sulphuric acids the general form of the curves representing the change of electric conductivity and of refraction is similar; in the case of the latter there is a special depression during the rise of the conductivity curve which makes its appearance as a slackening of the rise in the curve representing change of refraction. This connection of the two phenomena is being further carefully examined at present. 2. Report on Electrical Standards.-See Reports, p. 195. 3. On the Choice of Magnetic Units. By Professor SILVANUS P. THOMPSON, F.R.S. Professor Silvanus Thompson pointed out that the giving of names was a detail, and that agreement was wanted upon the units themselves in which magnetic quantities were to be expressed. He agreed with the Standards Committee that the two most important units to be defined were those of magnetic flux and of magnetic potential, and urged that no other units should be defined until these had been tried. But he differed from the suggestion to take the weber as 108 C.G.S. lines as being a unit of too great an order of magnitude to suit practical needs. He preferred simply to take the line, with its natural multiples the kiloline, and the megaline as the unit of flux. If the name weber were given to the line itself the Committee's recommendation would then be identical, so far as this unit He is concerned, with that of the American Institute of Electrical Engineers. agreed with the propositions to adopt the name gauss for the C.G.S. unit of magnetic potential. 4. On some New Methods and Apparatus for the Delineation of Alternate Current Wave Forms. By J. M. BARR, W. B. BURNIE, and CHARLES RODGERS. 5. On Alternating Wave Tracers. By Professor W. E. AYRTON, F.R.S., and T. MATHER. 6. On the Relation between Speed and Voltage in Electric Motors. 7. On some recent Improvements in Measurements of High Temperatures. Illustrated by Apparatus recently acquired by the Kew Observatory Committee. By E. H. GRIFFITHS, F.R.S. |