Beats of Sounds led to the Two Ears separately. When two approximately pure tones, of equal intensity and of approximately equal frequency, are conveyed to one ear, beats are perceived according to a well-known elementary theory, the frequency of the beats being the difference of the frequencies of the tones. When the beats are somewhat slow, the phase of silence is distinctly recognizable, and indeed the moment of the occurrence of this phase is capable of being fixed with great accuracy.

The question whether the beats are still audible when one sound is led to one ear alone, and the second sound to the second ear alone, is of great importance. A careful experiment of this sort is described by Prof. S. P. Thompson*, in which the sounds were conveyed to the ears by rubber tubes; and the conclusion was that in spite of all precautions the beats were most distinctly heard, although there was no phase of "silence," such as is perceived when both sounds are conveyed to the same ear.

I have lately tried a somewhat similar experiment, using telephones and electrical conveyance, by which perhaps the risk of the sounds reaching the wrong ears is reduced to a minimum. Two entirely independent, electrically driven, forks of about 128 vibrations per second were the sources of sound. Near the electromagnet of each fork was placed a small coil of wire in connexion with a telephone. The higher harmonics were greatly moderated by the interposition of thick sheets of copper; but the sounds were doubtless no more than rough approximations to pure tones. Both forks were placed at a great distance from the observer; and in one case the doubled connecting wire was passed through a hole in a thick wall specially arranged many years ago for this sort of experimenting. When the telephones were pressed closely to the ears, the utmost possible was done to secure that each sound should have access only to its proper ear.

The results depended somewhat upon the frequency of the beats. When this exceeded one per second, the beats were very easily audible. When, on the other hand, the frequency was reduced to or beat per second, the beats were not easily perceived at first. After a little while the attention seemed to concentrate itself upon the variable element in the aggregate effect, and the cycle became clear. But even after some practice neither Mr. Gordon nor I could hear slow beats during the first 10 or 15 seconds of observation.

* Phil. Mag. vol. iv. p. 274 (1877).

The general results of the experiments do not appear to me to exclude the view that the comparatively feeble beats heard under these conditions may be due to the passage of sound from one ear to the other through the bones of the head or perhaps through the Eustachian tube.

Loudness of Double Sounds.

Observations upon the double syrens (with separate horns) used by the Trinity House have given the impression that as heard from a distance the two syrens are no better than one, even though the horns are parallel, and the observer situated in the direction of the axis. Dr. Tyndall's experience was similar. In his Report of 1874 he remarks (June 2), "There was no sensible difference of intensity between the single horn and the two horns ;" and again (June 10), "Subsequent comparative experiments even proved the sound of the two horns to be more effective than that of the three."

These conclusions are rather startling, suggesting the query as to what then can be the use of multiplying pipes in an organ or voices in a chorus. In order to clear the ground a little, I have recently tried some small-scale experiments with organ-pipes.

Two stopped pipes of pitch about 256 were mounted near the window of a room on the ground-floor. When the window was open the sounds could be heard (over grass) to about 200 metres; but when the window was closed the range was much less. Some difficulty was experienced in getting equal effects from the two pipes. According to the instructions of the observer, one or other supply-pipe was more or less throttled with wax.

With approximate equality of intensities and with such tuning that the beats were at the rate of about two per second, the results were very distinct. The beats were much more easily audible than either of the component sounds. Doubtless part of the advantage was due to the contrast provided by the silences; but it was thought that, apart from this, the swell of the beat was distinctly louder than either sound alone.

The result of the experiment is, of course, just what was to be expected from a mechanical point of view. According to theory the intensity (reckoned according to energy propagated) at the loudest part of the beat should be four times that of the (equal) component sounds heard separately.

In another set of experiments the pipes were mistuned until the interval was about a minor third, no distinct beats being audible. In this case the intensity of the compound sound

might be expected to be double of that of the (equal) component sounds. The impression upon the observer hardly corresponded to this anticipation. It was difficult to say that the compound sound was decidedly the louder; although the accession of the second sound as an addition to the first could always be distinguished, and this whether the higher or the lower sound were the one added. It may be remarked that the question involved in this experiment is partly physiological, and not merely mechanical as in the case of sounds nearly in unison,

Terling Place, Witham.

XXVI. The Electromagnetic Effects of Moving Charged
Spheres. By EDWIN P. ADAMS *.

THE A

HE magnetic effects due to moving charges were first shown experimentally by Professor Rowland † in 1876. Dr. E. Lechert, in 1884, thinking that the importance of the experiment made it desirable to repeat, it did so, but with negative results. Insufficient data regarding his experiment make it difficult to point out the cause of his failure to obtain the effect. Since then, Professor Rowland's results have been fully confirmed by Professor W. C. Röntgen § in 1885, by Rowland and Hutchinson || in 1889, and by Professor F. Himstedt ¶ later in the same year.

The next experimental attack upon this problem was by M. V. Crémieu ** at Paris in 1900-01. His experiments were originally undertaken to determine whether a changing magnetic field exerts a mechanical force upon an electrically charged body. The negative results obtained led him to undertake a series of experiments on the magnetic effect of moving electric charges. The results of these experiments are apparently all opposed to the results obtained by the above observers. The data which have thus far appeared do not give sufficient details to render it certain that positive results should have been expected. Crémieu himself states that he is convinced that the effect does not exist.

The great importance of the experiment would seem to

* Communicated by Prof. J. Trowbridge. + Am. Jour. Sci. (3) xv. p. 30 (1878). Rep. d. Phys. xx. p. 151 (1884).

§ Sitz. d. Berlin. Akad. p. 198 (1885).

Phil. Mag. [3] vol. xxvii. p. 445 (1889).

¶ Wied. Ann. xxxviii. p. 560 (1889).

** Comptes Rendus, cxxx. p. 1544; cxxxi. pp. 578, 797 (1900); cxxxii. p. 327 (1901).

render further investigation desirable, and the present paper contains a description of an experiment with this end in view.

All previous experiments have been made with rotating disks. With one exception, the direct effect of a charged rotating disk upon a magnetic needle has been examined. The exception referred to is the method employed by Crémieu in one of his experiments, where he sought to observe the inductive effect of charging and discharging a rotating disk upon a neighbouring coil of wire in circuit with a sensitive galvanometer.

The use of charged spheres was suggested by Professor J. J. Thomson* in 1881, and he gives a calculation of the magnetic force which would be produced by the motion of a sphere charged to the highest possible potential. In many respects this seems the most natural method of procedure and was adopted in this experiment. The description of the apparatus employed follows.

A hollow brass shaft, A A (figs. 1 & 2), is separated into two portions by the hard-wood bar B. The shaft turns in fibre bearings, and the pulley and belt at D communicate power from the countershaft F F. The spheres which carry the electric charges are spun out of sheet copper into hemispheres, soldered together. There are two sets of spheres, sixteen in each set. Brass rods pass through the hollow spheres, and are soldered to them. These brass rods are screwed into collars carried on the axle. The two sets of spheres are thus insulated from each other by the hard-wood bar, 8 cms. in length. The electricity is communicated to the two portions of the axle by the copper brushes CC, and the spheres, being connected with the axle, thus become charged themselves.

The speed-counter, E, is directly attached to one end of the axle. The gearing is so proportioned that one revolution of the crown-gear and dial I corresponds to 500 revolutions. of the axle.

The magnetic system upon which the direct effect of the moving charged spheres is observed, is inclosed in the brass tube H, closed on the bottom by a glass plate coated with tinfoil. This tinfoil is cut into strips parallel to the axle for the purpose of preventing conduction-currents flowing in it in a direction in which they could cause a deflexion of the needle. The needle-system is carried on a piece of mica 7.5 cms. long and 1.25 cms. wide. Pieces of wellhardened magnetized watch-spring are cemented at the lower *Phil. Mag. vol. xi. p. 236 (1881).