Marage1 has reproduced vowels with great success by means of a specially constructed siren, fortified by resonance cavities moulded from the oral cavities. To imitate A, the siren disc had triangular slits arranged in groups of three each. The number, n, of groups passing the air supply per second gave the frequency of the fundamental note. The oral resonator was tuned to 3n, the number of slits passing. To imitate O large slits were arranged in pairs, and to imitate E narrow slits were arranged in pairs. If n' pairs passed the air-supply per second, the oral cavity was tuned to 2n'. To imitate OU and I the slits were equidistant, large for OU, narrow for I, and the oral cavity was tuned to unison with the number of slits passing the air supply. These experiments lean towards the "relative pitch" theory, but an essential part of Marage's explanation is in the intermittent, the "group" character of the disturbance. The subject still obviously requires further research.

GRAPHIC MODE OF RECORDING VIBRATIONS.

The Phonautograph.-This is an instrument for recording the vibrations corresponding to various sounds, writing them down entire without any analysis. It consists of a parabolic reflector, with a membrane stretched across the narrow end at the focus. At the centre of the membrane, on the side away from the reflector, is a style which writes on a revolving drum covered with smoked paper, the motion of the style being parallel to the axis of the drum. The axle of the drum being cut with a screwthread, the drum moves forward as it is turned round. Any series of waves sent into the cavity of the reflector set the membrane vibrating, and its motion is recorded on the drum. This instrument is chiefly of interest now in that it was a great step towards a most remarkable invention, viz.—

Edison's Phonograph. This is an instrument which not only records the sounds which it receives, but also reproduces them.

Like the phonautograph, it consists of a stretched membrane, with a style at its centre recording on a drum; but instead of making the vibrations parallel to the axis, the style moves in and out, ploughing its way through the surface of the material of the drum, and recording the vibrations by a furrow of varying depth. In the original form of the instrument, Fig. 32, the drum had a screw-thread on it of the same pitch as that on the axle, and a sheet of tinfoil was wrapped round the drum. The style, which had a blunt point, indented the foil along the screw-thread, and the varying depth of the indentation recorded the vibrations. The drum was brought back to its original position and the style was allowed to press against it, so as always to be at the bottom of the indentation. On turning the drum at the same rate as

1 M'Kendrick, Nature, loc. cit.

FIG. 32.-Original form of the Phonograph, showing the principle. a a, axis with screw-thread turned in bearings 11 by the handle K; W, cylinder with screw-thread of same pitch as a a, and covered with tinfoil; m, mouthpiece, of which a side view is given in the lower figure; n, membrane connected with style p.

But

before, the membrane was made to execute the vibrations recorded on the foil, and so gave out again the waves at first put in. the membrane did not record all the original vibrations in the same proportion, so that the reproduced sounds were very different in quality from the original ones, though words could often be easily recognised. In later forms of the instrument devised by Tainter and Bell, as well as by Edison, the tinfoil cylinder is

FIG. 33.-Pendulum hung by string so as to execute two vibrations at right angles, one in the plane of the figure corresponding to length CD, the other in a perpendicular plane corresponding to length ED. The times are in the ratio/CD/ED.

FIG. 34.-Diagram of Harmonograph. The pendulum A vibrates on the knifeedge K1 (support not shown), giving a harmonic motion to a cylindrical table T at its upper end, K1 being the axis of the cylinder; B vibrates on the knifeedge K,, which is at right angles to K1, and gives its motion through a jointed framework to the siphon pen p. This moves harmonically in a line parallel to K1, and therefore, relative to the table, has a motion compounded of the two motions. The paper is attached to the surface of T.

replaced by one of wax, say 4 inches long and 2 to 4 inches in diameter, which can be turned by a motor at a very exactly governed speed. A mouthpiece closed by a disc of glass, say 20 inch thick, is placed in position in front of the cylinder. One end of a short lever is fixed to the centre of the disc, and the other end, on which is a minute cutting tool of sapphire, is

weighted so as to press against the wax. The cylinder is then set in motion, revolving and at the same time travelling forward along its axis, the tool ploughing a furrow in it. The words to be recorded are spoken into the mouthpiece, and duly imprinted by a series of varying depths in the furrow. To reproduce the sounds a second disc of glass, with a lever as before but with attached rounded sapphire point in place of a cutting tool, is brought into the position formerly occupied by the recording disc, the point being pressed slightly against the bottom of the furrow in the wax. When the cylinder is set in motion the original sounds are reproduced, and, under favourable circumstances, the original quality is exceedingly well preserved.

The Composition of Vibrations at Right Angles.-This subject may be usefully entered into at this point, though it has no direct connection with the foregoing discussion of the analysis

of vibrations. We shall see, however, that it leads to a method of detecting the form of vibrations which is of considerable importance that of the Vibration Microscope.

A simple mode of exhibiting the composition of two vibrations at right angles is shown in Fig. 33. A string, ACB, is fastened to two points, AB-say two nails in a doorway-and to its middle point C a third string CD is attached, having a heavy weight, D, at its lower end. If the weight is drawn aside in the vertical plane through AB, it vibrates as a pendulum of length, CD, and if in the vertical plane through ECD, it vibrates as a pendulum of length, ED. If drawn aside in any other way, and either released or projected, its motion is a resultant of these two motions. The motion may be recorded by making the weight D hollow, with a small hole at the bottom, and filling it with sand. The sand trickles out, and marks, on a sheet placed below,

the curve traced out. By varying the ratio ED/CD and the relative phases and the relative amplitudes, the figures obtained

&

D

B

B, Resultant when the two start

of a b.

oc. C, D, E. F, successive forms as the phase of c d lags

behind

that

FIG. 36.-A, Two directions of vibration; frequency in a b supposed greater than that in cd.

in the same

phase

along o a and

may be varied to any extent.

The Harmonograph. — This instrument was devised by Mr. A. E. Donkin for the purpose of compounding two harmonic vibrations. These may be either in the same line or at right angles to each other. By means of wheelwork and cranks a glass siphon pen has one harmonic vibration communicated to it, while a table on which it writes has the other. If the two harmonic motions are in one line the table is drawn along with uniform speed in a direction at right angles to the line of the harmonic vibrations. If either harmonic motion exists alone the pen describes a curve of sines (p. 65) of period and amplitude equal to those of the harmonic vibration. If the two coexist the pen describes a curve the resultant of the superposition of the two described separately. Fig. 34 shows a simple form of harmonograph in which the two vibrations to be compounded are at right angles to each other. Two pendulums, A and B, vibrate in perpendicular planes. A carries a small table on which is fixed a card, and B is connected by a lever to a pen p which writes on the card. Owing to the gradual decrease of the vibrations and the consequent interlacing of the successive curves traced out, it is most interesting to watch the gradual development of the figures. These are often of great beauty. Fig. 35 represents an imperfect unison and an imperfect fifth.

The Kaleidophone. If a knitting-needle is clamped in a vice and a small bead is fixed on its free end, the image in the bead

1 Proc. Roy. Soc., xxii., 1873-74, p. 196; Donkin's Acoustics, 2nd ed.