February 5, 1920] 



NATURE 



605 



tion, thus causing its curvature to change. The 

 result was that the reflected beam became alter- 

 nately more and less divergent, so that the 

 amount of light incident on the selenium executed 

 fluctuations of the original frequency and ampli- 

 tude, and the speech sounds were reproduced in 

 the telephone receiver. Other forms of transmitter 

 are also described by Graham Bell, but it is 

 doubtful whether they were actually successful 

 in practice. 



Little further work on the subject appears to 

 have been done until about 1900, when Ernst 

 Ruhmer carried out for the German Government 

 a long series of experiments. He approached the 

 problem from a different point of view. Instead 

 of seeking to impose fluctuations of intensity on a 

 beam of light from a constant source, as Graham 

 Bell had done, he arranged to control the bright- 

 ness of the source itself by means of the vibrations 

 of speech. The sensitive or speaking arc was 

 already known, and Ruhmer improved it for the 

 purpose of light telephony. Briefly, the principle 

 lunounts to this. The current in an electric arc 

 controls the brightness of the arc. Variations of 

 current produce variations of brightness. By 

 means of a transformer, the fluctuations of current 

 in a microphone actuated by speech can be intro- 

 duced into the arc circuit, and thus produce 

 changes of brilliancy corresponding to the speech 

 vibrations. Ruhmer succeeded in perfecting this 

 system, and claims to have communicated speech 

 over several miles by projecting the beam from 

 the fluctuating arc, by means of a searchlight 

 reflector, on to a distant selenium cell. 



This method suffers from several disadvantages, 

 of which the chief is that it is limited to the arc 

 as a source of light, and rules out the use of that 

 much more efficient source — the sun. It is difficult 

 also to maintain the arc in the necessary sensitive 

 condition ; it requires continual adjustment. It 

 was these considerations which caused the present 

 writer, in working for the British Admiralty on the 

 subject in 1916, to revert to the general method 

 adopted by Graham Bell — namely, to interrupt the 

 light after it had left the source. The essential 

 point which had to be borne in mind was that the 

 vibrations which it is possible to impart by speech 

 to a diaphragm are of very small amplitude — a few 

 tiiousandths of an inch only. In order to use 

 these vibrations for producing large fluctuations 

 of intensity in a beam of light, magnification is 

 necessary. In the transmitter about to be de- 

 scribed, it will be seen that the magnification is 

 optical. There are many possible variations of 

 the apparatus, but the essential features are shown 

 in Fig. 2. Speech sounds enter the trumpet and 

 fall upon the diaphragm of a gramophone sound- 

 box. To the lever of this sound-box, at the place 

 which the needle ordinarily occupies, is attached 

 a small galvanometer mirror. The vibrations of 

 the diaphragm cause this mirror to execute small 

 angular oscillations about an axis perpendicular 

 to the plane of the diagram. Light from a suit- 

 able source, such as an arc, or, it may be, the 

 NO. 2623, VOL. T04] 



sun, is focussed by means of the first lens upon 

 the vibrating mirror; thence it is reflected through 

 the second lens. The focus of this lens coincides 

 with the vibrating mirror, so that the emergent 

 beam is a parallel one. 



A grid consisting of equal and parallel strips 

 alternately opaque and transparent is placed close 

 to the first lens, and a second equal grid near the 

 second or projecting lens. The result is that the 

 light from each point of the source is split up into 

 segments indicated by the unshaded portions, and 

 the extent to which the light penetrates the second 

 grid depends on the momentary position of the 

 vibrating mirror. As shown in Fig. 2, about 

 50 per cent, of the maximum is being projected, 

 but evidently if the mirror turns through a small 

 angle in a clockwise direction the reflected seg- 

 ments will turn also, and the light penetrating the 

 second grid will increase ; a counter-clockwise 

 movement of the vibrating mirror will, on the 

 other hand, diminish the light projected. 



Thus, in so far as the mirror copies the vibra- 

 tions of speech, and provided that the amplitude 

 is not allowed to be greater than that correspond- 

 ing to the width of one space of the grids, a 

 fluctuating beam of the desired character is ob- 



Diablii"ft«>n 



Fig. 2. 



tained. It may be projected on to the receiving 

 apparatus shown in Fig. i and used for the trans- 

 mission of speech. By making the width of the 

 grid spaces small in comparison with the distances 

 between the grids and the vibrating mirror, ade- 

 quate control of the light intensity is secured, 

 even though the movements of the diaphragm are 

 so small. It is, in fact, easy to reach the stage 

 when the grids must be made no narrower, other- 

 wise the amplitude of movement of the segments 

 of light is excessive, and the frequency of inter- 

 ruption becomes doubled or even trebled — to the 

 detriment of articulation in the received speech. 

 It should be pointed out that the diagram shows 

 only the light proceeding from a single point of 

 the source. Actually, every source is finite in 

 size, and in order to provide for this it is neces- 

 sary to use as the vibrating mirror a concave 

 reflector, the radius of curvature of which is equal 

 to the distance between the grids and the mirror. 

 A real image of the first grid is thus obtained on 

 the second, and this image moves in the manner 

 of a shutter when the mirror oscillates. 



