July 18, 1902.] 



SCIENCE. 



91 



limit. Such a Nichol could probably be 

 rotated at a speed exceeding two thousand 

 per second. Polishing machines are now 

 run above one hundred thousand a minute. 

 Foucault's mirror rotated eight hundred 

 times a second. At such a speed and sensi- 

 bility a velocity in light could be detected 

 over a distance of about twenty centi- 

 meters. 



Another system depending upon electric 

 oscillations would reqi;ire much the same 

 optical system. A prism of glass between 

 two electrodes as arranged for the Kerr 

 'effect' in a dielectric is placed beyond a 

 half-wave plate and a Nichol prism which 

 is fixed. If now the glass is subjected to 

 electric strain by rapid oscillations, the 

 fields from each mirror are lighted up dif- 



of the velocity of that property to the wave- 

 velocity would have to be determined. 



No method for determining the wave- 

 velocity of light seems yet to have been 

 proposed. The following arrangement oc- 

 curred to the writer in 1890 while experi- 

 menting with the phototachometer as al- 

 ready described. Suppose that in two in- 

 terfering systems of rays we could alter the 

 length of the path of one or both of the rays 

 during their transit from their co mm on 

 source to their final point of meeting, there 

 would be a displacement of the bands de- 

 pending on the relative retardation intro- 

 duced into the paths during this interval. 

 Figure 3 illustrates the first conception of 

 the system. A beam of parallel light, from 

 a lens, say, strikes the two adjacent faces 



ferently and the limiting distance at which 

 this difference is observable will be the 

 same as stated previously. Probably oscil- 

 lations several hundred millions a second 

 would be possible with such a condenser 

 system. Forty millions a second was the 

 limit with a rotating mirror and grating 

 as described above. This could probably 

 be increased to one hundred millions a sec- 

 ond with a suitable rotating mirror. In- 

 stead of using the Kerr 'effect' a piece of 

 Faraday glass within a single turn of foil 

 could be used for very high frequencies 

 and the Faraday 'effect' employed instead. 

 Probably the same order of sensibility 

 could be obtained as with the former. It 

 is hoped to make preliminary experiments 

 on these promising methods which would 

 probably give shorter groups or types of 

 waves than any of the others. Here we 

 should have not an intermittence but a 

 property impressed at intervals upon a 

 continuous train of waves, and the relation 



of a rectangular mirror, each face of which 

 reflects the rays in opposite directions to 

 the mirrors 31 and M', each of which reflects 

 the corresponding rays to the other and 

 thence both return to the mirror, thus trav- 

 ersing each other's paths. If now the mir- 

 ror m is displaced to the position of m' in 

 this interval the path of one ray will exceed 

 that of the other by twice the distance 

 through which the mirror has moved. 

 Knowing this distance, by means of the 

 wave-length and the time it takes to dis- 

 place the mirror a given distance, we have 

 at once the time of displacement of the 

 mirror from the position m to m' and thus 

 the time of transit of any one ray around 

 this path, and hence the wave-velocity. 

 Considerable difficulty was experienced in 

 devising a method of displacement of suffi- 

 ciently high speed. If the mirror m is 

 mounted on a rotating disc the rays would 

 be reflected beyond the mirrors M and M' 

 and the interference would be changed by 



