July 18, 1902.] 



SCIENCE. 



89 



an aperture of ten centimeters. A narrow 

 strip across this mirror was divided into 

 equal bright and dark spaces of ten to the 

 millimeter. This was accomplished by 

 means of a diamond making five deep 

 strokes in each alternate space. In this way 

 normally incident light would be returned 

 over the same path from the bright spaces, 

 but be scattered from the cuttings in the 

 adjacent spaces, so that very little of it 

 would be returned in the direction of inci- 

 dence. A lens converged the sun's rays 

 on the slit S from which a beam passed to 

 the mirror P through the collimator lens T 

 to the upper part of the adjoining face of 

 the rotating mirror R from which it was 

 reflected at 45° through one quarter of the 

 lens L and brought to a focus on the grat- 

 ing mirror G. From this it was reflected 

 through the lower opposite quarter of the 

 lens to the lower part of the next face of E 

 and thence reflected at 45° to the plane 

 mirror M as parallel light. This reflected 

 it to the upper part of the same face of B 

 thence through L to G and back to the lower 

 half of the face of B upon which it was 

 first incident and thence through the ob- 

 ser\'ing telescope T' below the collimator, 

 to the eye. It is clear from the diagram 

 and from the principle that an even num- 

 ber of reflections of a ray system from a 

 moving reflecting system does not alter their 

 direction that the incident and the reflected 

 rays from the rotating mirror B will re- 

 main parallel to each other and hence will 

 always meet the mirror M at the same point 

 diiring the rotation of B. For the moment 

 let us assume the image of the slit S just 

 covers one of the bright spaces. By proper 

 adjiTstment of M the return image can be 

 made to coincide with it. Usually it was 

 displaced slightly below it so as to observe 

 the relative positions of the two when the 

 mirror B was rotated so as to carry the 

 images across the grating. If now, during 

 the time of transit from G and back, the 



mirror has rotated through an odd number 

 of spaces no appreciable light will be re- 

 flected from G through T' to the eye. If 

 the rotation corresponded to an even num- 

 ber of spaces, the eye would see an enfeebled 

 image of the slit ;S'. If the mirror were 

 varying in speed the eye would see this 

 image pass successively through maxima 

 and minima, depending on the rate of 

 change of the rotating mirror B. Suppose 

 now the image of 8 covers any niunber of 

 spaces on G the eye will see an image of 8, 

 but crossed by bright and dark spaces cor- 

 responding to those on G. With the corre- 

 sponding variations in the speed of the 

 rotating mirror the eye will see correspond- 

 ing fluctuations in this image. In this way 

 the eye may be able to determine the 

 minima by comparison with the darker 

 spaces which remain of constant intensity. 



The aperture of the 'sending' telescope 

 was 4.5 cm. but the effective aperture with 

 the rotating mirror at 45° was only 2.5 

 cm. The actual spacing was .02 cm. be- 

 tween the bright lines, which is Avell wdthin 

 the limits of good definition. The mirror 

 could be driven up to 250 revolutions or 

 more per second without serious vibration. 

 Thus the ray could be interrupted about 

 10,000,000 times in a second. This would 

 give a group of waves about fifteen meters 

 long. If the limits of resolving power and 

 speed were used 40,000,000 interruptions 

 could be obtained and the length of the 

 groups could be reduced to less than four 

 meters. Thus there would be about 6,000,- 

 000 waves in each group. As the eye can 

 observe a change in intensity of less than one 

 per cent., the method would thus be capable 

 of detecting the existence of a velocity if the 

 total distance of the mirror M were less 

 than 2 cm. from the grating. This shows 

 the sensibility which the combination of 

 the methods of Fizeau and Foucault may 

 give under very favorable conditions. 



If the velocity were different for difi'er- 



