MOUNT WILSON OBSERVATORY. 243 



ample electric current for the arc lamp used as the source of light was available, 

 and where the supply of compressed air for driving the rotating mirror could 

 be maintained conveniently. 



The first trials were made with the apparatus arranged essentially as 

 described in the last report. For the distant mirror, however, a 22-inch con- 

 cave of 30 feet focal length was substituted, and mounted at its focus was a small 

 concave mirror of 30-feet radius, placed perpendicular to the axis of the large 

 mirror. The image of the source at Mount Wilson was thus focussed ac- 

 curately on the surface of the small mirror, thence returned to the 22-inch 

 concave, and so back to Mount Wilson as a parallel beam. The adjustment 

 of these mirrors remained surprisingly constant, and after the second trial 

 no further expeditions to the distant station were found necessary. On the 

 last trial of the apparatus the return light was found to disappear at a distance 

 of less than 1 foot from the edge of the concave mirror at the home station. 



When the rotating mirror was set in motion it was found at once that the 

 intensity of the return-beam was far too small for measurement. Accordingly, 

 Michelson had recourse to a totally different arrangement of apparatus, which 

 is essentially a combination of the Foucault and the Fizeau forms. The light, 

 after passing through a slit, is thrown by a lens upon the lower half of a face 

 of the octagonal rotating mirror. From this it is reflected and an image of 

 the slit is formed on the surface of a concave mirror of short focal length (in 

 practice a series of 10 small mirrors placed in tandem was used). It is then 

 reflected back to the upper half of the same face of the rotating mirror, thence 

 to the large 22-inch concave of 30-feet focal length, and so, as a very nearly 

 parallel beam, to the distant station. From this the light retraces its path 

 accurately to the source. A plane- parallel plate of glass is placed in the return 

 beam near the slit and observations are made with a low-power eyepiece. 



Under these conditions, when the octagonal mirror rotates slowly, the 

 intensity of the return beam, apart from losses by reflection and transmission, 

 will be proportional to the angle subtended by the battery of short-focus 

 concave mirrors. As the speed of the rotating mirror is increased, the 

 effective arc will become less until it is exceeded by the angle through which 

 the mirror turns during the time required for the light to pass to the distant 

 mirror and return, when the intensity will fall to zero. As soon, however, as 

 the speed of the mirror becomes so great that a second face replaces the 

 first, the intensity suddenly returns to its full value. The observer, ac- 

 cordingly, varies the speed of the mirror, 



V 186,000 „~n 

 n= 8^2D = 8^46 =5 ° 6 turnS a SeCOnd ' 



within narrow limits until the light suddenly reappears. 



With good seeing this reappearance should be very definite. The variation 

 in mirror speed, due to poor seeing, diffraction, etc., may be estimated at 



0.01 mm.: \ irr, or 1:55,000 

 where r, the focal length of the small concave mirrors, is 700 mm. This will 

 be the probable error of the result if the speed of the rotating mirror can be 

 measured with a corresponding order of accuracy. In some previous work the 

 error of this measurement was estimated at 1 part in 200,000, but the speed 

 was much less than that required for these observations. 



