94 REPOKT OF THE SECRETARY. 



practice, for the artificial star itself could be obscured by a shutter immediately in 

 front of it, which could be removed instantaneously by the electric signal from the 

 personal-equation machine. The same thing could, however, be done at the focal 

 plane of the telescope in real use. 



Upon trial this device justified all hopes. Three observers whose habit is to 

 observe, one too early, one too late and the other very close to the true time, were 

 found all to observe as close to the truth as the accidental errors would admit, 

 which in the case of the 3-foot focus transit instrument employed was generally 

 within 0.03 second. In other words, personal equation seems to be wholly eliminated 

 by this procedure. The general design may be applied without great expense to any 

 outfit of transit and chronograph. 



Provininn for- great solar image.— As already said, the experiments on solar absorp- 

 tion, nature of sun spots and other phenomena required for their successful continu- 

 ation a large solar image free from rotation and as free as possible from "boiling" 

 and from ojjtical defects. Among these latter it will be evident that variations in 

 absorption and in magnification at different portions of the image were quite serious, 

 as well as the ordinary defects of definition. To magnify a small image to the 

 required diameter seemed to be very undesirable, for several reasons, chief of which 

 were the optical defects just spoken of and the increased "boiling" due to the 

 heating at the small focal image. The use of the siderostat was objectionable on 

 account of its rotating the field. 



You have therefore decided to use the coelostat and a concave mirror of long 

 enough focus to form the image of the desired size without a second magnifica- 

 tion. The coelosfat, which I understand you to have been first employed by you 

 as a working instrument on a large scale (in 1882), is essentially a plane mirror 

 rotating on an axis parallel with the axis of the earth at the rate of one revolution 

 in forty-eight hours. It is unable to send the reflected beams from objects at dif- 

 ferent declinations in the same direction, and if used to provide a horizontal 

 reflected beam from the sun, must send it in a more southerly direction in summer 

 than in winter. In this latitude the extreme directions reached by a horizontal 

 coelostat beam would be, respectively, about 30° north and 30° south of an east and 

 west line. To avoid moving the concave mirror and other apparatus to suit the 

 shifting direction of the coelostat beam, it has been determined to place a second 

 moval)le plane mirror close to the coelostat to reflect the beam to the concave mirror, 

 retaining the latter fixed. This mirror is so mounted that it can be wheeled north 

 or south along a track situated as close to the coelostat as possible. This track is 

 designed to curve round the south end of the coelostat so that the second mirror can 

 be either east or west of the coelostat as desired, and it is intended to use it east in 

 the morning and west in the afternoon. A concave mirror of 18 inches aperture and 

 140 feet focus which you have ordered will be placed about 60 feet north of the 

 coelostat. Its beam on the way to its focus will pass directly under the coelostat 

 mirror and between the morning and afternoon positions of the second plane mirror. 

 The coelostat mirror is thus about 2 feet higher than the concave mirror and about 

 1 foot higher than the second mirror, so that the latter casts no shadow on the 

 coelostat mirror except for very low sun at the times of the equinoxes. On account 

 of the unfavoral:)le inclinatioji of the coelostat mirrors at certain times of the year 

 and day, it is determined also to provide for the use of the two plane mirrors as a 

 polar sidero.stat, since only slight changes are required to alter a polar siderostat into 

 a coelostat, and the reverse. If used as a polar siderostat, the gain in effective mirror 

 surface will be offset by having a rotating image. 



In either case, as thus arranged, it is possible to c(jmpletely inclose the beam of 

 light in suitable tubes after it reaches the coelostat, and even before, by a tube moved 

 to follow the sun. The only use of such tubes is to jirevent the "boiling" or appar- 

 ent motion around the edges of the image due to air waves in the path of the beam, 



