STELLAR PHOTOGRAPHY. 209 



the spectra being formed of the trails of the stars. In the spectrum of the Pole-star 

 over a dozen lines could be counted. In the spectrum of a Lijrw the characteristic 

 lines were sliown very clearly. Exposures of two or three minutes were usually 

 employed, although one minute gave an abundant width. In the spectrum of 

 a Aquilce, besides the lines seen in a Lz/rce, some of the additional faint lines noticed 

 by Dr, Draper were certainly seen. 



In the autumn of 1885, two prisms were constructed, having clear apertures of 

 20 cm. and angles of about 5' and 15°. They could be placed over the object-glass 

 of the photographic telescope without reducing the aperture. The second of these 

 prisms was that actually employed in the experiments described below. 



The prism was always placed with its edges horizontal Avhen the telescope was 

 in the meridian. The spectrum then extended north and south. If clockwork 

 was attached, a line of light would be formed too narrow to show the lines of the 

 spectrum satisfactorily. The usual method of removing this difficulty is the employ- 

 ment of a cylindrical lens to widen the spectrum; but if the clockwork is discon- 

 nected, the motion of the star will produce the same effect. Unless the star is 

 very bright, the motion will, however, be so great that the spectrum will be too 

 faint. It is only necessary to vary the rate of the clock in order to give any 

 desired width to the spectrum. A width of about one millimetre is needed to 

 show the fainter lines. This distance would be traversed by an equatorial star 

 in about twelve seconds. The longest time that it is ordinarily convenient to 

 expose a plate is about an hour. If then the clock is made to gain or lose 

 twelve seconds an hour, it will have the rate best suited for the spectra of the 

 faintest stars. A mean- time clock loses about ten seconds an hour. It is only 

 necessary to substitute a mean-time clock for the sidereal clock to produce the 

 required rate. It was found more convenient, however, to have an auxiliary 

 clock whose rate could be altered at will by inserting stops of various lengths 

 under the bob of the pendulum. One of these made it gain twelve seconds in 

 about five minutes, the other produced the same gain in an hour. The velocity of 

 the image upon the plate when the clock is detached could thus be reduced thirty 

 or three hundred and sixty times. This corresponds to a difference of 3.7 and 6.1 

 magnitudes respectively. Since the spectrum of a star of the second magnitude 

 could be taken without clockwork, stars of the sixth and eighth magnitudes respect- 

 ively could be photographed equally well with the arrangement described above. 



A number of photographs were taken of various portions of the sky, and to 

 secure images of all the brighter stars the following system was also adopted. 



