FIG. 196.—Plan of Spectroscope. T, eyepiece end of telescope, B interior tube, carrying A, cylindrical lens; D, slit of spectroscope; G, collimating lens; _h h_, prisms; Q, micrometer. (Huggins.) ] [Illustration: FIG. 197.—Cambridge Star Spectroscope Elevation. ] [Illustration: FIG. 198.—Cambridge Spectroscope Plan. ] These are shown in Fig. 195. _e_ is a reflecting prism, and F is another movable reflector to reflect the light from a spark passed between two wires of the metal to be compared, and to throw it on the prism, which reflects the light through the slit of the spectroscope to the prisms and eye; if the instrument were in perfect adjustment and turned on a star, and a person were to place his eye to the spectroscope, he would see in one-half of the field of view the spectrum of the star with dark lines, and in the other half the spectrum of the vapour with its bright lines; and if he found the bright lines of the vapour to correspond with any particular dark line of the spectrum of the star, he would know whether the metal exists at that star or not; so this little mechanical arrangement at once tells him what there is at the star, whether it be iron or anything else. In Figs. 197 and 198 is shown another form of stellar spectroscope, that of the Cambridge (U.S.) observatory; it is the same in principle as that just described. A direct vision star spectroscope is shown in Fig. 199. [Illustration: FIG. 199.—Direct-vision Star Spectroscope. (Secchi.) ] A new optical contrivance altogether has to be used when star spectra are observed. The image of a star is a point, and if focussed on the slit will of course give only an extremely narrow spectrum; to obviate this a cylindrical lens is employed, which may be placed either before the slit or between the eyepiece and the eye. If placed before the slit, it draws out the image of the star to a fine line which just fits the slit, so that a sufficient portion of the slit is illuminated to give a spectrum wide enough to show the lines, or the slit may be dispensed with altogether. In stellar observations, when the cylindrical lens is used in front of the slit, special precautions should be taken so as to secure that the position of the cylindrical lens and slit in which the spectrum appears brightest should be used. In any but the largest telescopes the spectra of the stars are so dim that unless great care is used the finer lines will be missed. A slit is not at all necessary for merely seeing the spectra; indeed they are best seen without one. If a slit be used, it should lie in a parallel and not in a meridian; under these circumstances slight variations in the rate of the clock are of no moment.