In this and in other observational matters it is good to know what to look for, and there are great generic differences between the spectra of the various stars. In Fig. 200 are represented spectra from the observations of Father Secchi. In the spectrum of Sirius, a representative of Type I., very few lines are represented, but the lines are very thick; and stars of this class are the easiest to observe. Next we have the solar spectrum, which is a representative of Type II., one in which more lines are represented. In Type III. fluted spaces begin to appear; and in Type IV., which is that of the red stars, nothing but fluted spaces is visible, and this spectrum shows that there is something different at work in the atmosphere of those red stars to what there is in the simpler atmosphere of the first—of Type I. These observations were first attempted, and carried on with some success, by Fraunhofer, and we know with what skill and perseverance Mr. Huggins has continued the work in later years, even employing reference spectra and determining their chemical constitution as well as their class. [Illustration: FIG. 200.—Types of Stellar Spectra (Secchi). ] We need scarcely say that the same arrangement, minus the cylindrical lens, is good for observing the nebulæ and such other celestial objects as comets and planets. For all spectrum work, it has to be borne in mind that the best definition is to be had when the actual colour under examination is focussed on the slit. With reflectors, of course, there is no difference of focus for the different colours. As the best object-glasses are over-corrected for chromatic aberration, the red focus is generally inside and the blue one outside the visual one. It is not necessary to move the whole spectroscope to secure this; all collimators should be provided with a rack and pinion giving them a bodily movement backwards and forwards. This precaution is of especial importance in the case of solar observations, to which we have next to refer.