In the past, nearly all spectrographs mounted on telescopes have had certain almost universal characteristics. The dispersing unit consisted of one or more glass prisms, the cameras used had lenses with glass optical components, and the diameter of the beam from the collimator was seldom more than two and onehalf inches. Although such instruments could produce excellent definition in the ordinary photographic region, these structural features caused some severe limitations. Below about À 3800

more »

... ectrographs of this type become virtually opaque, although this difficulty was overcome in a few cases through the use of quartz optical parts. In the red region of the spectrum the dispersion rapidly decreased and the definition generally deteriorated. Finally, the small diameter of the beam precluded the large ratio of collimator focal length to camera focal length that is necessary, if a wide slit, with corresponding gain in speed, is to be used without loss of definition. The older spectrographs were limited rather effectively, therefore, to the wavelength range 713800-5000, and in the sixty years or so since photography came into general use for stellar spectroscopy, this region has been thoroughly worked over. The design of a stellar spectrograph today should aim at removing these restrictions and at opening a much wider spectral range for investigation. At the same time the instrument should have maximum versatility. The term is here used in the sense that a variety of dispersions should be available and that for a given object and spectral region the appropriate dispersing medium and camerá can be quickly and easily put into use. Recently a new Cassegrain spectrograph for the 60-inch telescope has been placed in operation which was designed to overcome the limitations noted above. Instead of prisms, reflection gratings produce the dispersion and in order that maximum speed 346