On the theory of planar spectrographs

Abstract
The theory of planar spectrographs is presented. It is proven that by no means all aberrations up to the order of x/sup 4/ can be corrected over a finite spectral range. A general procedure to construct gratings with two stigmatic points of (nearly) arbitrary wavelength and location is proposed. Rowland-type and flatfield spectrographs are discussed as numerical examples.< >