From the Monthly Notes of the Royal Astronomical Society (1984) Vol. 210 pp 597-609
The Mersenne-Schmidt: a three-mirror survey telescope. Accepted 1984 March 20, Received 1984 February 10.

1. Introduction
      The Schmidt camera (Schmidt 1931) is an almost ideal survey instrument, especially if it is fitted with an achromatic corrector (Hendrix and Christe 1939; Wynne 1981). The performance of this type of instrument has been demonstrated by the UK Schmidt which was the first large one to be fitted with an achromatic corrector.

      Nevertheless, there would be problems in building a substantially larger camera of this type. Glass for the corrector might not be available in sufficiently large pieces of suitable homongencity. The tube length, more than twice the focal length, demands a large and costly dome.

      Even when fitted with an achromatic corrector of Schott UBK7 and LLF6 glass as in the UK Schmidt (Wynne 1981) the Schmidt camera suffers from chromatic variation of spherical aberrating, which amounts to 0.5 arcsec on axis at 370nm, 520nm and 1.0um if it is designed to be diffraction-limited at 400nm and 800nm and its aperture is F2.5. Reducing the size and cost of the dome by reducing the focal length (at constant aperture) would result in larger images on axis or a more restricted spectral range. Off-axis the image quality would also be affected (Bowen 1967).

      Triple lens correctors have been designed for the hyperboloidal primary mirrors of Ritchey-Chretien telescopes by Wynne (1965, 1968) who was also the first to design a three lens corrector for paraboloidal mirrors that would give images well under 1 arcsec in diameter over a field of 1º (Wynne 1974). The corrector for the 4.2 meter F2.5 William Herschel telescope is based on a design by Wynne, modified by Bingham, and (in perfect seeing) will cover a field of 42 arcmin in diameter, with images under 0.25 arcsec at the best wavelength, and under 0.5 arcsec over the range 365 to 1014nm (R. G. Bingham, private communication). However, fields of 1º or smaller are not suitable for a telescope intended for survey work. The Ritchey-Chretien telescope (Chretien 1922, Danjon and Couder 1935) is limited by astigmatism to fields of about 40 arcmin in diameter and works well at about F/8. The modified form of Ritchey-Chretien (Gascoigne 1965) has been shown in the case of the Irenee DuPont telescope (Bowen and Vaughan 1973) to cover a field of 2.1º in diameter, but still is about F/7.5, and baffles of an unusual form are require to prevent stray sky light from reaching any part of the field.

      The Schwarzschild (1905) telescope has a flat field but substantial astigmatism, resulting in an image 18 arcsec in diameter at the edge of a field of 3º, when working at F/3 (Danjon and Couder 1935). The couder (1926) telescope, sometimes known as the Schwarzschild anastigmat, can be modified to give images smaller than 0.4 arcsec over 3.6º at F/2.5, or smaller than 1 arcsec over 4.5º at F/2.0 (Willstrop 1983) but the curvature of field is slightly more than twice as great as in a Schmidt of the same focal length, and the usable field is therefore limited to about 3º square if glass plates re to be bent to fit the focal surface (Rule 1975). The tube length, including a sky fog baffle is 2.5 times the focal length, leading to a dome of approximately the same size and cost as for a large Schmidt.

      Other forms of coma-free, two-mirror systems considered by Schwarzschild (1905) have the convex mirror facing the incident parallel light, and a larger concave secondary mirror. These system from the basis of reflecting microscope objectives, when used with the light reversed, but they do not lead to any practical form of telescope.

      It therefore appears to be necessary to consider three-mirror forms of telescope to obtain a compact, fast, wide-field instrument with good optical performance.