1 Introduction

At sufficiently large angular distances from the centre of any galaxy image, the surface-brightness contribution due to the galaxy becomes, at some point, indistinguishable from the surface-brightness of the surrounding sky. The limit on reliable observation is determined primarily by the noise and often corresponds to those points at which the galaxy's surface brightness has fallen to several percent of the sky, though limits in the region of 0.1% of the sky are measurable on occasions.

Although total magnitudes are required for many astrophysical applications, it is therefore not possible to measure them directly. Instead, estimates are normally obtained by means of extrapolating model profiles fitted to those parts of the galaxy-light profiles (whether surface-brightness or integrated) that can be measured reliably. However, galaxies of different morphological type have very different profile shapes, and a wide range of models have generally had to be invoked. In Kron's (1980) system though, the flux due to a galaxy is measured to very large radial distances, so that different models do not need to be invoked. Such a procedure has its advantages, but at the very large radial distances involved, the signal due to the galaxy is often such a small fraction of the noise that large random errors cannot be avoided.

An alternative approach to profile extrapolation is very-low-resolution imaging so that, in theory at least, all target galaxies essentially become point sources and yield images of almost identical structure. Total magnitudes then become a simple function of the full-width half maximum (FWHM) of the image point-spread function. In practice, the FWHM of the point-spread function has to be much larger than the intrinsic angular sizes of the target galaxies for this method to yield reliable magnitudes. However, the wider the point-spread function is, the lower the mean surface-brightness of each image becomes and the greater the errors due to the noise become. In the compilation of their Catalog of Galaxies and Clusters of Galaxies, Zwicky et al. (1961, 1963, 1965, 1966, 1968) measured their total magnitudes visually from out-of-focus photographic plates. Out-of-focus images generally have a complicated point-spread function, but as long as they exhibit little variation across a single plate or CCD frame, this should not be a problem.

The new system presented in this paper combines both approaches. We believe that it can yield very large numbers of reliable total-magnitude measurements very efficiently, and that it is therefore particularly suitable for galaxy-survey work.