Numerous studies have been attempted to determine the stellar content of the nuclei of galaxies. In the case of elliptical galaxies observations show that there is a change in spectral type from later to earlier types correlated with a variation from high to low luminosity (e.g. Faber 1977). This has been interpreted for the most part as being due to differences in metal abundances in an old stellar population, although some recent studies suggest a variation in stellar age may be important as well (e.g. O’Connell 1980, Heckman 1980). The nuclear bulges of spiral galaxies also display a change from late to early spectral type along a sequence from high to low luminosity bulges. In this case, however, the assumption has almost universally been made that these differences are due to differences in age of the stellar population. In particular, most stellar models for the nuclear bulges of spiral galaxies have used solar neighbourhood metal- rich stars, and assumed that the strong hydrogen lines and weak metal lines in late-type spiral nuclei are due to the predominance of a very young main-sequence population rather than old metal-poor stars.