An optical imaging study of 0.4

We performed optical imaging of 102 radio-loud and radio-quiet quasars at $z=0.4{-}0.8$, of which 91 fields were found suitable for host galaxy analysis after the deselection of saturated and otherwise flawed images. The data sets were obtained mainly in the R band, but also in the V and I or Gunn i band, and were presented in Rönnback et al.(1996, MNRAS, 283, 282) and Örndahl et al. (2003, A&A, 404, 883). In this paper we combine the two above-mentioned samples and also separately discuss additional hosts, extracted from data taken by Wold et al. (2000, MNRAS, 316, 267; 2001, MNRAS, 323, 231). The joint sample forms a sizeable fraction of the to-date total number of observed sources at intermediate redshifts and increases the number of resolved radio-quiet hosts at $z>0.4$ considerably. Equal numbers of radio-loud and radio-quiet objects were observed, resulting in a detection rate of 79% for the radio-loud hosts and 66% for the radio-quiet hosts. Profile fitting could only be carried out for a minority of the sample, but it results in predominantly elliptical morphologies. This is consistent with the mean values of the axial ratios, for which we find $b/a\ga0.8$ for both radio-quiet and radio-loud hosts, just as in the case of normal elliptical galaxies. The mean absolute magnitudes of the radio-loud and radio-quiet hosts is MR=-23.5 in both cases. This similarity between the mean magnitudes of the two types of host galaxy is also seen in the other imaged bands. While the radio-loud host absolute R magnitudes are correlated with redshift, only a weak trend of the same sort is seen for the radio-quiet host magnitudes. Note, however, that the sample is not fully resolved and that the detection limit, in combination with the relationship between host and nuclear luminosity, may conspire in creating the illusion of an upturn in magnitude. The average nucleus-to-host galaxy luminosity ratios of the radio-loud and radio-quiet objects do not differ significantly in any band, nor is the difference between the average luminosity ratios of flat spectrum and steep spectrum radio-loud quasars larger than $1.5\sigma$. Thus, no effect of beaming (as expected in the unifying scheme) is seen. The colours of both radio-loud and radio-quiet host galaxies are found to be as blue as present-day late-type spirals and starburst galaxies. These blue colours are most likely due neither to galaxy evolution over the range, which only gives rise to a colour shift of ~0.2 mag, nor to scattered nuclear light, since colours determined from annular apertures yield very similar results. Since close companions in projection are not uncommon (and a few sources even exhibit tidal tail-like features and other signs of interaction), ongoing star formation is a reasonable explanation of the blue host colours. As multiple-band imaging primarily was carried out for quasars showing indications of the presence of a host galaxy, the colour analysis results are valid for host galaxies which are large, bright, have low nucleus-to-host luminosity ratios, and/or display large scale disturbances, but cannot however safely be generalised to hold for the quasar host galaxy population at intermediate redshift as a whole.