Satellites in the field and lens galaxies: SDSS/COSMOS versus SLACS/CLASS

The incidence of subgalactic level substructures is an important quantity, as it is a generic prediction of high-resolution cold dark matter (CDM) models which is susceptible to observational test. Confrontation of theory with observations is currently in an uncertain state. In particular, gravitational lens systems appear to show evidence for flux ratio anomalies, which are expected from CDM substructures although not necessarily in the same range of radius as observed. However, the current small samples of lenses suggest that the lens galaxies in these systems are unusually often accompanied by luminous galaxies. Here, we investigate a large sample of unlensed elliptical galaxies from the Cosmic Evolution Survey (COSMOS) and determine the fraction of objects with satellites, in excess of background counts, as a function of satellite brightness and separation from the primary object. We find that the incidence of luminous satellites within 20 kpc is typically a few tenths of 1 per cent for satellites of a few tenths of the primary flux, comparable to what is observed for the wider but shallower Sloan Digital Sky Survey. Although the environments of lenses in the Sloan Lens ACS survey are compatible with this observation, the Cosmic Lens All-sky Survey (CLASS) radio lenses are significantly in excess of this.