1 Introduction

The use of gravitational lensing statistics as a cosmological tool was first considered in detail by Turner et al. (1984); the influence of the cosmological constant was investigated thoroughly by Fukugita et al. (1992), building on the work of Turner (1990) and Fukugita et al. (1990). Kochanek (1996a, and references therein) and, more recently, Falco et al. (1998, hereafter FKM) have laid the groundwork for using gravitational lensing statistics for the detailed analysis of extragalactic surveys. Quast & Helbig (1999, hereafter Paper I) reanalysed optical surveys from the literature, for the first time exploring a range of the $\lambda_{0}-\Omega_{0}$ parameter space large enough to enable a comparison with other cosmological tests. Here, we use the formalism outlined in Paper I to analyse the Jodrell Bank-VLA Astrometric Survey (JVAS), the largest completed gravitational lens survey to date.

Radio surveys offer several advantages over optical surveys (see, e.g., FKM): one doesn't have to worry about systematic errors due to extinction or a lens galaxy of apparent brightness comparable to that of the lensed images of the source, the resolution (of followup observations if not of the survey proper) is much smaller than the typical image separation, parent catalogues in the form of large-area surveys exist from which unbiased samples can be selected and relatively easily observed. Disadvantages in the radio are due to our relatively poor knowledge of the flux density-dependent redshift distribution or equivalently the redshift-dependent number-magnitude relation.

For a description of our method see Paper I. The plan of this paper is as follows. Section 2 describes the JVAS gravitational lens survey. In Sect. 3 we describe the calculations we have done based on the JVAS data. Section 4 presents our results, using both the JVAS data alone and in combination with the results from the optical surveys analysed in Paper I. Finally in Sect. 5 we compare our results to those of Paper I and present our conclusions and our prognosis for the analysis of future large surveys such as CLASS.