ABSTRACT This paper reviews the question of whether the wide separation double quasar Q2138-431 is a gravitational lens. From early work, the two quasar images are known to have almost identical spectra and redshifts, but no lensing galaxy has so far been detected. In this paper, we used recent deep surveys in infrared and optical bands to search for the presence of a galaxy with the expected properties of a gravitational lens. The search revealed a 5σ detection of a faint galaxy between the two quasar images on a deep J-band frame from the VISTA Science Archive, with apparent magnitude J = 20.68. Non-detection in the I-band implied a redshift z > 0.6, and mass modelling of the quasar system gave a mass of $1.31 \times 10^{12} \, \mathrm{M}_\odot$ for the lensing galaxy, with mass-to-light ratio M⊙/L⊙ = 9.0. Archival photographic data from the UK 1.2m Schmidt telescope covering 25 yr were used to construct light curves for the two quasar images, which were then cross-correlated to measure any time lag. This showed image B to lead image A by around a year, consistent with 410 d from the mass model. Although the similarity of the spectra and the detection of the lensing galaxy are the most compelling arguments for the classification of Q2138-431 as a gravitational lens, the time delay and mass-to-light ratio provide a consistent picture to support this conclusion. The wide separation of the quasar images and the simplicity of the mass model make Q2138-431 an excellent system for the measurement of the Hubble constant.