VLA observations of the X-ray selected BL Lac H0414 + 009 at 6, 20 and 90 cm show that it has a steep-spectrum (|$\alpha \simeq 1; f_\nu \propto\nu^{-\alpha}$|) radio tail. The tail extends ∼ 40 arcsec from the core, in a gentle curve. There is no evidence of extended emission on the opposite side of the core. This is by far the clearest example of a single-sided radio tail associated with a BL Lac source. The low radio luminosity (|$1.6\times 10^{25} \text {W Hz}^{-1}$| at 20 cm for H0 = 50 km s−1 Mpc−1 and q0 = 1/2 at z = 0.287) and morphology of the extended emission strengthens the proposed association of BL Lacs with Fanaroff–Riley type I (FR I) radio sources. Deep V, R and I CCD images show that the BL Lac is located in a galaxy whose surface brightness and colours are best fitted by an elliptical model with absolute total magnitude Mv = − 23.53. The elliptical host galaxy is the brightest of a poor cluster. We confirm the recently published redshift (z = 0.287) for the BL Lac and find similar redshifts for five of the surrounding, faint galaxies, supporting the conclusion that the BL Lac and the cluster are physically associated and are not merely a chance super-position. We conclude, contrary to earlier suggestions, that BL Lacs do not avoid the cluster environment but are found in groups and clusters just as often as are normal radio galaxies, consistent with the hypothesis that BL Lacs are simply normal radio galaxies seen ‘end-on’. The appearance of the cluster is striking, containing a number of ‘twin’ galaxy pairs. It has been suggested that the pairs could be the result of gravitational lensing by a foreground cosmic string, but our observations do not support this hypothesis. We have analysed the cluster properties and present VRI CCD photometry for 108 galaxies fainter than the BL Lac host galaxy in the CCD images (1.1 × 1.2 Mpc2). In the Bautz–Morgan morphological classification scheme we classify this cluster as BM I. The cluster could be as rich as Abell class 0, however, the richness may be lower due to considerable uncertainties in the corrections for the field galaxy background and the small field of view of the CCD. We have searched deep EXOSAT images for evidence of extended X-ray emission from a hot intracluster medium, but find none. The upper limit to the luminosity of any cluster-type emission from 0414 + 009, however, is still rather high, |$L_\text x(2-10\text {keV})\le 4 \times 10^{37}\text W$|, comparable with that of moderately X-ray luminous Abell clusters. If hot gas emitting at that luminosity does exist in the cluster, it could help explain the unusual structure of the radio source.