Errors of touch localization after hand nerve injuries are common, and their measurement is important for evaluating functional recovery. Available empirical accounts have significant methodological limitations, however, and a quantitatively rigorous and detailed description of touch localization in nerve injury is lacking. Here, we develop a new method of measuring touch localization and evaluate its value for use in nerve injury. Eighteen patients with transection injuries to the median/ulnar nerves and 33 healthy controls were examined. The hand was blocked from the participant's view and points were marked on the volar surface using an ultraviolet (UV) pen. These points served as targets for touch stimulation. Two photographs were taken, one with and one without UV lighting, rendering targets seen and unseen, respectively. The experimenter used the photograph with visible targets to register their locations, and participants reported the felt position of each stimulation on the photograph with unseen targets. The error of localization and its directional components were measured, separate from misreferrals-errors made across digits, or from a digit to the palm. Nerve injury was found to significantly increase the error of localization. These effects were specific to the territory of the repaired nerve and showed considerable variability at the individual level, with some patients showing no evidence of impairment. A few patients also made abnormally high numbers of misreferrals, and the pattern of misreferrals in patients differed from that observed in healthy controls. NEW & NOTEWORTHY We provide a more rigorous and comprehensive account of touch localization in nerve injury than previously available. Our results show that touch localization is significantly impaired following median/ulnar nerve transection injuries and that these impairments are specific to the territory of the repaired nerve(s), vary considerably between patients, and can involve frequent errors spanning between digits.