Pain memory is thought to affect future pain sensitivity and thus contribute to clinical pain conditions. Systematic investigations of the human capacity to remember sensory features of experimental pain are sparse. In order to address long-term pain memory, nine healthy male volunteers received intradermal injections of three doses of capsaicin (0.05, 1 and 20 microg, separated by 15 min breaks), each given three times in a balanced design across three sessions at one week intervals. Pain rating was performed using a computerized visual analogue scale (0-100) digitized at 1/s, either immediately online or one hour or one day after injection. Subjects also recalled their pains one week later. Capsaicin injection reliably induced a dose-dependent flare (p<0.001) without any difference within or across sessions. The strong burning pain decayed exponentially within a few minutes. Subjects were able to reliably discriminate pain magnitude and duration across capsaicin doses (both p<0.001), regardless of whether first-time ratings were requested immediately, after one hour or after one day. Pain recall after one week was similarly precise (magnitude: p<0.01, duration: p<0.05). Correlation with rating recall after one week was best when first-time ratings were requested as late as one day after injection (R(2)=0.79) indicating that both rating retrievals utilized similar memory traces. These results indicate a reliable memory for magnitude and duration of experimentally induced pain. The data further suggest that the consolidation of this memory is an important interim stage, and may take up to one day.