SUMMARYTwelve white and 12 dark (unpainted) hives were set up in a sunny location in Baton Rouge, Louisiana, USA, on 7 May 1992. Each hive received a uniform colony of 5 048±125 (mean±s.d.) honey bees (Apis mellifera) that had been established from a single, artificially mixed population; 38% of the bees were infested with tracheal mites (Acarapis woodi). During June and July, bees produced more brood and honey in white hives than in dark hives. On 5 August, dark hives contained fewer tracheal mites than white hives; mites were found in 1 dark and 10 white hives (P<0.001). In summer, the temperature in the space between brood combs in the dark hives often exceeded 40°C; maximum temperatures in the brood areas of dark and white hives, were 45°C and 38°C respectively. Laboratory tests showed that heat can kill tracheal mites inside live bees. A single six-hour exposure of bees to 42°C, a condition comparable to the short periods of high temperature encountered in field colonies, significantly reduced mite populations. In a second test, three-day-old worker bees were kept at 34.5°C or 39°C for 48 h and evaluated four days later (both groups of bees were stored at 34.5°C during the last four days). Bees kept at 34.5°C (controls) contained 2.7 larval mites per adult mite; bees exposed to 39°C contained 0.01. Controls contained nearly twice as many eggs (2.8vs.1.5 eggs per adult mite). This suggests that existing eggs died or did not develop during the 48 h at 39°C and that viable eggs were not produced during that period.