SUMMARY Honeybee colonies maintain brood nest temperatures of 33–36°C. We investigated brood nest thermoregulation at the level of individual worker behaviour and the transfer of heat from workers to the brood. Worker bees contribute to the regulation of brood nest temperature by producing heat while sitting motionless on the caps of brood cells. We report here an additional,newly observed heating strategy where heating bees enter empty cells between sealed brood cells and remain there motionless for periods of up to 45 min. Individually marked worker bees on the surface of sealed brood cells maintained thorax temperatures ( T th ) between 32.2±1.0°C and 38.1±2.5°C (mean ± s.d. ; N =20 bees) with alternating warming and cooling periods. Most of the observed bees made one or several long-duration visits(>2 min) to empty cells within the sealed brood area. T th at the time bees entered a cell[ T th(entry) ] was 34.1–42.5°C ( N =40). In 83% of these cell visits, T th(entry) was higher (up to 5.9°C; mean 2.5±1.5°C; N =33) than the mean T th of the same bee. High values of T th(entry) resulted from preceding heating activity on the comb surface and from warm-ups just prior to cell visits during which T th increased by up to +9.6°C. Bees inside empty cells had mean T th values of 32.7±0.1°C (resting bees) to 40.6±0.7°C (heat-producing bees) during long-duration cell visits without performing any visible work. Heating behaviour inside cells resembles heating behaviour on the brood cap surface in that the bees appear to be inactive, but repeated warmings and coolings occur and T th does not fall below the optimum brood temperature. Bees staying still inside empty cells for several minutes have previously been considered to be `resting bees'. We find, however, that the heating bees can be distinguished from the resting bees not only by their higher body temperatures but also by the continuous, rapid respiratory movements of their abdomens. By contrast, abdominal pumping movements in resting bees are discontinuous and interrupted by long pauses. Heat transfer to the brood from individual bees on the comb surface and from bees inside empty cells was simulated under controlled conditions. Heating on the comb surface causes a strong superficial warming of the brood cap by up to 3°C within 30 min. Heat transfer is 1.9–2.6 times more efficient when the thorax is in touch with the brood cap than when it is not. Heating inside empty cells raises the brood temperature of adjacent cells by up to 2.5°C within 30 min. Heat flow through the comb was detectable up to three brood cells away from the heated thorax.