Sixteen 20-year-old Scots pine (Pinus sylvestris L.) trees growing in the field were enclosed for 4 years in environment-controlled chambers that maintained: (1) ambient conditions (CON); (2) elevated atmospheric CO2 concentration (ambient + 350 micro mol mol-1; EC); (3) elevated temperature (ambient +2-6 degrees C; ET); or (4) elevated CO2 and elevated temperature (ECT). The dark respiration rates of 1-year-old shoots, from which needles had been partly removed, were measured over the growing season in the fourth year. In all treatments, the temperature coefficient of respiration, Q10, changed with season, being smaller during the growing season than at other times. Respiration rate varied diurnally and seasonally with temperature, being highest around mid-summer and declining gradually thereafter. When measurements were made at the temperature of the chamber, respiration rates were reduced by the EC treatment relative to CON, but were increased by ET and ECT treatments. However, respiration rates at a reference temperature of 15 degrees C were reduced by ET and ECT treatments, reflecting a decreased capacity for respiration at warmer temperatures (negative acclimation). The interaction between season and treatment was not significant. Growth respiration did not differ between treatments, but maintenance respiration did, and the differences in mean daily respiration rate between the treatments were attributable to the maintenance component. We conclude that maintenance respiration should be considered when modelling respiratory responses to elevated CO2 and elevated temperature, and that increased atmospheric temperature is more important than increasing CO2 when assessing the carbon budget of pine forests under conditions of climate change.