Laboratory incubation studies evaluating the temperature sensitivity of soil respiration often use measurements of respiration taken at a constant incubation temperature from soil that has been pre‐incubated at the same constant temperature. However, such constant temperature incubations do not represent the field situation where soils undergo diurnal temperature oscillations. We investigated the legacy effects of constant and diurnally oscillating temperatures on soil respiration and soil microbial community composition. A grassland soil from the United Kingdom was either incubated at a constant temperature of 5°C, 10°C, or 15°C, or diurnally oscillated between 5°C and 15°C. Soil CO2 flux was measured by temporarily moving incubated soils from each of the above‐mentioned treatments to 5°C, 10°C or 15°C, such that soils incubated under every temperature regime had CO2 flux measured at each temperature. We hypothesised that, irrespective of measurement temperature, CO2 emitted from the 5°C to 15°C oscillating incubation would be most like the soil incubated at 10°C. The results showed that both incubation and measurement temperatures separately influence soil respiration. Oscillations between 5°C and 15°C resulted in significantly greater CO2 flux than constant incubations at 10°C or 5°C but were not significantly different from the 15°C incubation. The greater CO2 flux from soils previously incubated at 15°C, or oscillating between 5°C and 15°C, coincided with a depletion of dissolved organic carbon and a shift in the phospholipid fatty acid profile of the soil microbial community, consistent with stress associated with substrate depletion and microbial starvation when incubated at higher temperatures. Our results suggest that daily maximum temperatures are more important than daily minimum or daily average temperatures when considering the response of soil respiration to the diurnally asymmetric warming that is expected to occur as a result of climate change. [ABSTRACT FROM AUTHOR]