In a number of recent field studies, the positive response of soil respiration to warming has been shown to decline over time. The two main differing hypotheses proposed to explain these results are: (1) soil microbial respiration acclimates to the increased temperature, and (2) substrate availability within the soil decreases with warming so reducing the rate of soil respiration. To investigate the relative merits of these two hypotheses, soil samples (both intact cores and sieved samples) from a 3-year grassland soil-warming and shading experiment were incubated for 4 weeks at three different temperatures under constant laboratory conditions. We tested the hypothesis that sieving the soils would reduce differences in substrate availability between warmed and control plot samples and would therefore result in similar respiration rates if microbial activity had not acclimated to soil warming. In addition, to further test the effect of substrate availability, we compared the respiration rates of soils taken from shaded and unshaded plots. Both soil warming and shading significantly reduced respiration rates in the intact cores, especially under higher incubation temperatures. However, sieving the soil greatly reduced these differences suggesting that substrate availability, and not microbial acclimation to the higher temperatures, played the dominant role in determining the response of heterotrophic soil respiration to warming. The effect of shading appeared to be mediated by reduced plant productivity affecting substrate availability within the soil and hence microbial activity. Given the lack of evidence for thermal acclimation of microbial respiration, there remains the potential for prolonged carbon losses from soils in response to warming.