Soils are frequently exposed to drying and wetting events and previous studies have shown that rewetting results in a strong but short-lived flush of microbial activity. The aim of this study was to determine the effect of the water content during the dry period on the size and duration of the flush and on the rate of recovery. Two soils (a sand and a sandy loam) were maintained at different water contents (WC) 30, 28 and 25 g water kg−1 soil (sand) and 130, 105 and 95 g water kg−1 soil (sandy loam) for 14 days, then rewet to the water content at which respiration was optimal [WC 35 (sand), WC200 (sandy loam)] and maintained at this level until day 68. Ground pea straw (C/N 26) was added and incorporated on day 1. The controls were maintained at the optimal water content throughout the 68 days. Respiration rates during the dry phase (days 1–14) decreased with decreasing water content. The flush of respiration after rewetting peaked on day 15 in the sandy loam and on day 16 in the sand; it was greatest in the soils that had been maintained at the lowest water content [WC25 (sand) and WC95 (sandy loam)]. Cumulative respiration during the remainder of the incubation period in which all soils were maintained at optimal water content increased more strongly in the soils that had been dry compared to the constantly moist control. On the final day of the dry period (day 14), cumulative respiration in the dry soils was 29–65% (sand) and 67–94% (sandy loam) of the constantly moist control whereas on day 68 it was 80–84% (sand) and 86–96% (sandy loam). The greater increase in cumulative respiration in the previously dry soils can be explained by the reduced decomposition rates during the dry period which resulted in higher substrate availability on day 14 compared to the constantly moist control. Microbial community structure assessed by phospholipid fatty acid analyses changed over time in all treatments but was less affected by water content than respiration; it differed only between the highest and the lowest water content. These differences were maintained throughout the incubation period in the sandy loam and transiently in the sand. It can be concluded that the soil water content during the dry phase affects the size of the flush in microbial activity upon rewetting and that microbial activity in previously dried soils may not be fully restored even after 54 days of moist incubation, suggesting that drying of soil can have a significant and long-lasting impact on microbial functioning.