A soil disturbance experiment was performed during two seasons in degraded, calcareous sandy grassland in southern Sweden. pH, extractable phosphorus, plant species richness and vegetation composition were analyzed. Mechanical soil disturbance had no effect on pH, and caused only a minor increase in extractable phosphorus. Positive effects compared to control plots were seen on plot scale (360 m2) in species richness and Shannon index 2 years after treatment, while on a smaller scale (1 m2), species richness increased only in rotavated plots. Plant community differences were mainly found between disturbed and control plots, and some positive effects were detected on early establishing species, but the two disturbance techniques favored different species. It is concluded that mechanical soil disturbance, and a variation in techniques and intensities, may be an important measure for preserving diversity in grasslands that have a history of mechanical soil disturbance.