Abstract: Despite the attention currently given to the potential environmental benefits of large-scale forest planting, there is a shortage of clear observational evidence regarding the effects on river flows, and what there is has often been contradictory or inconclusive. This paper presents three independently conducted paired-catchment forestry studies covering 66 station-years of flow measurements in the UK and Ireland. In each case coniferous evergreen trees were removed from one catchment with minimal soil disturbance while the adjoining control catchment was left unchanged. Trees were removed from 20% – 90% of the three experimental basins. Following woodland removal there was an increase in dry weather baseflow at all sites. Baseflows increased by about 8% after tree removal from a quarter of the Hore basin and by 41% for the near-total cut at Howan. But the changes were more complex for peak flows. Tree harvesting increased the smallest and most frequent peak storm flows, indicating that afforestation would lead to the suppression of such events. This was however restricted to events well below the mean annual flood, indicating that the impact of forests upon the largest and most damaging floods is likely to be limited. Whilst a forest cover can be effective in mitigating small and frequent stormflows it should never be assumed to provide protection against major flood events.