1. Monitoring at fortnightly to monthly intervals of a very shallow, lowland lake over 24 years has enabled the time course of recovery from nutrient enrichment to be investigated after high external P loading of the lake (>10 g P m )2 year )1 ) was reduced between 1977 and 1980. 2. The lake showed a relatively rapid response during the spring and early summer, with a reduction in phytoplankton biomass occurring after 5 years when soluble reactive phosphorus concentration was <10 lg L )1 . 3. However, during the later summer the response was delayed for 15 years because of sustained remobilisation of phosphorus from the sediment. The greater water clarity in spring and a gradual shift from planktonic to benthic algal growth may be related to the reduction in internal loading after 15 years. 4. Changes in the phytoplankton community composition were also observed. Centric diatoms became less dominant in the spring, and the summer cyanobacteria populations originally dominated by non-heterocystous species (Limnothrix/Planktothrix spp.) almost disappeared. Heterocystous species (Anabaena spp. and Aphanizomenon flos-aquae) were slower to decline, but after 20 years the phytoplankton community was no longer dominated by cyanobacteria. 5. There were no substantial changes in food web structure following re-oligotrophication. Total zooplankton biomass decreased but body size of Daphnia hyalina, the largest zooplankton species in the lake, remained unchanged, suggesting that the fish population remained dominated by planktivorous species. 6. Macrophyte growth was still largely absent after 20 years, although during the spring water clarity may have become sufficient for macrophytes to re-establish.

1. This paper advocates a catchment‐scale perspective for river restoration and for individual rehabilitation works even though, at present, such works are often small‐scale and ad hoc in nature. The catchment‐scale approach is the logical consequence of the application of fundamental principles of river science to the philosophy of river restoration. 2. The five principles that river restoration should incorporate are: i) the hierarchy of river systems; ii) the proportional relationships between discharge and channel dimensions; iii) the importance of the physical and biological continua of natural rivers; iv) the four‐dimensional nature of river systems; and v) the role of heterogeneity in maintaining biodiversity and that of disturbance in maintaining heterogeneity. 3. Two sets of examples are given to illustrate the need for these five principles. One set relates to the role of trees in river processes (and in river restoration), while the other relates to the rehabilitation of physical structures in rivers at a different scale. 4. The value of trees both for the maintenance of water quality and for their conservation value in upper‐order rivers is demonstrated. This, combined with evidence elsewhere for trees as an integral part of the river‐riparian ecotone, suggests the restoration of lowland headwater streams as being totally tree‐influenced. In middle and lower reaches of rivers, it is important to restore the river–floodplain interactions. Continuity with groundwater through the hyporheos as a result of riffle‐pool rehabilitation and flood‐regeneration of meadows and alluvial forests should be the long‐term vision for lower‐reach restoration. 5. In the interim, piecemeal rehabilitation of the physical heterogeneity of the bed and banks in large rivers can be locally successful provided that it restores coarse particles and interstices where they are absent through artificial material and/or siltation, but not if it fails to recognize fully the importance of replacing lost heterogeneity. Copyright © 1999 John Wiley & Sons, Ltd.

No abstract

No abstract