Journal of Fish Biology (2010) 76, 1751–1769 doi:10.1111/j.1095-8649.2010.02614.x, available online at www.interscience.wiley.com Effects of discharge and local density on the growth of juvenile Atlantic salmon Salmo salar M. A. K. Teichert*, E. Kvingedal, T. Forseth, O. Ugedal and A. G. Finstad Norwegian Institute for Nature Research, Tungasletta 2, N-7485 Trondheim, Norway (Received 12 December 2008, Accepted 5 February 2010) The study explored the combined effects of density, physical habitat and different discharge levels on the growth of juvenile Atlantic salmon Salmo salar in artificial streams, by manipulating flow during both summer and winter conditions. Growth was high during all four summer trials and increased linearly with discharge and mean velocity. Differences in fish densities (fish m −3 ) due to differences in stream volume explained a similar proportion of the variation in mean growth among discharge treatments. Within streams, the fish aggregated in areas of larger sediment size, where shelters were probably abundant, while growth decreased with increasing densities. Fish appeared to favour the availability of shelter over maximization of growth. Mean growth was negative during all winter trials and did not vary among discharge treatments. These results suggest that increased fish densities are a major cause of reduced summer growth at low discharge, and that habitat-mediated density differences explain the majority of the growth variation across habitat conditions both during summer and winter. © 2010 The Authors Journal compilation © 2010 The Fisheries Society of the British Isles Key words: density dependence; habitat; salmonids; seasonal variation; substratum. INTRODUCTION Variation in the annual regime of water discharge affects growth and survival of stream salmonids (Elliott et al., 1997; Jensen & Johnsen, 1999; Nislow et al., 2004; Harvey et al., 2006). In view of ongoing modification of natural hydrological regimes (Poff et al., 1997; Maddock, 1999; Gordon et al., 2008) and the resulting habitat degradation (Giller, 2005), it is necessary to understand the effect of hydrological change on fish performance, to effectively conserve and restore native fish popula- tions (Nislow et al., 2004). Until recently, effects of changes in hydrological regime on fish production have predominately been assessed by empirical correlation between habitat variables and fish distributions in the wild (Bovee, 1986; Fausch, 1988). Availability of suit- able habitat at different discharges is subsequently predicted from the resulting in situ habitat preference curves (Heggenes et al., 1991, 1996) and physical habitat *Author to whom correspondence should be addressed. Tel.: +47 7380 1463; fax: +47 7380 1401; email: maxim.teichert@nina.no 1751 © 2010 The Authors Journal compilation © 2010 The Fisheries Society of the British Isles