Abstract
Longfin eels Anguilla dieffenbachii caught by electrofishing, in Lee Stream, a tributary of the Taieri River, New Zealand, were significantly longer in pools than in riffles, particularly in summer. Longfin eels were generally shorter in sites with willow. More longfin eels were caught at pasture sites than tussock sites with fewest at sites with willow. Sites judged difficult to access by commercial fishermen had longfin eels that were significantly longer than those in easy‐access sites. More than 80% of longfin eels were captured within 270 mm of banks. The first three components of a principal component analysis explained 40·3% of the variability in microhabitat character recorded at both the capture location and within 0·5 m of capture, and from random locations in each riparian category. Variables strongly associated with longfin eel capture locations were maximum depth (longfin eels used microhabitat of intermediate depth), distance to pool (intermediate), sediment size (finer), sediment depth (deeper), stream idth (intermediate) and flow velocity (slower). Overall there was a greater icrohabitat range in tussock sites but diurnal microhabitat use was most similar to what was available in pasture streams. Winter longfin eel locations were characterized by deeper and finer sediments in contrast to summer and spring locations.