Hydrobiologia 506–509: 443–449, 2003. © 2003 Kluwer Academic Publishers. Printed in the Netherlands. 443 Comparison of gillnet and trawl in diurnal fish community sampling M. Olin 1,2,∗ & T. Malinen 1 1 University of Helsinki, Department of Limnology and Environment Protection, P.O. Box 65, FIN-00014 Helsinki, Finland E-mail: mikko.olin@helsinki.fi 2 Finnish Game and Fisheries Research Institute, P.O. Box 6, FIN-00721 Helsinki, Finland ( ∗ Author for correspondence) Key words: diurnal, fish community, Gillnet, sampling method, trawl Abstract The fish community in a shallow, eutrophic lake basin in southern Finland was sampled diurnally with gillnets and trawl. The differences in species number, relative abundances and length distributions were considered. The fish density estimations differed notably depending on the gear and diurnal period. The most abundant species in the trawl catch, smelt, was almost totally missing from the gillnet catch. The proportions of perch, roach, white bream and asp were higher in the gillnet catch. Gillnets regularly underestimated the proportion of small (<10 cm) individuals in size distributions. The trawl probably underestimated the density of piscivores. In the two gears, diurnal changes, in both fish catch and species distribution, were considerable. Both trawl and gillnets are needed to get more reliable figure of fish communities in shallow eutrophic lakes. Introduction To successfully manage fisheries or lake environment, one needs to know the properties of the whole fish assemblage: species number, abundance and size dis- tributions. To attain this, representative and reliable fish community samples are needed. However, none of the existing sampling methods give non-biased es- timates of true values (e.g. Bagenal, 1979). Gillnets, widely used in fish monitoring, are passive and se- lective (Hamley, 1975; Backiel & Welcomme, 1980; Kurkilahti, 1999). The probability of a fish to en- counter and retain in a net increases with swimming distance and speed, and discontinuities of body out- line (e.g. spines). The size distribution estimates are skewed because small individuals move less and when encounter the net are caught less effectively due to slower speed and lower flexibility of small mesh sizes. A trawl, as an active gear, should be less selective pro- ducing more reliable estimates of species abundance and length distributions. However, the trawl is not use- ful in small, shallow or rough-bottom waters (Backiel & Welcomme, 1980). In addition, trawl catchability can be relatively low for large individuals (Bethke et al., 1999; Hjellvik et al., 2001), which may avoid the trawl but not gillnets (Richardson, 1956). Further to gear problems, a fish community is un- der continuous spatial change even in the short run. Number of species have diurnal vertical or horizontal migrations due to feeding or predation avoidance (Bohl, 1980; Helfman, 1981). Fish activity changes diurnally (Helfman, 1981) and affects the encounter probability in passive gears. Given the shortcomings of gears and the dynamic nature of fish assemblages, reliable results are attained only by standardising gears and fishing time, and by combining different methods. Studies including sim- ultaneous diurnal fish sampling with different gears help us to understand when it is profitable to use a certain method, and how the methods can supplement each other. Such studies are, however, rare, as stated by Peltonen et al. (1999) and Pierce et al. (2001). The aim of our study is to compare fish community data collected at different time of day with two gears: gillnets and the trawl. We will consider the differences in species number, relative abundances, and length distributions. We hypothesise that due to the passive- ness of gillnets (1) the proportion of small individuals