Imprecision in estimates of abundance and population parameters obtained from trawl surveys could be reduced if the sources of variability were understood and properly described. Fish behaviour in relation to sampling gears has been widely studied, with the main goal of providing information which might lead to improved reliability of abundance estimates. However, the general applicability of the results may in many cases be questionable due to unidentified effects of natural behaviour of the fish on catchability. Variability of natural behaviour may affect catchability directly and, therefore, if aspects of natural behaviour are taken into account in design and analysis of bottom trawl surveys, the reliability of the information obtained from these surveys could be improved.