A lightweight sorting grid was developed to reduce bycatch in the Danish small-meshedtrawl fishery (22 mm full mesh in the cod end) for Norway pout in the North Sea.Experimental fishing with the grid demonstrated the possibility to capture Norway poutwith only a minimum of unintended bycatch. Fishing with two different grid orientations,backwards and forwards-leaning, in distinct day and night hauls, resulted in an estimatedrelease of between 88.4 and 100% of the total number of haddock (Melanogrammusaeglefinus) and whiting (Merlangius merlangus) entering thetrawl. However, bycatch reductions were not significantly different between day and nightor between grid orientations, indicating that the grid rejection of haddock and whiting isnot influenced by fish behaviour. The loss of the target species, Norway pout, was low(between 5.6% and 13.7%) in comparison with the bycatch excluded, and clearly lengthdependent. Consequently, loss of target species would vary with the size structure of thepopulation fished. Although results were not statistically significant, length-basedanalyses indicated that the grid rejection likelihood for particularly smaller Norway pout(<16 cm) was higher when fishing with the forwards-leaning grid during the night;this might be explained by behavioural and visual aspects of the fish-grid encounterprocess for Norway pout.

In this paper we consider three measures of overlap, namely Matusia's measure ρ, Morisita's measure λ andWeitzman's measure Δ. These measures are usually used inquantitative ecology and stress-strength models of reliabilityanalysis. Herein we consider two Weibull distributions havingthe same shape parameter and different scale parameters. Thisdistribution is known to be the most flexible life distributionmodel with two parameters. Monte Carlo evaluations are used tostudy the bias and precision of some estimators of these overlapmeasures. Confidence intervals for the measures are alsoconstructed via bootstrap methods and Taylor series approximation.

For the infinite time ruin probability in the classical risk process, efficient estimators are proposed in cases in which the claim amount distribution is unknown. Confidence intervals are computed which are based on normal approximations or on the bootstrap method. The procedures are checked in a Monte-Carlo study.