Abstract
Diamond-mesh codends are the most commonly used in demersal trawls. However, mesh geometry tends to vary in these codends during fishing, which leads to a less well-defined size selection process. This leaves one questioning the rationality of regulating exploitation patterns based on mesh size when size selection and/or variation between hauls is highly variable. While it has been speculated and theoretically investigated how much the variability in mesh geometry may contribute to the variability in size selection, it remained to be quantified experimentally. Therefore, we compared the size selectivity of a simple diamond-mesh codend, where meshes are subjected to variation in geometry, with a rigid diamond-mesh codend, where the geometry of the meshes were kept constant. For Atlantic cod (Gadus morhua) the simple diamond-mesh codend was found to have 45% more variation in size selection than the codend with constant mesh geometry. This confirms theoretical predictions and may guide research towards codend designs with more well-defined size selection properties.