Three manipulative experiments were done to estimate the selectivity of conventional and new sizes and configurations of mesh for school prawns, Metapenaeus macleayi, in three south-eastern Australian recreational fishing gears (haul, push, and scoop nets). The treatment meshes examined were: (1) conventional-sized, diamond-shaped mesh used in all gears (20 mm in scoop nets and 30 mm in push and haul nets); (2) 30 mm in scoop nets; and (3) 40-mm diamond-and (4) quantities of school prawns (6-24 mm carapace length (CL)) were placed in purpose-built enclosures, monitored to ensure no experimental-induced stress (as measured by changes in L-lactate in their haemolymph) and the replicate treatments of the various gear configurations deployed. Escapees from the various treatment nets were collected from the enclosures using fine-meshed nets. Logistic selection curves were derived for all treatment nets and specific comparisons made within and among gears. All nets had 50% retention lengths (L 50 ) comparable to other penaeid-catching gears with similar mesh sizes, but most had selection ranges (SRs) that were atypically inflated. The large SRs were attributed to a combination of factors that included the mesh geometry and towing speed of the gears and the behaviour of school prawns. The 20-mm scoop net had the smallest selection parameters, retaining >99% of individuals larger than 13 mm CL. Mesh size in this gear would need to be increased to at least 30 mm to allow some maturing prawns (>18 mm CL) to escape.