The two experimental net designs were tested against a standard design using the paired tow method (Wileman et al. 1996; DeAlteris et al. 1999). Under this protocol, tows were repeated as similarly as possible in pairs, with one tow using an experimental net and the other a standard net. When possible, tow position, tow duration, and tides were duplicated, with one leg of the pair immediately following the other leg.

To allow simultaneous testing on two vessels, two standard nets were constructed of 15.2 cm 3 mm diamond polyethylene (PE) mesh throughout, with 50 mesh diameter 16.5-cm square knotless mesh in the codend, 20 meshes wide on the top and bottom. The headrope and footrope lengths were 12.2 and 18.3 meters (Figure 1). The fishing circle in the standard nets was 140 meshes. The Ribas (Figure 2) and topless (Figure 3) experimental nets were also constructed of 15.2-cm PE mesh with 75 mesh diameter 16.5-cm square mesh in the codend, 25 meshes wide on the top and bottom. The Ribas net replaced 15.2-cm diamond mesh on the top middle of the net with 25 meshes wide of 20.3-cm square mesh from the headrope to the codend. This design had a headrope and footrope length of 18.3 m. The two experimental nets had fishing circles with 100 meshes. All three codends had chaffing gear on the bottom half, consisting of unbraided strands of PE twine.

The topless net has no top wings, allowing the headrope to follow a taper of the net's gore into the top belly, reaching a length of 27.1 m. The bottom half of this net and the Ribas net are identical. Two sweeps were used during this study; a sweep with 15.2 cm cookies and a sweep with 10.2 cm cookies was used to increase flatfish catch rates.

Catch sampling was conducted by trained observers following protocols established by NMFS and modified by DMF for this study. For each tow, position, time, duration, weather conditions, catch composition and weights of all species, and length frequencies of target species were recorded. In cases of large volumes, subsampling of catch and extrapolation were performed using standard DMF protocols.

Underwater video of cod and flatfish behavior was acquired with a third-wire pan-and-tilt system developed by DMF. The camera was mounted on the headrope of the experimental nets. This system allows for viewing of the net while it is being towed. The pan-and-tilt unit has a 360-degree range of motion, allowing an ondeck operator to point the camera in a useful direction. Performance of the modified gear was viewed by the operator during filming and again later by DMF personnel at least once. Cod and flatfish reactions were logged and classified into two basic categories: fish that escaped and fish that are caught in the net. Flatfish species could not be distinguished due to inadequate video resolution and were grouped together.

cod_redux_catch.csv