Threadfin shad Dorosoma petenense populations were sampled quarterly from 2010 to 2011 to determine appropriate sampling techniques for this species in tropical reservoirs of Puerto Rico. Offshore gill netting and night trawling were compared in terms of catch per unit of effort, size distribution, sampling precision, and bycatch. In total, 90 gill net–trawl pairs of catches were compared, which collected more than 80,000 threadfin shad. Gill-net and trawl catches were not correlated in either number or biomass. Coefficients of variation were greater in gill-net sampling (0.761 for numbers, 0.747 for biomass) than in trawl sampling (0.433 and 0.465, respectively) and were not dependent on reservoir, sampling season, reservoir section, or any combination of these factors for any gear. There was no correspondence in size distributions between gill-net and trawl catches. Gill nets collected threadfin shad in the range of 23–169 mm total length (TL) and displayed strong mesh size selectivity, causing distinct unrealistic peaks in size distribution. Gill nets underestimated threadfin shad smaller than 35 mm TL and overestimated shad bigger than 90 mm TL. The size range of threadfin shad collected via trawl was 10–108 mm TL, and trawls did not appear to be as size selective for fish up to 80 mm TL. Trawling was found to be the superior sampling technique for shad populations in pelagic habitats of Puerto Rican reservoirs, which consisted mainly of fish up to 80 mm TL. Trawling provided a more realistic picture of size distributions, collected far less bycatch, was less affected by the schooling of shad, and was less laborious and more cost-effective than gill nets.
Keywords: Dorosoma petenense; threadfin shad; trawl