Baited and unbaited hoop-nets commonly are used to capture catfish in lotic and lentic systems. Turtle bycatch and post-capture mortality has been problematic during catfish surveys in Missouri, USA, most recently in the Gasconade River, Gasconade and Osage counties. We evaluated 3 modified hoop-net designs that would reduce turtle bycatch without reducing catfish capture in the Gasconade River during 15 May–15 July 2006 after pilot study evaluation of 5 hoop-net designs in April 2006. We deployed modified and control-nets in blocks for 48 hours to evaluate differences in turtle and catfish catch rate, as well as abundance, size, and mortality rate of turtle bycatch. The chimney design reduced turtle bycatch by 84% when compared to the control, without decreasing the number or average size of captured flathead catfish (Pylodictis olivaris). Environmental conditions that affected turtle mortality included Secchi disc transparency, temperature, dissolved oxygen, and stream river depth. This is the first known attempt to create turtle exclusion or escapement devices for hoop-nets deployed in freshwater systems. Biologists using hoop-nets to sample aquatic vertebrates in moderate to large river systems will benefit from our study. The application of this methodology will reduce turtle bycatch mortality, especially when sampling is conducted in high water temperatures.