Negative binomial distributions were fitted for ten sets of complete trip creel survey data from the New and Shenandoah Rivers in Virginia and West Virginia. The fits for eight of ten datasets were not significantly different from the observed distributions (P< 0.05). The negative binomial provides a good alternative when complete trip data are unavailable or unreliable; however it was still more accurate to use the observed frequency distributions.

An age-structured, non-linear population model employing instantaneous mortality rates was developed to compare the merit of various regulation schemes. Model outputs included parameters commonly measured in recreational fisheries management, including structural indices, catch, harvest, and yield. Eight creel limits (1-8 fish/angler-day) and ten length limits (six minimum limits, three slot limits, and no length limit) were simulated in all possible combinations. A catch-and-release scenario was also simulated, increasing the total number of simulations to 81. As a case study, data were used from the smallmouth bass fishery in a pool of the New River in West Virginia. Principal components analysis revealed that 91% of the total variation in the output variables among the 81 simulations was represented by the first three components. A set of seven variables maximized the differences among simulations: harvest (numbers), size harvested, catch (numbers), size caught, proportional stock density, relative stock density of trophy fish, and yield (weight). Model results for these individual variables revealed that, while harvest and yield were maximized by liberal scenarios, all other variables were improved by the more restrictive regulations. The results also indicate that minimum length limits have a much greater impact than creel limit reductions.