Key points are not available for this paper at this time.
Two estimates of trout populations were obtained in a large stream, the St. Mary River (mean annual flow, 54 m 3/s) in southeastern British Columbia. Underwater counts of trout by seven swimmers were stratified by bank and midchannel zones, and expanded. Mean expanded counts were compared to mean Petersen (mark-recapture) estimates that also were obtained by underwater observation of fish marked during angling. The expanded count from the underwater enumeration underestimated the population of cutthroat trout (Salmo clarki lewisi over 20 cm in length by 26%, thus requiring a correction factor of 1.35 to adjust the expanded count. Species composition and estimated size distributions of observed fish were similar to those in a sample captured by angling. Replicated counts of trout over 30 cm in length were homogeneous, but counts of 10-30-cm trout were heterogeneous. Expanded underwater counts obtained from as few as two swimmers were reliable if they were replicated, conducted in both representative nearshore and midchannel zones, and multiplied by a factor of 1.35.
Slaney et al. (Thu,) studied this question.