Chuck Bronte, biologist at the Green Bay Fish and Wildlife Conservation Office, working with USGS researchers, had a paper accepted for publication in the Transactions of the American Fisheries Society, the oldest journal dedicated to fishery research.  Analysis of fish shape is used for stock identification, description of new species, and assessment of condition, but is frequently done on sample sizes smaller than those recommended in the literature.  The authors examined the stability of satistical outcomes of analysis of shape measures using different sample sizes from data sets of yellow and white perch from Lake Erie, and siscowet lake trout from Lake Superior.  The analysis suggested that stable outcomes occurred when the number of specimens was 3.5 to 8 times the number of shape measures taken--anything less would lead to erroneous conclusions regarding the shape differences among groups of fish studied. The results varied by species and led to recommendations that similar evaluations be conducted for other species.  Past work that used insufficient samples sizes may require re-evaluation of those results.  Authors, reviewers, and editors of future work should consider the effects of small sample sizes when conducting and evaluating fish morphology measures. The paper will be published in spring of 2009.