THE PERFORMANCE OF QUANTITATIVE SCALE PATTERN-ANALYSIS IN THE IDENTIFICATION OF HATCHERY AND WILD STEELHEAD (ONCORHYNCHUS-MYKISS)

Scale pattern analysis techniques used to distinguish between hatchery and wild steelhead (Oncorhynchus mykiss) in high-seas mixtures have r equired subjective determination of freshwater age. We use scale measu rement data from adult hatchery and wild steelhead returning to the Co lumbia River in 1991 and 1992 to develop and evaluate the performance of a quantitative technique that is independent of age determination. The best performance was achieved using a six-variable, direct maximum likelihood estimator (point estimates averaged 94% in test samples co ntaining 100% hatchery or wild fish). Errors of the estimates varied d epending on the composition and size of the mixed sample. The maximum likelihood estimator (using pooled 1991-1992 Columbia River baseline d ata) correctly allocated a sample of wild steelhead returning to Kamch atka, Russia, in 1994 and was slightly biased (underestimated the cont ribution of hatchery fish by 5%) in allocating a mixed sample from the Gulf of Alaska in 1995. The inclusion of additional baseline data to account for a wider range of variation in scale growth patterns of Nor th Pacific populations may further reduce bias in the Gulf of Alaska e stimates. We conclude that this technique has good potential for estim ating proportions of hatchery and wild steelhead in high-seas mixtures .