Two techniques of stock discrimination both utilizing scales, i. e. termed as classical technique and shape analysis, are compared for discriminating between North American and European Atlantic salmon (Salmo salar L.) caught off West Greenland. Three scales from each of 60 salmon were analyzed: 30 from North American and 30 from European-origin salmon collected from home rivers in 1982. The best results indicated misclassification and error rates of 0% for shape analysis and 5.0% and ± 1.7%, respectively, for the classical technique. A nested ANOVA indicated that variation was greater between scales of salmon of North American and European origin than from several scales from the same salmon. It was concluded that shape analysis with computer-aided devices may be superior to the classical technique and it is recommended that it be examined for use at West Greenland.
