Evaluation of Long-Term Retention and Detection of Oxytetracycline Marks in Walleye Otoliths Using Genetic Methodology

Because neither genetic population assignments nor oxytetracycline (OTC) mark identifications are 100% accurate, genetic methods were modeled alone and in combination with OTC mark observations to identify the source of walleyes Sander vitreus in a supplemented population and to evaluate the accuracy of OTC mark examination in distinguishing stocked fish that had been treated with OTC from those produced naturally in the lake. Readers examined otoliths from 3,663 fish from two year-classes that were sampled up to 5 years after stocking. A subset of 1,578 of the examined individuals were also assigned to source populations based on genotypes at eight microsatellite loci using baseline genetic data from the stocking source and the recipient population (the measure of genetic divergence [F(ST)] = 0.07 between populations before stocking). The models indicated that OTC mark observation errors occurred in all samples and that the error rates varied among samples and readers. Model estimates of overall false-negative error rates ranged from 11.4% to 17.2%, and estimates of overall false-positive error rates ranged from 8.9% to 16.7%. Over the ages of walleyes inspected (3 months to age 5), there was neither a systematic increase in mark errors with age nor a decrease in mark intensities; thus, the OTC marks present at 3 months appeared to be permanent. Any true mark absence was due to inefficient OTC immersion treatment or the loss of a successfully applied mark before the first sample was inspected as fingerlings. The presence of both false positives and false negatives in all samples indicates that mark detection error did consistently contribute to the overall mark error rates. Thus, changes in quality control are justified for both OTC immersion treatment and mark detection procedures.