• pink salmon;
  • electronic noses;
  • seafood quality;
  • headspace analysis


Volatiles in canned pink salmon, produced from different degrees of skin watermarked raw material and stored for 2 and 9 mo, were characterized and compared using static headspace gas chromatography analysis coupled to a mass spectrometer (SHGCMS). Sulfur-containing compounds comprised 30% to 50% of the total volatiles and tended to decrease with increasing degrees of skin watermarking, and dimethyl sulfide was the most abundant compound of this class of molecules. A few alcohols, aldehydes, ketones, and furans were also identified. Forward stepwise general discriminant analysis (FSGDA) was used to investigate prediction models based on degree of skin watermarking. The 2- and 9-mo models using SHGCMS showed 92.5% and 93.75% correct classifications, respectively. The ability of the Cyranose 320, a hand-held electronic nose (EN), to differentiate these grades of watermarking in the canned samples was also tested. EN analysis using FSGDA resulted in models with 90% and 92.5% correct classifications for the 2- and 9-mo samples, respectively. Overall, results indicate that the watermarking grades studied are not readily distinguishable from each other by either method of analysis.