SPECIES DIFFERENTIATION BY MULTIVARIATE-ANALYSIS OF PHOSPHOLIPIDS FROM CANNED ATLANTIC TUNA

Multivariate statistical analyses have been applied to different param eters regarding phospholipid classes with the aim of differentiating b etween three species of canned Atlantic tuna: albacore (Thunnus alalun ga), bonito (Sarda sarda), and big eye tuna (Thunnus obesus). Using fo rward discriminant function analysis to generate classification functi ons, 14 variables were chosen, showing the content of the 20:4n-6 fatt y acid in the phospholipid fraction, the highest value of Wilk's lambd a. An excellent percentage of right classification in the three specie s groupings was obtained. The efficiency of the generated functions wa s tested using a set of canned tuna samples. Commercial albacore and b onito samples, samples caught during different years, and overprocesse d samples were successfully identified employing this technique.