“…When fish products have been studied, phenotyping of LAB has resulted in unreliable species identification. LAB originating from vacuum-packaged, salted, smoked herring gave atypical profiles (Gancel et al, 1997), and in a study of the spoilage flora of vacuum-packaged, salted, cold-smoked rainbow trout fillets (Lyhs et al, 1998), we also concluded that phenotyping is not always reliable for Several molecular methods have been used for strain typing or characterisation of LAB (Hill and Hill, 1986;Tannock et al, 1990;Welsh and McClelland, 1990;Williams et al, 1990;Duffner et al, 1993;Dykes et al, 1993;Jaho et al, 1993;Roussel et al, 1993;Björkroth et al, 1996aBjörkroth et al, , 1996b, but only some of these applications have been used for establishing species identification libraries. Paludan-Müller et al (1998), studying the role of Carnobacterium species in the spoilage of vacuum-packaged, cold-smoked salmon, used whole-cell-protein patterns for identification of the carnobacteria.…”