Despite these general results, the importanceofgene
flow among several widelydistributed marine fishes remains controversial. In the case of the Atlantic cod,Gadus morhua, the controversy arises from the ambiguous results provided by different types ofgenetic markers. Highly significant differences have been observed
between Baltic and North Sea populations and betweenNorwegian coastal and arctic stocks at blood protein loci such as hemoglobin-1 (Hbl) and transferrin ( T j )(FRYDENBERGet al. 1965; SICK1965; MOLLER 1968;DAHLEandJORSTAD 1993), but not among populations sampled throughout the entire species' range and scored for a larger set of routine electrophoretic loci(MOM et al. 1985). Theheterogeneity observed among protein-coding loci has been interpreted as reflecting
the action of selection at a subset of loci, but this is incompatible with the continuing action of gene flow suggested by tagging studies (e.g., RASMUSSEN 1959;GULLANDand WILLIAMSON1962; TEMPLEMAN1974),which should eliminate differences among populations unless selection is particularly strong. Recent studies examining patternsof mtDNAvariation in the Atlantic
cod have corroborated the electrophoretic andtagging results in revealing limited population differences(SMITHet al. 1989;CARRand MARSHALL1991;ARNASON and RAND 1992). Here again,however, exceptions have been reported (DAHLE1991). Thelow levels of mtDNA polymorphism observed in the Atlantic cod relative to other marine fish species has suggested the possibility that its limited population structuremay reflect a recent origin of populations, possibly involvinga recentbottleneck (see SMITHet al. 1989; ARNASON et d .1992). Determining whether gene flow has recently ceased among populationsor is continuing atlow levels is problematic, however, since both would produce patterns of high genetic similarity