Using samples of Atlantic cod Gadus morhua L. from the North Sea, and previously published genetic data from the Irish and Celtic Seas, Iceland, and Norwegian fjord and offshore populations, we describe striking macro-and micro-geographic patterns in pantophysin (Pan I) allele frequencies. The relatively abrupt discontinuity in Pan I allele frequency distribution at 2 different locations is not congruent with standard patterns of isolation by distance and could arise from population admixtures, historical or contemporary natural selection, behavioural segregation or a combination of these factors. Here, we examined the relationships between the distributions of Pan I alleles and temperature, salinity and depth. In the northeast Atlantic, temperature was highly correlated with Pan I allele frequency, even when the effect of geographic distance was removed. In the Norwegian fjords, partial Mantel tests indicated that temperature, salinity and depth all had a significant effect on Pan I allele frequency in juvenile fish. However, a sample from the brackish waters of the eastern Baltic Sea suggested that salinity may be linked to Pan I allele frequency distribution and that the relationship with temperature was weaker in areas of low salinity. Strong correlations between Pan I allele frequencies and key environmental variables, together with evidence from the available literature, suggested that environmental conditions play an important role in determining the distribution of different Pan I genotypes. The combined use of environmental data, Pan I genotyping and neutral markers may provide a valuable approach to examine local adaptation, levels of gene flow and stock structuring.