Up to now, relatively few bio-optical measurements have been made in the high northern latitude waters, which allow sound relationships for ocean color remote sensing to be determined. We collected optical and chlorophyll a concentration, Chl, data in the north polar region of the Atlantic in summer season. The investigated region includes subarctic and arctic waters between 70¿N and 80¿N within the meridional zone between 1¿E and 20¿E. Our measurements show that the current NASA global algorithms, OC2, OC4, and chlor-MODIS, generally overpredict Chl in the investigated waters by a factor of about 2 at low pigment concentrations (<0.2 mg m-3) and underpredict Chl at higher concentrations (20--50% at 2--3 mg m-3). For our data set, the best two-band algorithm for Chl involves the ratio of remote-sensing reflectance, Rrs(442)/Rrs(555), at 442-nm and 555-nm light wavebands. We found that the general trend of variation in the blue-to-green reflectance ratio, Rrs(442)/Rrs(555) or Rrs(490)/Rrs(555), with Chl was driven primarily by Chl-dependent change in the green-to-blue ratio of absorption by pure seawater and particles. The effect of the blue-to-green backscattering ratio was of secondary importance. We observed a characteristic optical differentiation of waters within the investigated region. The majority of waters, which are here hypothesized to be dominated by diatoms, exhibited a relatively high blue-to-green reflectance ratio. The waters at several other stations, presumably dominated by dinoflagellates and/or prymnesiophytes, showed much lower reflectance ratio. Our data also show that the seemingly random variations in particulate absorption and backscattering coefficients at any given Chl are significant (more than a factor of 2) in the investigated waters.