Longitudinal chromatic aberration (LCA) of the human eye has been studied repeatedly, but only at the fovea. Poor visual acuity prevents its subjective determination beyond a few degrees eccentricity. Consequently, we have used an objective approach, similar to that of Charman and Jennings [(1976). Vision Research 16, 999-1005], to measure ocular LCA across the visual field. To determine the validity of our double-pass approach, a direct comparison between objective and subjective results was established where possible, namely at the fovea and parafoveally (2.5 deg). In both cases we focused a monochromatic point source at four different wavelengths (458, 501.8, 543.5 and 632.8 nm). At the fovea, for a 3 mm pupil, we found a close match between subjective and objective results. However, as the subjective task became harder (off-axis or larger pupils), subjective results tended to yield slightly more myopic eyes than the results for objective refraction. In all cases, the offset was virtually independent of the wavelength used. Therefore, we have not found evidence of any biased estimates of the LCA, as determined objectively. Our foveal results show reasonable agreement with previous findings, except for slightly smaller amounts of LCA. Starting at the fovea, LCA tends to gradually increase with eccentricity, up to 40 deg, although such an increase is small, just approaching statistical significance. Computation of the LCA using a model eye predicts a slightly smaller increase with eccentricity.