Computer assisted perimetry has revolutionised the investigation of the visual field. Experience of central field assessment with the Octopus Automated Perimeter shows that sensitivity recorded with target size 3 across all age groups can frequently be greater than the published normative values. Use of the latter values can therefore provide a serious underestimation of field loss. Inter-individual variation in sensitivity is found within and between age groups. The limitations associated with the use of the measurement error to define abnormality and the additional problems of hypernormal thresholds and resolution of the blind spot are discussed. It is suggested that methods should be developed to evaluate sensitivity on an intra-individual basis.