Statistical analysis of visual-field data is permitted by the quantitative techniques used in automated perimetry. The visual-field indices mean defect (MD) and corrected loss variance (CLV) provide an opportunity to analyze and follow visual fields mathematically. We investigated the relationship of MD to CLV in 32 glaucoma suspects and 17 patients with visual-field defects due to chronic open-angle glaucoma. In glaucoma patients with early and moderate disease (up to a MD of 18 dB), the CLV statistically covaried with the MD (R2 = 0.85). Beyond an MD of 18 dB, increases in the MD seemed to cause a decline in the CLV. These results differ from prior reports on the relationship of MD and CLV. Reasons for this and implications for the use of these statistics in the clinical management of the glaucoma patient are discussed. These results may be helpful in the understanding of the pathophysiologic basis of damage in glaucoma.