The purpose of this study was to assess objectively the effects of Percoll separation on human sperm head morphometry. Semen samples were washed and smears were prepared on slides. An aliquot of each sample was centrifuged on a Percoll gradient and spermatozoa were prepared on slides. Dimensions of sperm heads from each sample were assessed by computer-aided sperm head morphometry analysis and manual sperm morphology was assessed for each sample. The percentage of normal sperm heads and morphometric dimensions from washed and post-Percoll separated samples were compared across all men by a paired t-test. Correlations between normal sperm head morphometry and manual sperm morphology were assessed in washed and Percoll-separated samples. The percentage of normal sperm head morphometry was significantly(p < 0.001) higher in Percoll-separated samples than in washed samples (23.6 vs. 12.6%). No differences (p > 0.1) in mean sperm head measurements were detected between washed samples and Percoll-separated samples. Coefficients of variation for mean sperm head measurements were significantly lower in Percoll-separated samples. No correlation (p > 0.1) in percentage normal was found between computer-assisted sperm head morphometry and manual morphology for washed and post-Percoll samples. These results indicate that percentage normal sperm head morphometry is increased by Percoll separation. While sperm head dimensions were unchanged, sample variability was decreased.