A cross sectional, prevalence survey of eye disease in the population over 75 years old of Melton Mowbray has been used to examine the accuracy and completeness of the Blind and Partially Sighted Registers. The Blind Register had high sensitivity and specificity but was found to underestimate the prevalence of blindness by a factor of 1.1. The Partially Sighted Register had high specificity, but the sensitivity was only 50% and it underestimated the prevalence of partial sight by a factor of 1.5. Seven persons eligible for registration, but previously not registered, were found, two as blind and five as partially sighted. This represented 21% of the registrable visually impaired population.