A nation-wide test on the completeness of birth registration was carried out in 1940 in connection with the Decennial Census of Population and Housing. The present paper discusses the comparison, based on this data, of two methods of obtaining per cent completeness of birth registration by states: 1) by relating a matched set of records for the state as a whole to the total group of matched and unmatched records combined; and 2) by a cumulative technique suggested by Chandra Sekar and Deming in the March issue of the Journal. Differences in the results by the two methods were minor except in those areas having a comparatively high degree of under-registration.