A new method to estimate the completeness of birth registration (1950 to 1970) and of enumeration of children and adolescents in recent censuses of the Soviet Union is based on the observation that primary school enrollment totals for 1959 and 1970 are virtually identical with the census counts of children in the relevant ages at these dates, and on the consequent assumption that such enrollment data therefore provide an accurate “census” for noncensus years. Projecting backwards and forward provides the basis for assessing completeness of birth registration and census enumeration of those under 25. Results suggest a four to five percent underregistration of births and some underenumeration of preschool children and, for 1970, of those aged 16-17 and 20-24.