The question of Census undercount has recently gained considerable attention. The 1990 Post-Enumeration Survey (PES) constituted the major vehicle for measuring coverage by area of the 1990 Decennial Census. It was designed to be used to adjust the Census enumeration. This article discusses the background of the survey, the sampling plan, the methods used to measure Census omissions and Census erroneous enumerations, the treatment of nonresponse, the use of dual-system estimation to estimate the total population by post-strata, and the use of these estimates to calculate adjusted Census data.