Most developing countries do not have complete registration of deaths on which to base mortality estimates. Four broad categories of unconventional methods have been developed to provide mortality estimates in such settings. The first consists of approaches for evaluation adjustment of incompletely recorded deaths by comparison with recorded age distributions. The second consists of alternative data collection methodologies collecting information about deaths by age. The third consists of approaches based on asking respondents about the survival or otherwise of close relatives. The fourth estimates mortality from changes in age distributions, interpreting cohort attrition as mortality. Methods in the first two categories offer the greatest potential for contributing information on developing country mortality to the Human Mortality Database. Methods in the first category are illustrated here by application to data from the Republic of Korea for the second half of the 20th century. In populations with good age reporting and little net migration, these methods work well and offer the opportunity to include developing country data in the HMD. © 2005 Max-Planck-Gesellschaft.