The Rivermead Behavioural Memory Test for Children (RBMT-C) was used to assess a population-based sample of 74 people with Down syndrome (DS) aged 30 years and over (range 30-65 years) living in the Cambridge health district. Given the high age-specific risk of Alzheimer-like neuropathology in people with DS, the aim of this study was to determine whether this test provided a suitable measure of memory function in an older population of adults with DS who were at the age of risk for Alzheimer's disease. We also investigated whether there was a significant difference in everyday memory performance between different age, gender and day-centre groups. The level of learning disability in our sample ranged from mild, through severe to profound. Individuals with severe or profound learning disability were virtually untestable on this and other cognitive tests, but for the remaining two-thirds of the group, the RBMT-C could be used to assess memory function. Significant differences in memory function were found between younger (30-44 years) and older (45+ years) participants. Performance also varied between groups (e.g. from different day centres), illustrating the importance of population-based samples when determining the value of such tests. Ongoing longitudinal studies are required to establish the extent to which the RBMT-C is useful in assessing change in retrospective and prospective memory.