Patients with Korsakoff's syndrome were evaluated with nine tests of new learning ability and three tests of remote memory to determine the correlation between anterograde amnesia and remote memory impairment. There was no correlation between the severity of anterograde amnesia and either the overall severity of remote memory impairment (1940s-1970s) or the impairment observed for more remote time periods (1940s-1950s). However, the correlation between remote memory impairment and anterograde amnesia became progressively stronger with the recency of the time period and was significant for the 1960s-1970s or the 1970s alone. The results support the view that the extensive remote memory impairment in Korsakoff's syndrome is, at least in part, distinct from and unrelated to anterograde amnesia. The more severe impairment observed for more recent time periods could be related to anterograde amnesia in that it reflects anterograde amnesia that was either already in place or progressively developing during recent years. The ability to recall very remote events seems therefore to depend on brain mechanisms distinct from those required for new learning or for recall of more recent events.