Sundowning refers to episodes of agitated behaviour that are more frequent or are more severe at night. Although the effects of different psychoactive medications on agitated behaviour in dementia patients have been documented in hundreds of reports over the last 30 years, less than 20 studies make explicit reference to time of day for which outcome measures were derived, and even fewer have also examined sleep as an outcome. Thus, despite varying claims of efficacy and effectiveness for various medications, there are few data to support informed management of disruptive nocturnal behaviour in these patients. In this brief article, we selectively review those few studies explicitly mentioning temporal dimensions of behavioural outcome, including some newer studies of unconventional types of treatment that may be useful for the treatment of sundowning. We conclude that future pharmacological studies should systematically assess behaviour throughout the 24-hour day to provide outcome data relevant to this phenomenon.