Current views on the relationship between dementia and cerebrovascular disease are in a state of flux. Even our labels are changing vascular dementia rather than multi-infarct dementia is now the preferred term. The NINDS-AIREN international workshop (Neurology, 43 (1993) 250-260), in proposing research criteria for the diagnosis of vascular dementia, recognized the heterogeneity in mechanisms, syndromes, and course of dementia from cerebrovascular disease. There has been a tendency to consider vascular dementia as a mimic or mask of the most common primary dementia, Alzheimer's disease, in which cellular failure leads to progressive mental decline. A currently evolving thesis is that cerebrovascular disease may also be the cause of a primary dementia, in which supply failure leads to progressive mental decline, resurrecting older views on the role of arteriosclerosis and its relationship to ischemic changes (now termed leuko-araiosis ) regularly seen in the cerebral white matter of aging individuals. In this review, the concept of a progressive dementia due to supply failure but without clinically-obvious strokes will be considered. Since direct and definitive data for this thesis are currently lacking, indirect evidence from studies of post-stroke dementia will be examined to argue for the biological plausibility of this concept.