Family history is perceived to be an important risk factor for stroke despite conflicting published data. We examined patterns of familial aggregation of stroke among three generations using data from the Framingham Study. Cox proportional hazards analyses, adjusting for known stroke risk factors, were used to examine familial concordance in three groups: (1) members of the original Framingham cohort using reported parental stroke death; (2) members of the Framingham Offspring Study and their parents (members of the original Framingham Study); and (3) sibships within the original Framingham cohort. We found no association between stroke or transient ischemic attack among original cohort members and their reported parental stroke death (n = 4933; relative risk [RR] = 1.07). Using verified cases of parental stroke or transient ischemic attack, the Offspring analyses revealed that both paternal (n = 1762; RR = 2.4; 95% confidence interval [CI], 0.96 to 6.03) and maternal (n = 2074; RR = 1.4; 95% CI, 0.60 to 3.25) histories were associated with an increased risk. Parental history of coronary heart disease was strongly associated with stroke or transient ischemic attack among Offspring Study members (RR = 3.33; 95% CI, 1.27 to 8.72). Sibling history of stroke or transient ischemic attack was not associated with stroke or transient ischemic attack among original cohort members, although a non-statistically significant increased risk associated with sibling history of atherothrombotic brain infarction was observed (RR = 1.8; 95% CI, 0.68 to 4.94). These analyses suggest that parental history of stroke may be a risk factor for stroke. As more stroke or transient ischemic attack events develop among the Offspring Study members, it will be valuable to reexamine these associations.