We reviewed epidemiological and experimental studies of female gonadal hormone replacement therapy (HRT) on cognitive function in post-menopausal women and carried out meta-analyses. In healthy ageing women, HRT has small and inconsistent effects that include enhancement of verbal memory, abstract reasoning and information processing. Epidemiological studies show larger effects than experimental studies, which is not related to sample size. Important confounds may be that women who start using HRT are healthier than women who do not. Also, controlling for socio-economic status diminishes the effect of HRT. The effects of HRT may depend on the age and type of menopause and the therapeutic intervention used, with the most widely used drug, Premarin, having least effect. However, the effects are independent of mood and climacteric symptom alleviation. There is a paucity of experimental studies that include healthy elderly women. The evidence for an estrogen deficiency in women with dementia and cognitive dysfunction is inconsistent. Nevertheless, epidemiological studies suggest that HRT protects against the development of clinically diagnosed Alzheimer's disease. However, poor recall of HRT use by patients and altered physician behaviour may have confounded the effects. Surprisingly, both healthy and demented women with low education seem to benefit most from HRT. Three recent controlled experimental studies using Premarin showed no effects of HRT in preventing further cognitive decline in women who already have Alzheimer's disease. Duration of treatment seems to play an important role, with beneficial effects declining-and even reversing-with longer treatment in women with Alzheimer's disease.Future research should further investigate the cognitive effect of different HRT preparations, serum estrogen levels, and the interactions of HRT with age, menopausal status and existing protective (e.g. education) and risk factors (e.g. smoking and apolipoprotein E genotype) for cognitive decline and Alzheimer's disease.