The present paper describes findings from a mixed-method, longitudinal cohort study into perinatal depression among black Caribbean women. Using symptom scores from the Edinburgh Postnatal Depression Scale (EPDS), 101 black Caribbean and 200 white British women accessing maternity services in Manchester, UK, were screened for depressive symptoms in the last trimester of pregnancy and 6 weeks following delivery. The purpose of the study, which was undertaken between February 2000 and February 2001, was to estimate the prevalence of depressive symptoms during and after pregnancy among black Caribbean women compared to white British women in the same geographical area, and to explore black Caribbean women's beliefs about perinatal depression and their attitudes to help-seeking. Despite higher levels of self-reported risk, black Caribbean women were less likely than white British women to score above threshold (EPDS > or = 12) during pregnancy (chi2 = 4.16, d.f. = 1, P = 0.041). Although equally likely to score above threshold postnatally, they were less likely to receive treatment (chi2 = 4.20, d.f. = 1, P = 0.040) and more likely to be referred to secondary care (Fisher's Exact Test, P = 0.049). Qualitative findings suggest important differences between black Caribbean women's beliefs, attitudes and associated help-seeking practices, and those previously reported. Lower rates of depressive symptoms might partly be accounted for by conceptualisations of mental illness which differ from those of white British and South Asian women, and mistrust of the mental health services. This study showed that black Caribbean ethnicity is an important dimension in understanding the social patterning of mental illness. The findings have implications for the equitable provision of primary care services since black Caribbean women experienced depressive symptoms in pregnancy and early motherhood, but were less likely than their white British counterparts to receive treatment.