The cultural validity of instruments to detect perinatal common mental disorders (CMD) in rural, community settings has been little-investigated in developing countries. Semantic, content, technical, criterion and construct validity of the Edinburgh Postnatal Depression Scale (EPDS) and Self-Reporting Questionnaire (SRQ) were evaluated in perinatal women in rural Ethiopia. Gold-standard measure of CMD was psychiatric assessment using the Comprehensive Psychopathological Rating Scale (CPRS). Community-based, convenience sampling was used. An initial validation study (n=101) evaluated both EPDS and SRQ. Subsequent validation was of SRQ alone (n=119). EPDS exhibited poor validity; area under the receiver operating characteristic (AUROC) curve of 0.62 (95%CI 0.49 to 0.76). SRQ-20 showed better validity as a dimensional scale, with AUROC of 0.82 (95%CI 0.68 to 0.96) and 0.70 (95%CI 0.57 to 0.83) in the two studies. The utility of SRQ in detecting 'cases' of CMD was not established, with differing estimates of optimal cut-off score: three and above in Study 1 (sensitivity 85.7%, specificity 75.6%); seven and above in Study 2 (sensitivity 68.4%, specificity 62%). High convergent validity of SRQ as a dimensional measure was demonstrated in a community survey of 1065 pregnant women. Estimation of optimal cut-off scores and validity coefficients for detecting CMD was limited by sample size. EPDS demonstrated limited clinical utility as a screen for perinatal CMD in this rural, low-income setting. The SRQ-20 was superior to EPDS across all domains for evaluating cultural equivalence and showed validity as a dimensional measure of perinatal CMD.