Background: Low birthweights as well as high perinatal mortality rates are common in most African populations. Little is known, however, about how low birthweight corresponds with higher mortality rates within African populations. Twins are known to have lower birthweights and higher perinatal mortality rates than singletons. If lower birthweights represent higher perinatal risk per se, small twins within a population with generally lower birthweights should have critically increased risks. Methods: In total, 15,255 births in a Tanzanian hospital during 1999—2006 were analysed to determine birthweight distribution and examine perinatal mortality rates (including stillbirths and neonatal deaths within 24 hours) by birthweight in twins and singletons. Referral births from outside the district where the hospital was situated were excluded from analysis. Results: The mean birthweight for births within an estimated normal distribution was 3172 g, with a standard deviation of 462 g. The overall perinatal mortality rate was 43.9 per 1000 births (95% confidence interval: 40.7—47.2). Perinatal mortality rates among twins and singletons were 91.0 and 41.1 per 1000 babies respectively, corresponding to a relative risk of 2.2 (95% confidence interval: 1.7—2.8). The birthweight distribution for twins was shifted to lower birthweights. Twins had a generally lower birthweight and an excess of extremely small births as compared to singletons. The increased mortality rate for twins appeared to be independent of birthweight. Conclusions: The two-fold increased risk of perinatal death for twins was observed across the whole birthweight distribution, and very small twins appeared to have an excess perinatal risk that was almost similar to that of larger twins.