Summary.Information concerning all 10 859 singleton deliveries in Greece in April 1983, were analysed to assess the contribution of socioeconomic factors to the perinatal mortality rate. Statistically significant associations were initially found with parental eduction, parental ages, duration of marriage, paternal occupation and parity. There was no association with maternal smoking habit, maternal occupation during pregnancy, type of health insurance or housing conditions. Once logistic regression analyses had taken account of the strong parity effect (P<0.0001), only a moderate association with maternal age (P<0.05) remained statistically significant, together with a marginally significant (P<0.05) association with maternal education level. Mothers who were moderately well educated had the lowest risk of loosing their baby. It is concluded that traditional measures of social deprivation appeared to have little effect on perinatal mortality in Greece in 1983.