A prospective study was conducted to compare the accuracy of visual estimation of blood loss (EBL) at delivery with laboratory determination of measured blood loss (MBL). It showed that EBL tends to be clouded by the conventional teaching that blood loss at delivery is usually between 200 to 300 mL. Women with MBL up to 150 mL were overestimated and the best correlation was in women with MBL between 150 to 300 mL. There was a tendency to underestimate blood loss when the MBL was between 301 to 500 mL. Of the 9 women with a primary postpartum haemorrhage, only one was correctly diagnosed as such and 3 women were estimated to have blood losses of at least 500 mL but the measured blood losses were all lower. It was concluded that visual estimation of blood loss is inaccurate, especially at the extremes of MBL and that primary postpartum haemorrhage is not detected by visual estimation of blood loss, unless there are associated signs of haemodynamic instability.