To determine whether the improved survival of very low birthweight (VLBW) infants (< 1500 g) born in Australia can be attributed to currently high rates of Caesarean section, we examined the associations between neonatal mortality and Caesarean section in singleton liveborn VLBW infants (500-1499 g) born during 1986-93 in Victoria, Australia, using data from the Victorian Perinatal Collection Unit. The infants included in this study had completed > 23 weeks of gestation, had no life-threatening malformations and had not been delivered by a repeat Caesarean without a trial of labour (n = 2763). For infants weighing 500-749 g, 750-999 g, 1000-1249 g and 1250-1499 g, the neonatal mortality rates were 56.1%, 25.7%, 13.0% and 4.3% respectively, and the Caesarean section rates were 33.1%, 42.3%, 54.8% and 55.8%. Nearly half of these births (n = 1269) were associated with one or more obstetric indications for Caesarean section (non-breech malpresentation, fetal distress, prolapsed cord, placenta praevia, pre-eclampsia and hypertension). Overall, the odds ratio (OR) for neonatal death associated with Caesarean section was 0.92 [95% confidence interval 0.60-1.41], after adjustment for gestational age, birthweight, year of birth, type of hospital, presence or absence of labour, presentation and obstetric indications for Caesarean section. However, when the vertex-presenting (n = 1702) and breech-presenting (n = 746) infants were considered separately, the adjusted ORs for neonatal death were 1.98 [0.96-4.10] and 0.52 [0.29-0.96] respectively. For those infants without obstetric indications for Caesarean section, the adjusted ORs for neonatal death in vertex-presenting (n = 950) and breech-presenting (n = 446) infants were 3.80 [1.11-13.0] and 0.47 [0.23-0.6]. These recent population-based data support the view that Caesarean section does not enhance the neonatal survival of VLBW babies when obstetric complications are absent.