BACKGROUND:Complex seasonal variations in abortions and seasonal trends have been reported worldwide, but there are no data from Australia. MATERIAL/METHODS:Hospital morbidity and birth data were modelled using time series and regression techniques to describe seasonal changes in births and abortions (1989-99) in NSW, Australia. Data were also analysed to determine demographic predictors of abortions. RESULTS:A higher proportion of births in NSW were recorded in March and September-October. Irrespective of the nuptiality the lowest seasonal indices were recorded in February and November. For married women aged 20-49 years, voluntary abortions were significantly higher in February (t=0.04) and significantly lower in April (t<0.001) and June (t=0.03) and non-voluntary abortions were significantly higher in February (t=0.002) and March (t=0.006). There was seasonality in voluntary abortions (F=4.3, p<0.001) and no seasonality in non-voluntary abortions (F=2.4, p=0.01) for never married women aged 20-49 years. As similar to married women, voluntary abortions among never married women were significantly higher in February (t<0.001) and significantly low in April (t=0.009). The lowest number of abortions were in April. Non-voluntary abortions occurred mostly in March for married women. The average number of voluntary abortions were significantly higher for never married women than for married women. CONCLUSIONS:There was a significant seasonality of births and abortions in NSW, Australia, with a peak of conceptions in December/January. Abortions were lowest in April and non-voluntary abortions were peaked in March. These data suggest that the human reproductive behaviour has a close relationship with environmental factors.