To determine the number of teenage births and abortions in Amsterdam, the Netherlands, in total and according to age and ethnic origin. Retrospective analysis of routinely collected data. Based on data collected from the municipal population register and the abortion clinics in Amsterdam, birth rates, abortion rates and abortion ratios were calculated for the period 1996-1998. The birth rate was the number of live births and the abortion rate the number of abortions per 1000 teenage girls. The abortion ratio was the number of abortions per 100 pregnancies. The average birth rate and abortion rate for the period 1996-1998 were 3.2 and 7.6 for 14- to 16-year-old girls respectively, and 25.0 and 29.5 for 17- to 19-year-old girls respectively. Among 14- to 16-year-old girls 70 pregnancies out of 100 ended in an abortion and among 17- to 19-year-old girls 54 pregnancies out of 100. In both age groups the birth rate for Surinamese, Antillian, Moroccan, Turkish and Ghanian girls was higher than for Dutch girls. There was one exception: Moroccan girls aged 14-16 years did not have a higher birth rate than Dutch girls of the same age. Surinamese, Antillian and Ghanian girls aged 14-16 years and 17-19 years also had a higher abortion rate compared with Dutch girls of the same age. The Turkish and Moroccan girls did not have (much) higher abortion rates. In both age groups the abortion ratio was lowest for Turkish teenagers and highest for Ghanian teenagers. The figures for second-generation immigrant teenagers were more akin to those of the native Dutch girls than those of first-generation immigrants. In Amsterdam unwanted pregnancies were most frequent in Surinamese, Antillian and Ghanian teenage girls. To prevent these pregnancies further research into the (cultural) determinants of sexual behaviour is necessary.