The aim of the study was to investigate the rationale of the current indications for fetal chromosome analysis. 5372 women had 5423 amniocentesis performed, this group constituting a consecutive sample at the chromosome laboratory, Rigshospitalet, Copenhagen from March 1973 to September 1980 (Group A + B). Pregnant women 35 years of age, women who previously had a chromosomally abnormal child, families with translocation carriers or other heritable chromosomal disease, families where the father was 50 years or more and women in families with a history of Down's syndrome (group A), were compared to women having amniocentesis, although considered not to have any increased risk of fetal chromosome abnormality (1390 pregnancies, group B). They were also compared with 750 consecutive pregnancies in women 25-34 years of age, in whom all heritable diseases were excluded (group C). The risk of unbalanced chromosome abnormality in group A (women with elevated risk) is significantly higher than in group B + C (women without elevated risk) (relative risk 2.4). Women with a known familial translocation and women 40 years or more have a relative risk of 5.7 of having an unbalanced chromosome abnormality compared with women without elevated risk. Spontaneous abortion rate and prematurity rate did not differ from rates expected without amniocentesis. It is concluded that current indications may be characterized as a mixture of evident high risk factors and factors with only a minor influence on risk. Indications for amniocentesis should therefore be reconsidered. Because it must be considered impractical and ethically wrong to limit amniocentesis to the two mentioned real high risk groups, and illogical to continue to present policy, which is not based on clearcut evidence, the possibility of offering amniocentesis to all who want it, is discussed. Screening for chromosome disease in all pregnancies is not without problems, but may be reasonable in some localities.