To provide data about the frequency of prenatal misdiagnosis in achondroplasia (Ach), we retrospectively abstracted data from 37 consecutive referrals of infants with Ach where ultrasound was performed prenatally. Nine of 37 (24 per cent) had a positive family history of Ach; all nine were correctly diagnosed prenatally. Of the 28 with no family history of Ach, 16 (57 per cent) were recognized to have abnormalities on ultrasound but none was given a definite diagnosis of Ach. Five families received an appropriate diagnosis of "most likely' Ach and four others were given a non-specific (but appropriate) diagnosis of some dwarfing disorder, not otherwise specified. In seven instances (25 per cent), an incorrect diagnosis of a lethal or very severe disorder was provided. These results illustrate the difficulty of making a specific prenatal diagnosis of Ach. In the face of the resulting uncertainty, physicians appear to elect to emphasize the most severe of alternative diagnoses. Given the homogeneity of mutations within the fibroblast growth factor receptor 3 (FGFR3) gene in the vast majority of patients with Ach, FGFR3 mutational analysis can be offered in every instance where a short-limb disorder is ultrasonographically detected in the latter stages of pregnancy. This would reduce the amount of incorrect and potentially harmful information provided to families.