A developmental toxicity study was conducted with ammonium perchlorate (AP) in the drinking water at doses of 0.0, 0.01, 0.1, 1.0, and 30.0 mg/kg-day beginning 14 days before cohabitation and continuing through sacrifice. Twenty-four rats/group were cesarean-sectioned on day of gestation (DG) 21 and fetuses examined for visceral and skeletal alterations. An additional 16 litters/group were sacrificed on DG 21 for maternal and fetal serum TSH, T(3), and T(4) (thyroid-stimulating hormone, triiodothyronine, and thyroxine) levels and thyroid histopathology. Clinical and necropsy observations, body weights, feed and water consumption, and cesarean-sectioning parameters were comparable among the groups with only delays in ossification observed in the 30 mg/kg-day group. Maternal thyroid weights were increased in the 30.0 mg/kg-day group. Decreased colloid was present in male and female fetal thyroids in the 1.0 and 30.0 mg/kg-day groups. Maternal TSH was increased and T(4) was decreased at all levels, and T(3) was reduced at 30.0 mg/kg-day. Fetal TSH was increased at 1.0 and 30.0 mg/kg-day, T(4) was reduced at 30.0 mg/kg-day, and T(3) was decreased at all levels. The maternal no-observable-adverse-effect level (NOAEL) was 1.0 mg/kg-day; exposures of 30.0 mg/kg-day increased absolute and relative maternal thyroid weights and histopathology findings. The developmental NOAEL was 1.0 mg/kg-day; developmental delays in ossification occurred in the 30.0 mg/kg-day group. The colloid depletion in the thyroids and increased TSH and decreased T(3) and T(4) levels at lower exposures were considered adaptive and not adverse. No adverse effects on development at occurred levels that did not cause maternal toxicity. AP is not a selective developmental toxicant.