Supernumerary ribs (SNR) of differing sizes are commonly observed in rodent developmental toxicity studies, and the significance of treatment-related increases in SNR in standard studies has been contentious. We induced dose-related increases in SNR in fetal CD-1 mice by treating on gestation days 7-8 with benomyl (BEN; 0, 75, 150 mg/kg/d), dinoseb (DIN; 0, 30, 50 mg/kg/d); 2-methoxyethanol (2-ME; 0, 75, 150 mg/kg/d), or valproic acid (VPA; 0, 125, 250 mg/kg/d). Incidences of SNR were 9.3-27.6% in controls and 19.3-84.4% in the high dosage groups. SNR length showed a bimodal distribution with peaks at 0.3-0.4 mm and 0.9-1.1 mm in both treated and control groups. Based on length distributions, we used an actual length of 0.6 mm to separate short (rudimentary) from long (extra) SNR. DIN, 2-ME, and VPA induced a dose-related increase of extra ribs, while the incidence of rudimentary ribs remained at control levels. There was no apparent correlation of the presence of either type of SNR in a fetus and the occurrence of other anomalies. These data support the idea that extra and rudimentary SNR may reflect separate developmental phenomena, and should be considered and reported separately in developmental toxicity studies for risk assessment.