AbstractA group of 88 spontaneously occurring variations of the skeleton were studied in adult CD‐1 mice which had been exposed in utero to the herbicides 2,4,5‐T or Trifluralin, administered to their dams by gavage, in an attempt to detect subtle biological effects of these compounds even in the absence of gross malformations. Sixty‐seven characters showed variation in this population. Of these, 19 were significantly different in frequency from untreated controls in mice having received a teratogenic dose of 2,4,5‐T, with an average difference of 23.7%. Seventeen of them were increased in frequency. By contrast, only three traits differed from untreated among mice receiving a no‐effect dose of 2,4,5‐T. Twelve traits differed significantly from untreated in the Trifluralin‐treated group; the average difference was 18.8%, with 10 of the traits being increased in frequency.Two clusters of affected variants specific to the 2,4,5‐T high dose group included frontal bone variants in the skull (presence of an interfrontal bone and fusion of the frontals) and variants in the cervical vertebrae (imperfect foramina in the first and second cervicals, dyssymphysis of the second cervical, and a shift of the arch foramen from the fourth to the fifth cervical). Two other effects peculiar to the 2,4,5‐T‐treated specimens were a loss of the prominent dorsal spine of the second thoracic vertebra and a reduction in the number of caudal vertebrae. The most obvious effects specific to the Trifluralin treatment were an increase in occurence of 14 ribs, an undoubled foramen ovale, and the occurence of accessory foramina in the cervical vertebrae. A striking increase in frequency of parted frontals was seen in both 2,4,5‐T high dose and Trifluralin‐treated groups.The skeletal variant assay system may be useful as a postnatal screen for detecting prenatal exposure to potenti