The joint action for malformations induced by 18 binary mixtures of four carboxylic acids was determined, using Xenopus embryos. The acids were selected to evaluate whether joint actions for mixtures changed when the data were analyzed by specific malformations or malformation syndromes as opposed to joint actions determined when any type of malformation was included in the analysis. Each of six binary combinations of the acids were tested three or four times in 96-h static-renewal tests, starting with late-blastula stage embryos. Toxic unit analysis indicated that all but five of 56 possible analyses showed concentration addition joint actions. Concentration addition was observed for all instances in which microcephaly was included as a malformation (with or without other malformations being included) and for all possible mixtures where only eye defects were evaluated. In five instances, evaluating only abnormal gut coiling, the joint action was response addition, including all three mixtures for the valproic acid and pentanoic acid combination. The other 29 analyses which included abnormal gut coiling showed concentration addition. The results indicate that, for chemicals that are structurally similar and induce the same types of defects in frog embryos, there is no significant change in the joint action when the data is analyzed by specific malformations or syndromes rather than by including all malformations together. Abnormal gut coiling, a malformation induced by most toxic chemicals, appears to be less reliable for joint action determination, especially if it is the primary manifestation of abnormal development.