As in other related disciplines, researchers in developmental toxicology have pursued the goal of defining structure-activity relationships within chemical classes. Due to the inherent complexity of developmental processes, however, it has often been difficult to find significant and predictive relationships between biological activity and chemical structure in the area. The review covers four examples of such efforts. Each approached the problem with different levels of experimental or statistical sophistication. Finally, comments are provided on how future research efforts may be constructed to improve the probability of overall success.