Developmental processes suitable for the detection of effects of drugs used in human therapy were investigated in the freshwater polyp Hydra and in the marine hydroid Hydractinia. The organisms were treated with the anticonvulsant drug valproic acid (VPA), a suspected human teratogen causing exencephaly in mice, and 12 related substances. In Hydractinia the acquisition of the normal larval shape in the course of embryogenesis and the developmental step from the larval to the adult stage (metamorphosis) were studied. When applied during embryogenesis, the test substances were found to affect cellular adhesiveness. The greater the concentration, the greater the number of groups of cells that failed to integrate within the spindle-shaped larval body. The most potent substance was octanoic acid, while diethyl-acetic acid had almost no effect. In experiments on metamorphosis, 2-propyl-pentanol was the most potent in reducing the frequency of induction of metamorphosis, whilst 4-pentanoic acid was the least effective. Long treatments (24 hr) with low concentrations of VPA in Cs(+)-enriched seawater reduced, and short treatments (3 hr) with high concentrations increased, the frequency of induction of metamorphosis. VPA itself failed to induce metamorphosis. A possible explanation for this may be that VPA and the other potent compounds affect the turnover of potassium ions in Hydractinia larvae. In Hydra, head regeneration and its inhibition by the drugs was studied. Head regeneration was inhibited by low concentrations (less than 0.1 mm) of VPA, 4-en-valproic acid and octanoic acid, while valproic acid amide (valpromide) and diethyl-acetic acid had almost no effect. The order of potency of these test substances varies greatly with different mammalian in vivo and in vitro test systems. Therefore, a correlation between potency in hydroids and in mammalian test systems is not possible. However, some similarities are explored.