A first experiment compared the behavior of goldfish injected with puromycin immediately after each of a weekly series of brief discriminative training sessions in the shuttlebox to that of appropriate controls. Discrimination was not prevented, nor was escape from shock impaired, but probability of response to the conditioned stimuli, both positive and negative, was reduced substantially. These results suggest that puromycin interferes with the consolidation of conditioned fear. The null outcome of a second experiment, in which all training was given in a single long session instead of a series of short sessions, suggests (contrary to recent indications) that consolidation begins in the training session. The conditioned-fear hypothesis is supported by the results of a third experiment in which the animals were shocked upon entering a goalbox to which they had previously learned to swim for food; animals injected with puromycin, immediately after the shock, entered the goalbox more readily 1 week later than did appropriate controls.