This work investigated the effect of the Hj receptor blockade in the forebrain of ablated Carassius auratus in a simple stimulus-response learning task using a T-maze test with positive reinforcement. The goldfish were submitted to surgery for removal of both telencephalic lobes five days before beginning the experiment. A T-shaped glass aquarium was employed, with two feeders located at the extremities of the long arm. One of the two feeders was blocked. The experimental triáis were performed in nine consecutive days. Each fish was individually placed in the short arm and confined there for thirty seconds, then it was allowed to swim through the aquarium to search for food for ten minutes (máximum period). Time to find food was analysed in seconds. Animáis were injected intraperitoneally with chlorpheniramine (CPA) at 16 mg/kg of body weight or saline after every trial, ten minutes after being placed back in the home aquarium. The results show that all the training latencies of the A-SAL group were higher than the latencies of the S-SAL group. The S-SAL group had decreased latencies from the second trial on, while the S-CPA group showed decreased latencies after the fourth trial. The A-SAL group showed reduced latencies after the fifth trial, but the A-CPA group mainteined the latencies throughout the experiment. This suggests that CPA impairs the consolidation of learning both on telencephalon ablated animáis and in sham-operated ones through its action on mesencephalic structures of the brain and/or on the cerebellum in teleost fish.