We conducted an exploratory study where we exposed four experimentally non-naive female Betta fish to a plus maze. Each arm of the maze was illuminated by either a red, yellow, blue or green LED light. Fish were trained to locate food in the arm illuminated by the target color. Once fish entered the arm corresponding to the target color, food was delivered at the end of that arm. A trial was completed once fish consumed the food or 300 seconds elapsed. Performance was considered stable when no more than two errors were made across two days of testing. Once stable, the target color was switched to one of the remaining three wavelengths. Only three fish met proficiency on the first target color. Although inconclusive, the data shows steady learning curves for all fish across wavelengths tested. We can infer that fish were capable of differentiating between some wavelengths because fish initially made more errors when the target color was changed. It is possible that distinguishing between four different wavelengths was too difficult for the fish. Future work should modify this task to include discrimination between two illuminated arms at once. Errors could also be attributed to lack of motivation; some fish did not eat all food in time allotted. We suspect fish may have demonstrated foraging behavior as several subjects showed arm checking behavior. An understanding of stimuli that control Betta fish behavior will open future lines of research investigating the reinforcing effectiveness of stimuli in the fish's natural environment.
