Many prey organisms will approach (inspect) a potential predator, primarily to assess local risk of predation. It has been argued that by avoiding the head region of predators during inspections, prey can reduce the risks associated with such behaviour (attack cone avoidance). Prior experiments, however, have not incorporated the combined chemical and visual predator cues. We conducted laboratory experiments to determine the effects of combined predator dietary (chemical) and visual cues on the form and location of predator inspection visits by glowlight tetras (Hemigrammus erythrozonus). Tetras were exposed to a live cichlid predator (Cichlasoma octofasciatum) which had been fed tetras (with alarm pheromone), swordtails (Xiphophorus helleri; lacking Ostariophysan alarm pheromones) or food deprived. There was no significant difference in the overall rate of predator inspection by tetras to tetra-fed, swordtail-fed or food deprived cichlids. There was, however, a significant effect of predator diet on the size of inspection shoals and the location of inspections. Tetras inspected a tetrafed predator in significantly smaller groups, more often as singletons and directed a greater proportion of their inspections towards the tail of the predator. Tetras exposed to swordtail-fed or food deprived cichlids inspected in larger groups and directed more inspections towards the head of the predator. When the predator dietary cue contained tetra alarm pheromone, there was a significant increase in the attack latency by predators. Taken together, these data suggest that tetras use both chemical and visual predator cues during inspection visits and will modify their behaviour based on the presence or absence of conspecific alarm pheromone in the diet of a potential predator.