Many animal taxa perform stereotyped displays during ritualized contests over territory and mates. Such displays facilitate assessment of an opponent's prowess and reduce the probability of physical injury. Color patterns likewise can serve as indicators of an individual's body condition. Male San Cristóbal Lava Lizards (Microlophus bivittatus) of the Galápagos Islands possess a black epaulet on each shoulder that contrasts with their surrounding body coloration. We created robots with the appearance of conspecific males to explore the potential function(s) of the epaulets in a series of four experiments. Our experiments focused on epaulet presence/absence, body location, and location relative to the axis of body motion (pectoral versus pelvic girdle) during pushups. Paired trials were conducted with adult males and females using a robot that performed M. bivittatus signature displays and that differed on its two sides in epaulet attributes. We quantified two kinds of bobbing displays in subjects: signature displays and two-bob displays. We also computed a composite response (CR) score, in which point values were assigned to types of nonbobbing displays and display combinations according to the apparent level of subject aggression. Results of Experiment 1 showed that males (but not females) performed more signature displays to the side of the robot with the epaulet present than to the side with the epaulet absent, suggesting that epaulets serve a signaling function. Composite response scores did not differ between treatments for either sex. In Experiment 2 neither males nor females performed more bobbing displays to an enlarged epaulet than to an epaulet reduced in size. However, males exhibited greater CR scores to the enlarged epaulet, suggesting that epaulet size or conspicuousness is important to its function. In Experiment 3 neither sex produced more bobbing displays to an epaulet on the robot's shoulder than to one on its hip. Yet, males accumulated significantly larger CR scores in response to the shoulder epaulet than to the hip epaulet, indicating that males were aware of normal epaulet location. Finally, in Experiment 4, where our robot performed hindlimb rather than forelimb pushups, subjects failed to exhibit more bobbing displays in response to a shoulder epaulet than to a hip epaulet. CR scores also did not differ between treatments for either sex, perhaps due to the striking abnormality of hindlimb-driven bobbing displays. Although the function of M. bivittatus epaulets remains elusive, these markings clearly are salient to male conspecifics.