Rescue behavior is considered a type of pro-social response, defined as a voluntary action directed to benefit another individual who is in a stressful or dangerous situation. In two experiments, we investigated whether dogs would rescue their owners when the person was trapped inside a wooden box and emitted clear signs of stress. The performance of these dogs was compared against that of a control group in which the owners remained calm while trapped. In addition, to assess if training modulated this behavior, we tested a group of dogs from the military trained in search and rescue tasks (Experiment 1). Results showed that dogs opened the box more frequently when the owner pretended to be stressed than when calm. Training shortened latencies to open the door but not the frequency of the behavior. In Experiment 2, we investigated if emotional contagion could be a possible mechanism underlying dogs' rescue responses by measuring dogs' behavior, heart rate, and saliva cortisol level in the stressed and calm conditions, and also controlled for obedience by having the calm owners call their pets while trapped. We replicated the findings of Experiment 1 as more dogs opened the door in the stressed owner condition than in the calm condition. In addition, we observed an increase in heart rate across trials in the stressed condition and a decrease across trials in the calm condition, but no differences in cortisol levels or stress-related behaviors between conditions. In brief, we found evidence that approximately half of the dogs without previous training showed spontaneous rescue behaviors directed to their owners. Neither was this behavior motivated by obedience nor by the motivation to reestablish social contact with the owner. We conclude that emotional contagion is a plausible mechanism underlying dogs' rescue behavior in the present protocol.

When confronted with a difficult or impossible problem, dogs tend to look back at humans and try to catch their attention, instead of trying to solve it themselves. This behavior has been interpreted as a help request, but it is debated whether dogs take into account prior experiences with people when selecting whom to turn to. In the present study, dogs were trained to discriminate between a generous experimenter who gave them food and a selfish one that took it away. After assessing that they had established a preference for the generous one, we exposed them to an unsolvable task in which food was locked inside a container, while the experimenters stood on each side of the apparatus. During this task, we measured their behaviors towards each experimenter. Results showed that dogs did not first turn to the generous experimenter. However, they gazed more at the generous experimenter during the task, which implies that they did, to some degree, selectively ask for help based on previous interactions. Moreover, they gazed more and spent significantly more time in contact with the female experimenter when she was generous, suggesting a possible synergic effect of the experimenters' ID (male/female) and their attitude (generous/selfish). All in all, these results suggest that, to some extent, dogs are able to use the information from previous interactions with unknown humans to selectively ask for help.Keywords Human reputation . Unsolvable task . Asking for help . Dogs Dogs are particularly skilled at communicating with people, and gazing at the human face plays a fundamental role in this ability (e.g., Prato-Previde & Marshall-Pescini, 2014). One pioneer study that focused on gazing as a communicative signal is that of Miklósi et al. (2003), in which dogs were exposed to two kind of problems: pulling a rope with food attached to