We conducted large-scale lost letter experiments in Beijing, a megacity with more than 21 million residents, to test if the observed altruistic attribute of the letter recipient would induce more passersby to return the lost letters. The treatment letters were addressed to a nationally renowned charitable organization in China, while the control letters were intended to an invented individual. A total of 832 ready-to-be-posted letters were distributed in 208 communities across eight districts in the city. The overall return rate was only about 13%. Yet, the return rate of the treatment letters (17%) was nearly twice as high as that of the control letters (9%). The finding adds large-scale field experiment evidence in support of the interdependent other-regarding preferences theory. In addition, we also found that the lost letters were more likely to be returned if they were dropped in communities with a relatively higher income or a postal box located closer.