The emergence of agents that play fair strategies is investigated in a simple bargaining model. The strategies played by the agents are constructed by evolutionary algorithms. Agents make offers to each other describing possible ways to share a certain commodity, until an offer is accepted. Finite-horizon bargaining models give an advantage to the first or last part making an offer, depending on the discount factor incurred by the players in each transaction. By introducing uncertainty regarding the playing order, i.e., who makes the first or last offers, experimental results show that evolutionary agents abandon greedy strategies, that attempt to obtain the whole commodity without sharing, for those that lead to more just divisions of the commodity.