The purpose of this study was to determine whether children can learn to cooperate in what has been described as the "minimal social situation." The research also compared the effectiveness of verbal instructions and a training task for teaching subjects the "win-stay, lose-change" rule. This rule has been used to explain the development of cooperation in the minimal social situation. Subjects were 19 teams of first-, second-, and third-graders. Five teams were composed of two girls; six were girl-boy teams; and eight were boy-boy teams. Ten of the 19 teams learned to cooperate in the minimal social situation without treatment. Two of four teams given the rule training procedure learned to cooperate after having failed to learn under typical minimal social conditions. Of five teams given verbal instructions, four learned to cooperate immediately. The probability of following the win-stay, lose-change rule was approximately 50% initially and did not increase significantly in later sessions. It is not clear then that following this rule is a prerequisite for the development of a cooperative exchange. Explanations in the literature which suggest subjects learn a single rule, i.e., win-stay, lose-change, may be misleading since children evidenced a variety of rules, any of which might have been reinforced or punished over the course of the experiment.