analyses of historical data together with experiments. Religious groups show higher levels of cooperation, often over longer periods of time, than many other organized, but non-religious groups. Religious groups also tend to last longer as groups than non-religious organizations or institutions. Cooperation among religious groups 1s often facilitated by punishment or the implication of punishment from a deity. In a study of 186 societies by the biologist Dominic Johnson, analyses showed that those who believed in a strong moralizing god, capable of doling out punishment, engaged in higher levels of cooperation, including paying taxes, complying with norms, and repaying loans. These observations are complimented by experiments showing that people are more generous about giving away their money ina bargaining game, and less likely to cheat, when they think about words associated with religion — divine, God, spirit, sacred, prophet — than when they think about neutral words. For example, in the dictator game, involving two anonymous players, one decides how much of a pot of money to give to the other. The recipient has no say, and is thus stuck with whatever the donor offers. In general, donors give either nothing or about half. When primed to think about religion, donors are more likely to give than keep the entire pot, and give more as well. The implication of these results is that the religiously-minded feel that they are being watched. Cueing up words that are indicative of their religious beliefs, heightens their vigilance and their moral obligations. Religion fuels altruism and fends off the temptation to cheat. All of the observations and experiments discussed above are fascinating and relevant to understanding the role of religion in past and present societies. But this evidence is irrelevant for understanding the evolutionary origins of religion. It is irrelevant because it can’t determine whether religion originally evolved to solve the problem of large sca