The resource dispersion hypothesis (RDH) asserts that, if resources are heterogeneous in space or time, group living might be less costly than was previously thought, regardless of whether individuals gain direct benefits from group membership. The RDH was first proposed more than 20 years ago and has since accumulated considerable support. However, it is sometimes discredited because a priori tests of specific predictions are few, relevant variables have proved difficult to define and measure, and because its assumptions and predictions remain unclear. This is unfortunate because the RDH provides a potentially powerful model of grouping behavior in a diversity of conditions. Moreover, it can be generalized to predict other phenomena, including spacing behavior in nonsocial animals and utilization of resources other than food. Here, we review the empirical support, clarify the predictions of the RDH and argue that they can be used to provide better tests.