|Abstract:||Most characterizations of the intergovernmental system portray it in unidimensional terms and assume that interactions occur between a relatively stable set of actors, pursuing identifiable goals, attending, at more or less constant levels, to all issues, and classified best by level of government. This analysis demonstrates the variability inherent in grant interactions and attempts to identify, operationalize, and examine the determinants of that variation. We begin by concluding that progress toward an explanatory model is not possible within the confines of the traditional assumptions. Interactions are better described as non-purposive, unstable, and non-levelled. These observations lead us to the identification of four sets of factors--environmental, actor, issue, and time--that have the greatest influence on interaction variation. Close scrutiny of each set yields a series of variables that are used to "discriminate" among five major types of observed grant interactions--disintegrative conflict, symmetric and asymmetric bargaining, mutual accommodation, indifference, and standard operating procedure (routine). In the end, two "solution" models are derived, each accounting for almost 60% of the observed variation in interaction type. These results move us one very significant step closer to an heretofore elusive empirical model of intergovernmental relations.