Recent applications of more elaborate quantitative methods (e.g., bloc cluster analysis and Guttman scale procedures) to legislative politics have sharply improved the ability of political scientists to specify significant dimensions of voting behavior. Party affiliation, constituency characteristics and cohesion within state delegations have been correlated with the congressional vote in a number of subject matter areas. Because of the masses of data to be handled, however, even the best of these studies have been limited in scope to a single year or to one Congress. This restriction has not prevented the authors of these works from demonstrating clearly the utility of their methodology, but it has limited correspondingly either the generality or the reliability of their conclusions. The present study, limited to the single issue of congressional voting on foreign aid, is an attempt to discover, by the application of quantitative methods to votes over longer time periods, how far the relationships previously suggested persist over time or are peculiar to the individual sessions in which they occur.