An equity index fund attempts to duplicate the performance of a selected benchmark as closely as possible. For practical reasons, it is often desirable to achieve this with a designated number of securities. In the operations research literature the selection of securities for index fund construction can be formulated as a quadratic 0-1 programming problem. Since such models are NP hard, heuristic methods are usually required to produce (approximate) solutions in a reasonable time. This article reports on an Ant Colony Optimization approach to selecting securities for inclusion in an S&P 500® index fund.