Two alternative trading mechanisms for securities markets are compared using laboratory experimentation and computer simulation. One mechanism is the floor-based specialist auction in place in most U.S. stock exchanges today, and the other is an electronic alternative employing automatic order matching. We conclude that transition from the established floor-based exchanges to potentially superior electronic alternatives is possible, despite the inertia resulting from the experience of benefits investors trading in active markets, and that current proposals for electronic markets are not demonstrably superior on generally accepted criteria used to assess market quality. This has clear implications for established stock exchanges, market regulators, and vendors of electronic trading systems.