In the 1990s the U.S. antitrust authorities began to analyze explicitly the innovation effects of mergers by applying the "innovation market approach" (mostly based upon innovation incentive arguments but also on diversity arguments). In this paper we present the results of our analysis of 89 merger cases, in which the FTC challenged mergers with innovation aspects between 1995 and 2008. In our investigation we ask to what extent innovation effects were taken into account, what theories were used ex- or implicitly, what assessment criteria were applied, and what kind of remedies were imposed. Beyond the analysis of the general policy in regard to the analysis of innovation effects of mergers, we are particularly interested in the use of arguments about the benefits of parallel research as a process of parallel experimentation (diversity), because we think that these arguments - based upon a clearer evolutionary reasoning - can help to reinvigorate the analysis of innovation effects of mergers.