Despite earlier evidence that the presence of 2 redundant cues can facilitate activation of a common response, T. C. Rickard and D. Bajic (2004) found no dual-cue facilitation in the case of cued recall, provided that each cue-response association was learned independently. In this study the authors investigated the generality of their results using a dual-task cross-talk design. There was no evidence of dual-cue facilitation for compatible cue trials in the case of associative independence. Race models as well as at least some limited capacity parallel retrieval accounts can be eliminated by these and related results. It appears instead that a preretrieval stage performance bottleneck precludes cued recall through more than 1 independently represented cue-response association at a time.