Research suggests that forcing participants to withhold responding for as brief as 600 ms eliminates one of the most reliable findings in prospective memory (PM): the cue focality effect. This result undermines the conventional view that controlled attentional monitoring processes support PM, and instead suggests that cue detection results from increased response thresholds that allow more time for PM information to accumulate. Given the significance of such findings, it is critical to examine the generalizability of the delay mechanism. Experiments 1-4 examined boundary conditions of the delay theory of PM, whereas Experiment 5 more directly tested contrasting theoretical predictions from monitoring theory (e.g., multiprocess framework) and delay theory. Using the same (Experiment 1) or conceptually similar (Experiment 2) delay procedure and identical cues (nonfocal "tor" intention) from the original study failed to show any influence of delay on performance. Using a different nonfocal intention (first letter "S") similarly did not influence performance (Experiment 3), and the difference between focal and nonfocal cue detection was never completely eliminated even with delays as long as 2,500 ms (Experiment 4). Experiment 5 did find the anticipated reduction in the focality effect with increased delays with a larger sample (n = 249). However, the focality effect was not moderated by attention control ability despite the fact that participants with impoverished attention control should benefit most from the delay procedure. These results suggest that any theory of PM that considers only a delay mechanism may not fully capture the dynamic attention processes that support cue detection.