Two modes of attentional deployment can be distinguished on the basis speed: slow (e.g. endogenous cueing) and fast (e.g. visual search) (Wolfe, Horowitz, & Alvarez, Nature, 2000). Is it possible to make fast deployments in the "dead time" between initiation and completion of slow deployments? Observers looked for a cue (a "2") in a 9.3Hz stream of digits presented to the right of fixation, then reported the first letter in a second stream to the left. The time of the reported letter indicated the time of the attentional shift. A 4-item search array centered on fixation was presented for 107 ms at various SOAs; subjects detected a T among Ls. Search performance during the cue to shift interval was superior to performance when the shift of attention occurred. These data suggest that rapid deployments of attention are possible between the planning and execution of a slow deployment of attention.