In an attempt to bridge the gap between search under time-limited (brief exposures) and list conditions, two experiments were designed to extend a previously obtained (Solman, 1975b) relationship between accuracy of search and exposure-time. In Experiment 1 subjects searched circular arrays of letters for a pre-specified target. Responses were sampled over an exposure-time range of 100 to 500 msec, and, in accord with Neisser's (1967) description of selection during visual search, it was anticipated that the available time would be sufficient for subjects to carry out a further selection on those occasions when the initial target candidate proved to be incorrect. The results however gave no obvious indication that subjects had attempted to compensate for incorrect selections, and it was suggested that as the physical stimulus was present for a maximum period of 1 SO msec (longer exposure-times were composed of an 180 msec presentation plus a period of blank field}, a second take might (a) have been ineffective due to the effects of image decay, or (b) have been impossible if associated with an eye movement. In Experiment 2 the time course data obtained in Experiment 1 was supplemented for short times (i.e., exposure-time varied from 15 to 400 msec), and the decay theory proposed as an explanation of the results obtained in that study was tested. The results ruled out image decay as an explanation of the previous findings. They also showed that the opportunity for eye movements need not prompt subjects to have a second go at selecting the target, although it was possible that the design of the study precluded effective use of this option. In discussion it was suggested that there may be fundamental differences between search under time-limited and list conditions.