Onsets do not override top-down goals, but they are responded to more quickly.

Do onsets automatically capture attention? Spatial-cuing experiments often reveal no capture by onset cues in searches for color targets. However, recent experiments have shown faster responses to an uncued color target presented as an abrupt onset than as a change to an existing item, which has been argued to reflect capture by the onset. In the present experiment, we tested whether this onset advantage reflects the capture of attention or processing independent of shifts of attention. In a modified spatial-cuing paradigm, noninformative color precues were paired with color targets presented as abrupt onsets or as no-onset characters. Critically, the number of other onset items in the target display was manipulated, which has previously been shown to disrupt attention allocation to any particular item. It was reasoned that if the onset advantage for uncued color targets reflects attentional capture, then the appearance of additional onsets should eliminate this advantage. The results showed that even with multiple onsets on the target display, the onset advantage remained additive with cue validity. The additive effects are inconsistent with automatic capture by onsets, suggesting instead that the onset advantage arises from an independent source.