If a facial expression of emotion is preattentively recognized it shou
ld be recognized independently of the direction of attention. In the p
resent experiment an outline face was located each side of a fixation
point. Either both faces were coloured gray or one was coloured gray a
nd one blue, and either both had a neutral expression or one had a neu
tral expression and one an angry expression. RT to the presence of a b
lue face in a display was slower when the blue face was neutral and th
e gray face was angry than when both blue and gray faces were neutral.
The irrelevant angry expression captured attention, indicating preatt
entive processing. There were no effects in a condition in which a hap
py face was substituted for the angry one.