41-54.]) to investigate the degree to which subjects find anger faces reinforcing. In the DILT, separate sequences of actions were paired with presentations of anger faces, neutral faces or a blank screen. After training, performance on the three sequences was measured in the absence of face stimuli. Saliva was collected for T measurement. Higher T predicted better learning on sequences paired with sub-threshold (i.e., presented too fast for conscious awareness) anger faces, suggesting that T is related to reinforcing qualities of these faces. In Study 2, we examined whether morning or afternoon T better predicted attention and vigilance to anger faces. Participants were tested at 9:00 and 15:00. At each session, saliva was collected for T measurement, and participants completed a Stroop task and a dot-probe task [Mogg K, Bradley BP, Hallowell N. Attentional bias to threat: roles of trait anxiety, stressful events, and awareness. Q J Exp Psychol A 1994;47(4):841-64.] with facial expression stimuli. Morning (peak) T was a better predictor of responses to anger faces than afternoon T. Morning T predicted greater Stroop-like interference to sub-threshold anger faces, as well as attentional orienting away from sub-threshold anger faces. These effects were not present for joy faces or for supraliminal anger faces. T may generally decrease aversion to threatening stimuli, and/or may specifically facilitate approach towards signals of dominance challenge.