Following reports on improved test–retest reliability of emotion-modulated startle during a 6-s picture presentation when different pictures are presented at each assessment (Larson et al., 2000) and data suggesting that brief picture presentations also elicit affective blink modulation (Codispoti, Bradley, & Lang, 2001), we assessed test–retest reliability of blink modulation during brief picture presentations. At two acoustic startle sessions (4 weeks apart) subjects viewed different IAPS pictures for either 6 s (long group) or 300 ms (short), with emotion modulation assessed at three different points in time during and following picture viewing. Group ANOVAs revealed emotion modulation for both short and long groups. In addition, comparable and, in some cases, greater stability of emotion modulation was found for short compared to long picture presentations. Stability was generally low for individual probe times for both groups.