In eyewitness research, presenting lineup photos sequentially (one at a time) is thought to be superior to presenting lineup photos simultaneously (all at once); however, recent research has identified a robust simultaneous lineup advantage. This simultaneous advantage has been explained by the feature-detection hypothesis, which posits that simultaneous presentation allows witnesses to discover that particular facial features are non-diagnostic of the perpetrator (i.e., features shared by all lineup members) because lineup members can be directly compared (Wixted & Mickes, 2014). In two experiments, the feature-detection hypothesis was examined with a load manipulation (Experiment 1) and using eye tracking measures (Experiment 2). Although there was a significant simultaneous advantage in both experiments, support for the feature-detection hypothesis was mixed. In Experiment 1, a load manipulation designed to interfere with the processes described by the feature-detection hypothesis, was ineffective, possibly due to poor performing lineups or to the unique processing inherent to faces. In Experiment 2, eye movements, as predicted by the feature-detection hypothesis, was expected to change as a participant focused in on useful features. Eye movements changed as predicted in the simultaneous lineup, but not in the sequential lineup. More work is needed to examine the plausibility of the feature-detection hypothesis. Several alternative experiments are discussed.