By means of a laboratory experiment with 508 participants, we study the impact of ego depletion on revealed distributional preferences. Subjects are exposed to a social preference identification procedure in 2 consecutive weeks. In the treatment intervention, they accomplish an ego-depletion task before being exposed to the procedure in 1 of the 2 weeks, and in the control intervention they accomplish a control task. Half of the subjects are exposed to the intervention in Week 1 and the other half in Week 2. Our design allows us to cleanly identify 3 separate effects on social preferences: (a) the effect of exposing subjects to the social preference identification procedure a second time, (b) the effect of the intervention per se, and (c) the effect of ego depletion in particular. We find that only the intervention per se has an effect on social preferences for some types, whereas the ego-depletion task does not have a significant effect compared with the control task, and preferences display a considerable degree of stability over time.