Witnesses sometimes report event details that are acquired solely from another witness. We reevaluated the potency of this memory conformity effect. After viewing a crime video, some participants learned about nonwitnessed details via discussion (dyad group), reading another participant's report (read group), or watching another version of the video (both-video group). In Experiment 1, these participants often reported nonwitnessed details, but on a source-judgment test most details were attributed primarily to the actual source rather than to the video. In addition, the dyad group was not more likely than the read or both-video groups to report nonwitnessed details. Participants in Experiment 2 were explicitly discouraged from providing details that were remembered from the secondary source only. These postwarning instructions substantially reduced the memory conformity effect, and a dyad group was not more likely than a read group to report nonwitnessed details. Encouraging source monitoring at test can reduce the negative consequences of co-witness collaboration.