Most legal systems require jurors to consider all the evidence presented at trial. Hence when there is uncertainty over aspects of evidence this should be factored into juror judgments. Two experiments examined how mock jurors used uncertain information in their ratings of defendant guilt and final verdicts. Participants read scenarios where an eyewitness expressed uncertainty about the identity of a critical piece of evidence (e.g. the object a defendant was holding could have been a knife or a mobile phone). The respective probability of these alternatives was varied, as was their association with the alleged crime. When the probability of the alternatives was varied between subjects (Experiment 1) there was only weak evidence that jurors considered both alternatives. When probability was varied within-subjects (Experiment 2), jurors did consider both alternatives in their guilt judgments. The implications for theories of reasoning with uncertain information and forensic practice are discussed.