In two experiments, this study compared anaphor resolution times in negative sentences (e.g., Either Peter does not catch a train, or it will arrive late in the evening.) with those in affirmative sentences (e.g., If Peter catches a train, then it will arrive late in the evening.). Sentences were read segment-by-segment, and segment reading times were being recorded. In line with the hypotheses, segment reading times following the anaphoric expression were longer in the negative than in the affirmative condition, but only when the critical entity was being referred to (e.g., the train as compared to Peter). When instead of a pronoun a repeated-name was being used for reference (e.g., the train as compared to it), resolution times were faster specifically in the negative condition. Implications for different accounts of language comprehension are discussed.