The opening of a scientific research article bids for readers to invest time and attention in a writer's new ideas. To make such investments more attractive, a conventional introduction portrays new ideas as filling a serious gap in the current literature on an important topic. Do scientists rely on such introductory moves to decide what's worth reading and how to read it? We analyzed introductions of four texts and observed twelve scientists thinking-aloud while reading them. The texts were journal articles on chaos theory, including two recent and two early articles to reveal effects of disciplinary history. The textual analysis revealed that recent articles adhered closely to four conventional introductory moves, but early articles modified the moves, included more equations and fewer citations. Analysis of comments recorded during reading indicated that scientists attend to introductory moves: comments were concentrated within certain moves, rather than distibuted evenly across sentences; reading activity intensified for recent texts that adhered to the conventions.