The present empirical analysis of the short texts accompanying research articles in the scientific journal "Nature" covering a period from 1991 to 2005, not only shows that these texts are significantly different from prescriptive models of abstracts, but that they have also recently undergone a further change. Up until 1996, in contrast to the traditionally viewed structure of abstracts (Introduction-Methods-Results-Conclusion/Discussion (IMRC/D)), the short texts in "Nature" vary considerably in structure with only 18% of those studied having the basic IMRC/D format and the Results being the only obligatory move. This manipulation of structure, accompanied by the predominant use of the Present and Present Perfect active, the use of modifiers, the apparent removal of hedging to strengthen claims all contribute to make these texts "eye-catching," to "advertise" a paper's contribution. With the introduction of the e-version of the journal in 1997, further changes occurred. Many promotional elements have been retained, and though the texts have become much more standardized in their structure, the Methods have been completely removed and the Results incorporated into the Conclusion which becomes the only obligatory move. This change in structure, combined with the inclusion of a greater amount of commentary, greater inclusion of information concerning the study's effect of the field as a whole, and the inclusion of explicit definitions, shows an evident concern for the "general reader" and indicates a kind of "democratization" of the scientific community. Technological advancements in the field of science appear to have also contributed to these modifications coming about.