We present evidence that, in some research fields, research published in journals and reported on the Web may collectively represent different evolutionary stages of the field with journals lagging a few years behind the Web on average; and that a “two-tier” scholarly communication system may therefore be evolving. We conclude that, in such fields, (a) for detecting current research fronts, author co-citation analyses (ACA) using papers published on the Web as a data source can outperform traditional ACAs using articles published in journals as data; and that (b), as a result, it is important to use multiple data sources in citation analysis studies of scholarly communication for a complete picture of communication patterns. Our evidence stems from comparing the respective intellectual structures of the XML research field, a subfield of computer science, as revealed from three sets of ACA covering two time periods: (1) from the field’s beginnings in 1996 to 2001; and (2) from 20...