During the period 1985–1995 Daniel Koshland was Editor-in-Chief of the journal Science. As such he exerted a huge influence on all aspects related to content and lay-out of the journal. This study compares Science’s bibliometric characteristics between three periods: a pre-Koshland (1975–1984) period, the Koshland period (1985–1995) and the post-Koshland period (1996–2006). The distributions of document types, the country/territory and institutional distribution of authors, co-authorship data and disciplinary impact measured by subject categories of citations are studied. These bibliometric characteristics unveil some of the changes the journal went through under the leadership of Daniel Koshland.