The Swedish industrial research institutes are research organisations that exist somewhat in between academy and industry, fulfilling an intermediary role as well as providing a space for research relevant to industrial companies, and they have a history of being funded by both state and industry as a way to support technical research in Sweden. The present study examines the history one of these institutes – the Institute for Surface Chemistry – with respect to three dimensions of its knowledge production: the role that basic and applied research has played for the institute, its external connections and the heterogeneity of its knowledge production, and how it has evaluated the quality of its research. The time period considered is 1975-2005, a challenging time for the Swedish institute sector, and the analysis is based on an interpretation of annual reports, research programs and newsletters from the period, as well as on interviews with institute managers and researchers.This work contributes to a wider research field in two respects. First, it provides input to the ongoing debate about how a changing research system is linked to changes in knowledge production. Second, it increases our knowledge of the Swedish industrial research institute sector, something interesting in its own right but that also can provide input to the ongoing policy reorientation vis-à-vis these institutes. The main novelty of the work is that it engagessystematically and historically with changes in knowledge production within an industrial research institute, something not done in previous studies of the sector. To briefly summarise the results, applied research gradually becomes more important than basic research at the institute, but basic research still keeps playing a rather large role for some time, even as this roleis downplayed in the official publications. At the same time, the institute becomes more heterogeneous in its knowledge production, associates closer with its industrial partners, and loses some of its independent knowledge production in favour of a more classic intermediary role. During the study period, the institute mainly ascertains the quality of its work through the use of traditional academic standards, and it retains a strong publication culture throughout.Three main conclusions are drawn: that the institute generally has oriented itself more towards its industrial partners; that this is the result of adapting to a situation in which the traditional state funding and political support appear ever more insecure; and that in spite of this general dynamic of adaption, the institute, thanks to a unique knowledge base or strong and well-connected actors, has sometimes been able to defend its preferred modes of knowledge production instead of adapting,something which also has had a notable impact on its development.