Organizing for interdisciplinary research must overcome two challenges to collaboration: the cognitive incommensurability of knowledge and the political economy of research based in the disciplines. Researchers may not engage in interdisciplinarity because they would have to invest in new knowledge unrelated to their discipline or risk losing career-related rewards. Our field study of a university interdisciplinary research center shows that boundary spanning occurred as students interacted with scientific instruments in a symbiotic relationship through what we call symbiont practices: matching disciplinary language and methods with the experimental possibilities of instruments, developing co-specialization between students and instruments, and engaging disciplinary actors to design experiments using instruments. Instruments engendered incipient interdisciplinary possibilities, but they required the students - engaging in symbiont practices - to actualize that potential. Simultaneously, students required instruments in order to be classified, staffed on projects and placed in jobs. These practices resolved both the cognitive and economic challenges of boundary spanning by mobilizing material resources that were costly (cognitively and politically) for actors on either side of the disciplinary divide to engage. In conceptualizing interdisciplinary research as occurring through symbiont practices, we develop a sociomaterial perspective on boundary spanning.