There is conflicting evidence about the capacity for scientific collectives (e.g., research teams, centers) to seed grand innovations. Although scientific challenges often require large numbers of specialized experts to work together, many large organizational groups are susceptible to weak member motivation and poor coordination. We recently concluded a six-year programmatic investigation into this organizational conundrum. Our research considered how best to organize and support collaboration for scientific innovation. Our findings, along with extant research on collaboration and innovation in the organizational sciences, have led us to draw three conclusions for the management of team science. First, we conclude that, rather than single “teams,” many of the collective entities addressing interdisciplinary scientific challenges are more appropriately labeled scientific “Multiteam Systems” (i.e., MTSs). Therefore, referring to all scientific collectives as “scientific teams” can sometimes lead to incorrect conclusions about the best ways to support collaboration. Second, we conclude that processes of interteam leadership and boundary spanning communication, which serve to connect different component teams to one another, are essential to the overall success of scientific MTSs. However, we caution that this second conclusion does not necessarily imply that managers should attempt to create one “big team” characterized by overly integrated subgroups that have lost sight of their unique team identities and subordinate goals. Rather, our third conclusion is that managing MTS collaboration is a balancing act, which involves both the integration of efforts across teams as well as the recognition of component teams’ unique contributions, identities, and subordinate goals. In this chapter, we elaborate these three conclusions and summarize five properties of effective MTSs which are important targets for intervention strategies designed to facilitate multiteam functioning.