Scientific synthesis is the integration of disparate research questions, theories, methods and data across scales, systems, scientific phenomena, and forms of expertise to increase the generality, parsimony, applicability, or empirical soundness of scientific explanations and science-based innovations. Synthesis generates “emergent” knowledge from incomplete theories and data, and attempts to explain phenomena that span disciplines or extend across spatial or temporal scales. Synthesis occurs both within and across disciplines and professional sectors, and so differs in substance, scope and method from simpler forms of interdisciplinary science. The concept and practice of scientific synthesis was catalyzed in ecology by the founding of the National Center for Ecological Analysis and Synthesis in 1995, and since that time the use of scientific synthesis as a strategy for producing transformative, integrative understanding of complex problems and processes has spread rapidly across disciplinary and national borders (e.g., the US Socio-environmental Synthesis Center, The Stockholm Resilience Center, and the French Center for Synthesis and Analysis of Biodiversity). Understanding the organization, processes, and accomplishments of synthesis is vital for informing policies for science, guiding leadership of groups and organizations, improving research assessment, and advancing fundamental knowledge, but our knowledge of the social conditions and the types of organizational environments and practices which best support synthesis are insufficient for these tasks. During this talk we will integrate existing theories of interdisciplinary knowledge production with data and findings from multi-method, longitudinal studies of two major environmental research organizations conducting synthetic science – the US National Center for Ecological Analysis and Synthesis and the Resilience Alliance – to distill lessons from these paradigmatic cases and develop a number of well-grounded principles that evidence indicates are important for achieving synthesis, while also considering how the goals, process and outcomes of synthesis may differ across disciplines.