In this study, we develop a conceptualisation of science as a complex adaptive system. Scientometric data is used to investigate empirically the emergence of ‘search regimes.’ Complex regimes can emerge when three or more independent sources of variance interact. In our conceptualisation, researchers can be considered as the nodes that carry the science system. Research is geographically situated in practices with site-specific skills, tacit knowledge, equipments, and tools. The emergent science level refers to the formal communication activities of the codified knowledge published in journals and books, and presented at conferences. The socio-economic dynamics refer to the ways in which knowledge production provides and receives resources to and from social and economic developments. As cases for the empirical operationalisation of search regimes, we chose Biotechnology, Genomics, and Nanotechnology. These ‘new leading sciences’ can be characterised by rapid growth and divergent dynamics of search. The three dynamics are treated as equivalent in the model, but they are substantially very different. The selection mechanisms are expected to operate asymmetrically. Regimes can thus be distinguished in terms of the extent to which a synergy is self-organized among the main subdynamics.