Scholars and policy makers consider knowledge accumulation a major driver of growth and regional development. During the past two decades, the geography of innovation literature has provided a rich and detailed account of the underlying processes of regional knowledge production. More recently, a growing body of empirical literature has analysed the specific knowledge bases of regions and their evolution over time. The aim of these studies is not to explain why some regions produce more knowledge outputs than others, but why some regions produce a specific type of knowledge. The author refers to this body of literature as the relatedness literature. In the chapter the author discusses the theoretical foundations of this literature, its methodological framework and recent empirical findings. Based on evolutionary thinking, the spatial dynamics of knowledge are understood as a cumulative, path-dependent and interactive process. As a result, a main driving force is relatedness, as new knowledge is expected to branch out from related, pre-existing knowledge. Empirically, relatedness has mainly been formalized as a network, the knowledge space. In this network, nodes are knowledge categories, such as technological classes or scientific fields, and the links between these knowledge types indicate their degree of relatedness. The empirical analysis of relatedness and the knowledge space allows the mapping of regions’ knowledge bases, explaining scientific and technological change and identifying further opportunities for recombination and innovation. After having reviewed the empirics on knowledge space, the author discusses implications for research and innovation policy and suggests some avenues for future research.