Much of current scholarly work argues that China and Chinese communities are distinguished by a culturally specific and unique pattern of networking based on personal relations (guanxi), but there is little agreement about whether such personal relations produce discrete factions or more disbursed, weblike connections. The literature on banking networks is similarly unclear, and most discussion has focused on regional groups, such as a clique made up of natives from Zhejiang and Jiangsu provinces, or shared professional values among bankers. None of these approaches adequately describes actual connections, and this case study provides a new and empirically broad approach that applies tools of network analysis to interlocking bank directorships in 1936. This analysis shows the existence of twenty-four isolated banks; three very small, discrete, and regionally based groups; and one huge, diffuse, and weblike bank network that included virtually all Chinese bank assets and extended to most of China's economically developed regions. This network had a multicenter core made up of densely linked large banks and a number of small start-ups associated with prominent individuals. Diffuse connections also characterized networks of individual bankers, though dense ties existed either among the most important bankers who each sat on numerous boards of directors (suggesting associations based on business ties, professional interests, and the importance of banks) or among individuals who shared common native place, defined at the level of adjacent counties (suggesting the simultaneous existence of strong but very local guanxi ties). In the end, bank networks arose as much from historical contingency as from cultural predisposition.