We analysed two large collaboration networks -- the Microsoft Academic Graph (1800-2020) and Internet Movie Database (1900-2020) -- to quantify network responses to major historical events. Our analysis revealed four properties of network-environment interaction. First, historical events can influence network evolution, with effects persisting far longer than previously recognised; the academic network showed 45\% declines during World Wars and 90\% growth during La Belle Epoque. Second, node and edge processes exhibited different environmental sensitivities; while node addition/removal tracked historical events, edge formation maintained stable statistical properties even during major disruptions. Third, different collaboration networks showed distinct response patterns; academic networks displayed sharp disruptions and rapid recoveries, while entertainment networks showed gradual changes and greater resilience. Fourth, both networks developed increasing resilience. Our results provide new insights for modelling network evolution and managing collaborative systems during periods of external disruption.