This article introduces CiteSpace II, a Java application designed to detect and visualize emerging trends and transient patterns in scientific literature. The approach conceptualizes a specialty as a time-variant duality between research fronts and intellectual bases, where research fronts are emergent and transient groupings of concepts and underlying research issues, and intellectual bases are the citation and co-citation footprints of these concepts in scientific literature. Kleinberg’s burst detection algorithm and Freeman’s betweenness centrality metric are used to identify emergent research front concepts and highlight pivotal points of paradigm shift. Two complementary visualization views—cluster views and time-zone views—are designed to improve the interpretability of the visualized networks. The contributions of the approach include algorithmically and temporally identifying the nature of an intellectual base, explicitly interpreting the value of co-citation clusters in terms of research front concepts, and reducing the complexity of visualized networks through the identification of pivotal points. The method is applied to two research fields—mass extinction (1981-2004) and terrorism (1990-2003)—and verified with domain experts. The practical implications and future challenges are discussed.This article introduces CiteSpace II, a Java application designed to detect and visualize emerging trends and transient patterns in scientific literature. The approach conceptualizes a specialty as a time-variant duality between research fronts and intellectual bases, where research fronts are emergent and transient groupings of concepts and underlying research issues, and intellectual bases are the citation and co-citation footprints of these concepts in scientific literature. Kleinberg’s burst detection algorithm and Freeman’s betweenness centrality metric are used to identify emergent research front concepts and highlight pivotal points of paradigm shift. Two complementary visualization views—cluster views and time-zone views—are designed to improve the interpretability of the visualized networks. The contributions of the approach include algorithmically and temporally identifying the nature of an intellectual base, explicitly interpreting the value of co-citation clusters in terms of research front concepts, and reducing the complexity of visualized networks through the identification of pivotal points. The method is applied to two research fields—mass extinction (1981-2004) and terrorism (1990-2003)—and verified with domain experts. The practical implications and future challenges are discussed.