Abstract. Category cocitation and its representation through social networks is proving to be a very adequate technique for the visualization and analysis of great scientific domains. Its combination with pathfinder networks using pruning values r = ∞ and q=n–1 makes manifest the essence of research in the domain represented, or what we might call the ‘most salient structure’. The possible loss of structural information, caused by aggressive pruning in peripheral areas of the networks, is overcome by creating heliocentric maps for each category. The depictions obtained with this procedure become tools of great usefulness in view of their capacity to reveal the evolution of a given scientific domain over time, to show differences and similarities between different domains, and to suggest possible new lines for development. This article presents the scientogram of the United States for the year 2002, identifying its essential structure. We also show the scientograms of China for the years 1990 and 2002, in order to study its particular national evolution. Finally, we try to detect patterns and tendencies in the three scientograms that would allow one to predict or flag the evolution of a scientific domain.