Anthropology has a longstanding interest in the evolution of social and political systems. Although human cultures are diverse due to their histories and environmental settings, comparative research has established the existence of predominant general evolutionary trends. Among these are: increasing scale of the maximally organized political unit; increasing structural and functional differentiation of organized groups, involving specialization and hierarchical decision-making; and increasing social and economica stratification. These macro-trends may be summarized by stating that the complexity of societies has increased during long time spans. In the cultures of the Pacific basin, Sahlins (1958) was one of the first to realize that the scattered islands of Polynesia offer advantages for investigating the evolution of societal complexity, particularly of stratification. Polynesians, he noted, are phylogenetically closely related: all Polynesian peoples shared the same biological, linguistic, and cultural ancestry. Their common ancestry allowed him to study the diversity of Polynesian social stratification as a process of adaptive variation (Sahlins 1958:ix) that followed the colonization of scattered archipelagos with diverse environments. The degree of stratification in precontact times, he concluded, varied directly with island productivity. Sahlins's study suffered from the problems inherent in his functionalist approach, and he was unable to measure productivity--his main ecological determinant--satisfactorily. But the advantages of comparative studies of phylogenetically related peoples such as Polynesians are now well known. Polynesians shared a common ancestor (at around 3000 B.P., most specialists now believe) and remained largely uninfluenced by other peoples until the time of Western contact. This common cultural ancestry and relative isolation offers a large benefit to evolutionary researchers: we may assume that any differences between these cultures probably arose within Polynesia itself. This makes possible a controlled study of specific evolution, or of adaptive radiation (Kirch and Green 1987). In anthropology, this method was systematized by Eggan (1954) and has come to be known as controlled comparison. One chooses a region where cultures are believed to be descended from the same ancestor and tries to determine the factors that led to the subsequent evolution of differences between these cultures. Controlled comparison is most useful for sorting out the factors responsible for the evolution of cultural differences. Knowing (or assuming with good evidence) that the cultures under study were once the same culture allows us to more rigorously analyze the factors that made them diverse from one another. Researchers can analyze the factors influencing the evolution of differences relatively free from concerns that such differences were caused by the multiple origins and distinct histories of the cultures studied. Conversely, features that are shared within the area are already known to be homologous (similar because of their shared ancestry) rather than analogous (independently evolved) (Kirch and Green 1987). Knowing that similar features are homologous increases our confidence that differences arose as a consequence of environmental conditions or of internal sociocultural dynamics, rather than as a consequence external contacts or fortuitous events. These and other benefits of this method has led to its application to Polynesia (Kirch 1984), Mesoamerica (Flannery and Marcus 1983), and Athapaskan cultures (Aberle 1974; Perry 1983). Compared to Polynesian, Mesoamerican, and Athapaskan peoples, the islands of Micronesia in the western Pacific are less familiar to anthropologists and the public. Yet linguistic, cultural, and limited archaeological evidence suggest that some of these islands offer the same research advantages as those of Polynesia: historical relatedness and relative isolation. …