Multilevel (or modular) societies are a distinct type of primate social system whose key features are single-male–multifemale, core units nested within larger social bands. They are not equivalent to fission–fusion societies, with the latter referring to routine variability in associations, either on an individual or subunit level. The purpose of this review is to characterize and operationalize multilevel societies and to outline their putative evolutionary origins. Multilevel societies are prevalent in three primate clades: papionins, Asian colobines, and hominins. For each clade, we portray the most parsimonious phylogenetic pathway leading to a modular system and then review and discuss likely socioecological conditions promoting the establishment and maintenance of these societies. The multilevel system in colobines (most notably Rhinopithecus and Nasalis) has likely evolved as single-male harem systems coalesced, whereas the multilevel system of papionins (Papio hamadryas, Theropithecus gelada) and hominins most likely arose as multimale–multifemale groups split into smaller units. We hypothesize that, although ecological conditions acted as preconditions for the origin of multilevel systems in all three clades, a potentially important catalyst was intraspecific social threat, predominantly bachelor threat in colobines and female coercion/infanticide in papionins and humans. We emphasize that female transfers within bands or genetic relationships among leader males help to maintain modular societies by facilitating interunit tolerance. We still lack a good or even basic understanding of many facets of multilevel sociality. Key remaining questions are how the genetic structure of a multilevel society matches the observed social effort of its members, to what degree cooperation of males of different units is manifest and contributes to band cohesion, and how group coordination, communication, and decision making are achieved. Affiliative and cooperative interunit relations are a hallmark of human societies, and studying the precursors of intergroup pacification in other multilevel primates may provide insights into the evolution of human uniqueness.

Primate home range sizes can vary tremendously as a consequence of the analytical technique chosen to estimate home range. This is exemplified by a recent dataset on free ranging snub-nosed monkeys (Rhinopithecus bieti) in Northwest Yunnan, China. Our findings show that the grid cell method cannot substitute for the minimum convex polygon (MCP) method and vice versa. MCP-based estimates are far too large, especially when the form of the home range is irregular due to forays into peripheral areas. Here, we propose an adjusted polygon method, whereby unsuitable and never visited areas are clipped out from the polygon, thus producing more accurate results. Compared to the grid cell method, the adjusted MCP is much more robust when the number of group relocations is limited; MCP turned out to be the method of choice for calculation of monthly and seasonal home ranges. The grid cell method on the other hand yielded the most precise estimates for total or annual home ranges. The style of ranging exhibited by a given primate taxon or population determines which analytical procedures should be applied to estimate home range size, and we would stress the need for thorough evaluation of the pros and cons of home range estimators before conducting field work and analysing data.