Behav Ecol Sociobiol (1998) 43: 327±331 Ó Springer-Verlag 1998 Original article S. O'Donnell Dominance and polyethism in the eusocial wasp Mischocyttarus mastigophorus (Hymenoptera: Vespidae) Received: 8 December 1997 / Accepted after revision: 28 March 1998 Abstract Dominance interactions aected patterns of ary advantage promoting the ecological success of eu- non-reproductive division of labor (polyethism) in the social insects (Wilson 1990). Identifying the factors that eusocial wasp Mischocyttarus mastigophorus. Socially regulate polyethism, and how they have evolved within dominant individuals foraged for food (nectar and insect social insect taxa, remain as central challenges in insect prey) at lower rates than subordinate individuals. In sociobiology (Oster and Wilson 1978; Jeanne 1986a; contrast, dominant wasps performed most of the for- Page et al. 1989). It has long been recognized that re- aging for the wood pulp used in nest construction. Social productive status in eusocial paper wasps (Vespidae: dominance also aected partitioning of materials col- Polistinae) is aected by dominance interactions (Pardi lected by foragers when they returned to the nest. Wood 1948; West-Eberhard 1969). Research on a number of pulp loads were never shared with nest mates, while food polistine species has recently demonstrated that domi- loads, especially insect prey, were often partitioned with nance interactions can also structure polyethism (Reeve other wasps. Dominant individuals on the nest were and Gamboa 1987; Jeanne 1991; O'Donnell and Jeanne more likely to take food from arriving foragers than 1995a; O'Donnell 1995, 1998a). Dominant individuals subordinate individuals. The role of dominance inter- may avoid performing particular tasks when those tasks actions in regulating polyethism has evolved in the eu- remove individuals from opportunities for direct repro- social paper wasps (Polistinae). Both specialization by duction (West-Eberhard 1981). foragers and task partitioning have increased from basal Dominance interactions can also in¯uence behavior genera (independent-founding wasps, including Mischo- within the worker force when many individuals have cyttarus spp.) to more derived genera (swarm-founding little or no chance of reproducing directly (O'Donnell Epiponini). Dominance interactions do not regulate 1998b). When workers compete for social resources, forager specialization or task partitioning in epiponines. dominant individuals may obtain greater control of both I hypothesize that these changes in polyethism were task performance and resource utilization. Workers can enabled by the evolution of increased colony size in the con¯ict over whether to perform certain tasks, especially Epiponini. those involving high energetic costs or risks of mortality (O'Donnell and Jeanne 1992, 1995b). Subordinate Key words Division of labor á Foraging á Social insects á workers are expected to perform risky or costly tasks at Task partitioning á Task specialization á Eusocial wasp higher rates. Foraging behavior entails great risks to workers, given its high energetic costs, wear and tear on the body that decreases longevity, and increased expo- sure to predators (O'Donnell and Jeanne 1995b). If Introduction dominance aects task performance, foraging should be performed by subordinate individuals. The predicted Division of non-reproductive tasks among nestmates, or relationship between dominance and foraging behavior polyethism, is thought to have been a major evolution- is often manifest in groups of reproductive females that cooperate in founding new nests (Gamboa et al. 1978; Litte 1979, 1981; Pfennig and Klahn 1985), but has rarely been tested among ospring workers (O'Donnell Sean O'Donnell 1998a). Department of Psychology 351525, University of Washington, Seattle, WA 98195, USA In addition to individuals performing dierent sets of e-mail: [email protected], tasks (specialization), polyethism comprises task parti- Tel.: +1-206-5432315, Fax: +1-206-6853157 tioning (Jeanne 1986a). Task partitioning involves the 328 performance of component tasks in a linked series by were located on the eves and rafters of a building at approximately dierent individuals. If there is no partitioning, indi- 1500 m elevation. Adult wasps present on the colonies (labeled A± F) were marked for individual identi®cation with paint pens. Be- viduals perform the entire task series themselves. For cause the subject colonies had already produced adult ospring example, if foraging is partitioned, foragers ®xate on when located, I could not determine whether worker females were collecting materials and pass their loads o to adult foundresses or daughters that had emerged on the nests. Behavioral nestmates for processing. If foraging is not partitioned, data were collected by an observer seated on scaolding or stan- workers keep and work with their loads after returning ding on a ladder within 0.5 m of the face of the nest. Behavioral observations were conducted for 3 continuous hours in the morn- to the nest. ing (between 0800 hours and 1140 hours local time) and two con- In an earlier study, I hypothesized that dominance tinuous hours in the afternoon (between 1300 hours and 1630 hours interactions in¯uence patterns of task partitioning local time) on two consecutive days per nest, for a total of 10 h (O'Donnell 1995). Nestmates can compete for access to observation time per colony (9 h total at colony E). All social in- teractions, forager departures and arrivals, and material transfers foragers' food loads for personal consumption. Control were recorded, noting the identity of interacting adults, and time to of food materials in¯uences individuals' own physiology the nearest minute. (Hunt 1994), as well as that of developing brood Behavioral acts coded as dominance interactions included bit- (O'Donnell 1998b). Wasps may also compete to control ing, chasing, displacing (one individual ¯ed another upon approach but was not followed) and darting (one individual leapt toward the processing of building material loads. The ability to another but did not make physical contact). In all cases, a subor- monopolize building materials allows individuals to dinate individual was clearly identi®ed when it crouched, ¯ed, or regulate nest growth. I predicted that more dominant turned away from the aggressor. See Itoà (1985) for descriptions of individuals would take and utilize the materials collected dominance interactions in primitively eusocial wasps, including by less dominant foragers. Conversely, dominant indi- Mischocyttarus spp. viduals should rarely give up their loads when they re- Statistical analyses turn to the nest from foraging. I studied the eects of dominance interactions on Contingency table analyses (testing relationships between categor- polyethism in the independent-founding wasp Mischo- ical variables) were performed with the likelihood ratio Chi-square cyttarus mastigophorus Richards. I quanti®ed domi- test (Fienberg 1989). I tested whether foraging trips diered in duration among materials using the survival analysis Wilcoxon test nance interactions, task specialization, and task (SAS 1985). Foraging trips that were initiated before the start of partitioning to address two questions. First, do domi- observations were included in the analysis as censored values (Pyke nance interactions in¯uence task performance? If so, I and Thompson 1986). predicted that less dominant individuals would perform I calculated a dominance index for each individual (male and risky tasks, such as foraging, at higher rates. Second, do female) as follows: dominance index (number of times individual was dominant±number of times individual was subordinate)/num- dominance interactions in¯uence how forager-collected ber of observation hours the individual spent on the nest. materials are partitioned at the nest? If so, I predicted This index (rather than relative dominance rank) was developed that dominant individuals would be more likely to take to assess the relationship between polyethism and dominance status materials from incoming foragers. I conclude by com- for two reasons. First, position in a linear dominance order could not be determined for many individuals because relatively subor- paring polyethism in M. mastigophorus with other spe- dinate individuals rarely engaged in dominance interactions with cies of eusocial paper wasps. The structure of polyethism each other. Second, this index includes information on the overall has evolved within the Polistinae (O'Donnell 1995), and rate at which an individual gave and received aggressive behavior. I comparison of polyethism among polistine genera can wanted to test the quantitative eects of dominance interactions on task performance and partitioning. Eects of individuals' domi- yield insight into how these evolutionary transitions nance status, as indicated by the dominance index, on task per- occurred. formance (food foraging rates) and task partitioning (rates of taking food from foragers) were tested using linear regression (Sokal and Rohlf 1981; SAS 1985). For these analyses, dependent variables were nested within colonies in the regression models to Methods control for inter-colony dierences. Variables coded as proportions were arcsine transformed to normalize their distributions (Sokal and Rohlf 1981). Study species M. mastigophorus is relatively common at higher elevations (above 1475 m) at the study site in Monteverde, Costa Rica (10°18¢ N, Results 84°49¢ W) (O'Donnell and Joyce, in press). Nests are