Reproductive concessions between related and unrelated members promote eusociality in bees.

The cooperation of organisms to form a higher level of biological organization represents a major evolutionary transition1. Maintenance of a stable social group demands specific benefits to offset the costs incurred by individuals that help others reproduce. Individuals may maximize their inclusive fitness by controlling reproduction despotically or helping relatives. Kin selection predicts that animals will act in ways that tend to maximize their inclusive fitness2. In social hymenopterans, relatedness asymmetries between nestmates produce conflicts of interest as individuals simultaneously attempt to maximize their own reproduction2. In small insect societies, the most obvious potential conflict between breeders concern the partitioning of reproduction (reproductive skew) in groups lacking morphologically differentiated castes, where more than one individual is capable of reproduction3. How conflicts are resolved depends on the payoffs of the different reproductive strategies to each individual4,5. Reproductive skew theory has provided an important framework for understanding these strategies6,7,8,9,10. This theory is particularly interesting because it is relatively simple, comprising some aspects of the payoffs involved in alternative social contexts and the mediation of these payoffs such as competitive ability and relatedness8. The models based on skew theory attempt to discuss the skew based on the trade off of reproductive benefits, the result of which is shaped by a number of different social and ecological factors, including relatedness, resource-holding potential, group productivity and constraints on independent breeding11. The theory provides a convincing explanation of how and why conflicts are resolved, and has been suggested as a general theory of social evolution8. Previous studies have shown that a positive or negative relationship between skew and relatedness could be used to support transactional or tug-of-war models12,13,14. However, the generality of each model is restricted by their assumptions. Transactional models assume that a single dominant individual has control over group membership and the fraction of total group reproduction obtained by the subordinate breeder15. The dominant breeder maximizes her own fraction of reproduction at the expense of a related subordinate, but concedes just enough reproductive output to the subordinate to make it favorable for this individual to stay in the group. As an unrelated subordinate lacks this indirect benefit of staying, the dominant female must grant her a share of direct reproduction to maintain the association (individuals can negotiate based on the threat of group dissolution - “outside option”)8. Thus, one of the main predictions of the model is that reproductive skew will be high when relatedness between breeders is high15,16,17. In the tug-of-war models, neither individual has control over the allocation of reproduction8,12,18 (individuals can negotiate based on the threat of costly competition – the inside option)8. In contrast with concessions models, tug-of-war models predict the absence of a relationship between relatedness and skew18. This assumption of costly competition by both individuals impedes the evolution of more efficient form of reproductive sharing. The solution to this problem can be the association of the assumptions of the models of reproductive skew using Hamilton’s rule to predict the conditions under which the assumptions of major classes of models (transactional and tug-of-war) consider8. Therefore, synthesizing the transactional and tug-of-war models, it is possible to determine the conditions under which individuals will negotiate based on their options to leave or to stay8. A previous study showed that females of the allodapine bee Exoneura robusta are able to assess pairwise relatedness, either directly or indirectly, and use this information to mediate ovarian development19. This study suggests a path for future developments in skew theory, drawing attention to what has been widely considered to be an obscure point: the ability of individuals to acquire and process the types of information required for models of skew theory to function19. Euglossa melanotricha nests are usually multivoltine. Solitary females found new nests or can re-use inactive nests by mixing new with old resin to build the new cells. The process of nest re-use can be initiated when two newly-emerged females remain in their natal nest and one begins to reproduce (Fig. 1). Previous studies have shown that the multifemale societies of this orchid bee are usually formed by a mother and her daughters (matrifilial nests), sisters (full sibling nests) or usurpers and resident females (unrelated female nests)20,21,22. Different from other bee species, all E. melanotricha females can mate, but egg laying is regulated by the dominant’s behaviour and chemical signalling22. These behavioural features provide a rare opportunity to test predictions of the skew reproductive theory. Figure 1 Life cycle and types of nest associations of Euglossa melanotricha. Here, we predicted that dominant E. melanotricha females may do better to concede a small and cheap share of reproduction rather than enter into an escalated contest with a highly motivated subordinate. We evaluated the benefits of direct and indirect reproduction related to the genetic structure within the nests. In this primitively eusocial bee, dominant females control reproduction and capitalize on the direct reproduction of related and unrelated subordinates according to their interests.