Repeated switches from cooperative to selfish worker oviposition during stingless bee evolution.

Reproductive division of labour is a defining feature of insect societies. Stingless bees (Meliponini) are an interesting exception among the highly eusocial insects in that workers of many species contribute significantly to the production of males. Since workers remain sterile in other species of this large tropical tribe, it has been hypothesized that, in the latter species, ancestral queens have won the conflict over who produces the males. The fact that sterile workers of some species lay trophic eggs to feed the queen and display ritualized behaviours towards her during oviposition has been interpreted as an evolutionary relic of this ancient conflict. Here, I used ancestral state estimation to test whether worker reproduction is indeed the ancestral condition and worker sterility a derived state in stingless bees. Contrary to this hypothesis, data suggest that trophic egg laying was the ancestral condition, whereas selfish worker reproduction in queenright colonies evolved subsequently during stingless bee diversification. The appearance of worker reproduction in queenright conditions was tightly linked to the laying of trophic eggs, which suggests that having activated ovaries in queen presence facilitates the evolution of worker reproduction. Worker reproduction is also linked to brood cell architecture, but surprisingly not to colony size or queen-worker dimorphism. The reason for this association between brood cell architecture and worker oviposition is currently unknown. These results suggest that trophic eggs are not a relic of an ancient conflict, but a sign of overlapping interests between the queen and workers about who produces the males.