Caste conflict theory predicts that worker-destined individuals in insect colonies may try to develop as queens in order to gain greater direct reproduction. In situations where females can determine their own caste fate (‘self-determination’), this is expected to lead to overproduction of queens. Theoretical predictions are supported by patterns of queen production in Melipona stingless bees. In Melipona, queens and workers are similar in size and develop in identical, mass-provisioned, sealed cells, a situation which permits self-determination. In line with predictions, many Melipona females, up to 16%, selfishly develop as queens. Although these observations fit the predictions of caste conflict theory, some of the underlying assumptions remain incompletely tested. In particular, whether immature females can actually determine their own caste fate and whether queen production is really excessive rather than just an insurance against accidental queen loss? Here we test these assumptions. Queens and workers in colonies of Melipona beecheii in Yucatan, Mexico, had the same dry mass, showing that queen development is not conditional on above-average food provisioning. This supports the assumption that individuals can completely control their caste fate. Observations of 30 introduced virgin queens in three colonies showed that queens were killed rapidly and had a life expectancy of just 47 h. A second method, using the number of virgin queens in natural colonies, also indicated a short life span, estimated at 27 h. Rapid and efficient culling of virgin queens supports the hypothesis that queen production is excessive and results from caste fate conflict. In addition, we provide, for the first time, detailed behavioural descriptions of queen killing in this species.