The survival of Apis mellifera Linnaeus, 1758 adults under laboratory conditions is not ensured only by a well-balanced diet during the larval stage and the first days after emergence; their population density in each experimental unit also affects the longevity of these insects, In this context, where statistical techniques allow considering random effects, fixed and different, covariance structures are interesting to build models that add information to bees confined to the laboratory. The goal of this paper was to study bee mortality in an experiment with different densities in each experimental unit observed over time. The statistical analyses used a mixed model with a covariance structure AR(1) and concluded that the average mortality rate for the bees is less progressive when considering densities with 50 individuals in each experimental unit.