The first experiments on sex determination in bees began with Dzierzon, Meves, Nachtsheim, Paulcke, Petrunkewitsch, Manning. Whiting, (1943) found multiple alleles in Bracon xo that are the Rosetta stone of sex determination in Hymenoptera. Whiting also discovered that some species of microhymenoptera do not possess xo sex alleles. Therefore, Hymenoptera apparently presents two types of sex determination superimposed on haplodiploidy. In the panmictic groups hemizygous (xo1, xo2,... xon) and homozygous (xo1xo1, xo2xo2... xonxon) are males while heterozygous (xo1xo2, ... xon-1xon) are females. There is no such series of xon in endogamous Hymenoptera, since the constant elimination of diploid males would be damaging to the population and the mutation of xo to xon would be quickly eliminated. Besides the Whiting hypothesis, four others are discussed. The new hypothesis of genomic imprinting, of Beukeboom, is eliminated since: a) spermatozoa that develop within the egg produce male tissue; b) telitokous parthenogenesis due to the fusion of two haploid cells develop into females; c) last instar larvae treated with juvenile hormone become queens. The Cunha and Kerr hypothesis (female determining genes are totally or partially additive and male determination is totally or partially nonadditive) explains all known cases. The xo is a female determining gene. Sex determination in social bees led to the gradual evolution of two systems of caste determination: one in which queens and workers are similar and males are very different (Apinae), and another in which workers and males are very similar and both very different from the queens (Meliponinae). This second system in stingless bees implies that many of the mutations that improve worker capacities also affect the males that will carry out some activities that in Apis are clearly female ones. Ten of these activities are described.