The fungal cultivars of fungus-growing ants are vertically transmitted by queens but not males. Selection would therefore favor cultivars that bias the ants' sex ratio towards gynes, beyond the gyne bias that is optimal for workers and queens. We measured sex allocation in 190 colonies of six sympatric fungus-growing ant species. As predicted from relatedness, female bias was greater in four singly mated Sericomyrmex and Trachymyrmex species than in two multiply mated Acromyrmex species. Colonies tended to raise mainly a single sex, which could be partly explained by variation in queen number, colony fecundity, and fungal garden volume for Acromyrmex and Sericomyrmex, but not for Trachymyrmex. Year of collection, worker number and mating frequency of Acromyrmex queens did not affect the colony sex ratios. We used a novel sensitivity analysis to compare the population sex allocation ratios with the theoretical queen and worker optima for a range of values of k, the correction factor for sex differences in metabolic rate and fat content. The results were consistent with either worker or queen control, but never with fungal control for any realistic value of k. We conclude that the fungal symbiont does not distort the ants' sex ratio in these species.