Recently Seeley (1982) claimed that bees working inside the hive exhibit three age castes: cell cleaning, brood care, and food storage. In a subsequent paper, Kolmes (1985) stated that “honey bees do not appear to pass through a temporal sequence of tasks while performing hive duties” and that previous findings of this age polyethism are due to subjective interpretations of data. Here we report new behavioral data and statistical analyses substantiating the claim that age polyethism exists in bees working inside the hive. We followed a cohort of bees until they were 15 days old; they underwent two behavioral transitions. The first took place when the bees were approximately 3–4 d old and was associated with a switch from cell cleaning to other activities, mainly in the broodnest. The second occurred when the bees were approximately 10–12 d old and was associated with a switch from brood care to food-storage activities and foraging. Thus we present strong statistical evidence of a cell-cleaning age caste and a brood-care age caste for bees, but further evidence is needed to demonstrate fully that bees also pass through a distinct food-storage age caste before starting to work outside the hive. In addition, we demonstrate that Kolmes‘s previous denial of age polyethism in hive workers was caused neither by unusual behavior of his bees nor by weakness of his Kolmogorov-Smirnov statistical test, but by sampling errors due to a small but critical flaw in his method of labelling bees.