An examination of the flight directionality, the change in direction between arrival and departure, of Apis mellifera (Apidae) workers foraging on artificial inflorescences and on inflorescences of Echinops sphaerocephalus L. (Compositae) is reported. Thirty-seven circular, artificial inflorescences, each with three feeding tubes separated by 120° located on the perimeter were used. At the beginning of each visit all tubes contained 2 ± 0.2 μL of 1.2 M sucrose solution. Bees locating the tubes would probe one or more of them, then depart to another artificial inflorescence and repeat the foraging sequence. Changes in flight directionality were generated during within-inflorescence locomotion and not after feeding had ceased. Directionalities of bees that circled a second time around these inflorescences were indistinguishable from bees departing the same locations prior to completing a single cycle. Hence, the directionality of bees visiting tubes in the sequence 1-2-3-1, 1-2-3-1-2, or 1-2-3-1-2-3 were indistinguishable from the directionality of bees visiting tubes in the sequence 1, 1-2, or 1-2-3, respectively. The direction faced by bees while probing the final tube or floret (last-faced direction) was most often the departure direction; thereafter, bees flew ahead to the nearest inflorescence. An energetic analysis of the artificial inflorescence study showed that directionality was independent of nectar uptake rates. The experiment using real inflorescences, some of which had been enriched with ca. 10 μL of 1.2 M sucrose solution, confirmed that directionality was generated during within-inflorescence locomotion and that the last-faced directon was the most frequent departure direction; furthermore, changes between the last-faced and departure directions were independent of nectar uptake. Evident in the search pattern of honeybees were the following: (i) a forward-moving tendency while on inflorescences, (ii) a departure decision, (iii) departure in the last-faced direction, and (iv) minimum deviation from the last-faced direction when flying directly to the next inflorescence. Although bees tend to move in a forward-facing direction during all movement phases, the floral architecture facilitates or induces turning. The extent of the turn is limited by the bees decision to depart. It is concluded that directionality is an epiphenomenon resulting from interactions among floral architecture, bee locomotion, floral nectar content, and the bee's departure decision. Memory of the arrival direction is not required.