Fig. 1 (A) Foraging set up for honeybees. (B) Bee looking at a target face. (C) The ability of bees to discriminate between images of human faces. The upper region shows task (target top and distracter bottom), and the column immediately below shows mean percentage of correct choices for five bees in non-rewarded tests (±1 S.E.M.). Bees were trained to the face at the top of column i versus a schematic face distracter, then to recognise the target face from the distracter in column ii. Bees then recognised the target face from novel distracters (columns iii, iv) but failed to discriminate faces rotated by 180° (column v). Pooled choices for non-rewarded tests in the respective conditions (i=166, ii=246, iii=199, iv=196, v=141).