摘要: Honey bees can locate an inconspicuous food source by using prominent landmarks to "triangulate" it (reviewed Gould and Towne, 1987). Doubtless this strategy is a necessity for animal with such poor visual resolution. Although Anderson (1977) argued that are remembered nonpictorially, while Cartwright Collett (1982, 1983) concluded eidetically, (1985a, 1987) has shown the memory photographic resolution of about 3 ~ This contrasts relatively low flower-shape memory, which 8-10 (Gould, 1985b, 1986a), suggesting these two systems pictorial stored separately. (Real-time acuity 1-2~ sun's apparent diameter, calibration, 0.4~ Since flower shape learned on approach 1988), finding landmark committed departure would further strengthen conclusion pictorial-memory separate. Opfinger (1931) performed some early experiments hint at departuredependent landmark-learning system. In one test, she trained collect from corner checkerboard, rotating 180 between arrival departure; returning landed both corners. could be interpreted as meaning learn departure, or they treated pattern flower-learning task, learning not one-trial task. second experiment, transferred location another 10-25 m away were feeding, returned hive. When feed, went training table, release intermediate point; majority,