Three-dimensional spatial learning in hummingbirds

To date, most studies of spatial learning have been conducted in the horizontal plane, with few addressing animalsʹ use of the vertical dimension. It is not clear, therefore, what difference adding a vertical component to the learning/memory of a spatial location makes. The aim of this study was to examine the learning by rufous hummingbirds, Selasphorus rufus, of a rewarded location in the vertical and horizontal axes. We found that their ability to learn the location of a rewarded flower in a linear flower array depended on its orientation and on the spacing distance between the flowers in the array. Birds chose the flower in the rewarded location more than any of the other four flowers in the horizontal array but only when flowers were spaced 30 cm apart. In contrast, in the vertical array or when the flowers were 5 cm apart horizontally, birds only avoided visiting the two distal flowers. In addition, the cues the birds used to learn the rewarded location depended on the spacing scale in the horizontal dimension: at 30 cm flower spacing, birds learned the absolute location whereas at 5 cm they chose the relative and absolute locations equally. In a diagonal linear array, birds chose the rewarded flower more than any other. When the array was oriented either vertically or horizontally for the test, the birds chose the flower in the same horizontal location as for the rewarded training location. The hummingbirds appear to favour the horizontal over the vertical component of a learned location.