Ants use the panoramic skyline in part to determine a direction of travel. A theoretically elegant, but as yet untested, way of delineating the skyline is to use the maximum difference in the ratio of green wavelengths of light to ultraviolet (UV) wavelengths to define where the tops of terrestrial objects are. Terrestrial objects generally reflect relatively more green wavelengths while the sky contains relatively more UV wavelengths. Using a ratio in opponent-process fashion gains constancy in the face of changes in overall illumination level. We tested the use of UV wavelengths in homing ants by using a clear plastic that blocked wavelengths below 400 nm completely. Desert ants, Melophorus bagoti, were trained to home repeatedly from the middle of a uniformly coloured arena providing an artificial skyline (Experiment 1) or under natural conditions with the natural panorama (Experiment 2). On a test, an ant was allowed to run home, but was captured just before she entered their nest. Such a zero-vector ant was then brought back to the middle of a replicate arena (Experiment 1) or the starting point (the feeder, Experiment 2) and released for a test. Blocking UV light led to deteriorations in orientation in both experiments. If the artificial skyline was transformed from opaque to transparent UV-blocking plastic (Experiment 3) on the other hand, the ants were still oriented. We conclude that UV wavelengths play a crucial role in determining direction based on the terrestrial surround.