Although UV vision was first demonstrated in birds in the early 1970s, its function is still unknown,. Here we review the evidence for UV vision in birds, discuss the special properties of UV light, lay out in detail hypotheses for the function of UV vision in birds and discuss their plausibility. The main hypotheses are that UV vision functions: (i) in orientation, (ii) in foraging and (iii) in signaling. The first receives support from studies of homing pigeons, but it would be unwise to conclude that orientation is UV's primary function in all birds. It is especially important to test the signalling hypothesis because bird plumage often reflects UV and tests of theories of sexual selection have virtually always assumed that birds perceive plumage "colours" as humans do. A priori this assumption is unlikely to be correct, for unlike humans, birds see in the UV, have at least four types of cones and have a system of oil droplets which filters light entering individual cones.