Measurements from Lauder, New Zealand, and from the high-altitude Mauna Loa Observatory, Hawaii, are used to determine the altitude effects on spectral UV irradiance and to relate these altitude differences to other factors that influence UV radiation. The measured ratios UVMauna Loa/UVLauder are complex functions of both wavelength and solar zenith angle (SZA). Spectrally, the ratios tend to increase toward shorter wavelengths through most of the UV-A region. For small SZA (SZA < ∼40°) the ratios continue to increase as wavelength decreases throughout the UV-B region. For SZA = 60° a local maximum occurs in the UV-B region. As the SZA increases, this turning point moves to longer wavelengths and its peak value decreases. For SZA >∼80°, local minima in the ratios are seen at shorter wavelengths in the UV-B region. For biologically weighted irradiances, the peak ratios occur near SZA = 70°, where UV-A, erythemally weighted UV, UV-B, and DNA-weighted UV irradiances at Mauna Loa Observatory exceeded those at Lauder by ∼17%, 26%, 27%, and 29% respectively. The ratios of irradiances at the two altitudes, as functions of SZA and wavelength, were related to differences expected from radiative transfer calculations. For small SZA, modeled and measured ratios agreed within the limits of experimental uncertainty without taking differences in altitude distributions of ozone and temperature into account. However, for larger SZA and shorter wavelengths these profile shapes had a significant effect. In the model calculations, satisfactory agreement with the measurements was achieved only when the contribution from radiation scattered from air or cloud tops below the observation height at Mauna Loa Observatory was included. To model this accurately, a three-dimensional radiative transfer code should be used in conjunction with a topographical model of the surrounding terrain.