In this paper we discuss and evaluate the systematic sources of bias in aerosol optical depth (AOD) values in the UV range due to (1) the entrance of diffuse light into the finite field of view, (2) diurnal atmospheric changes of ozone under urban conditions, (3) the influence of omitting the effect of NO2 absorption, and (4) stray light of a single monochromator. These error sources have been neglected before in Brewer AOD retrieval. However, if these bias estimates are added together, it appears likely that the main reason for the recent results, that is, an AOD wavelength dependency that is in contradiction to the ¿ngstrom law, lies in the omitted sources of systematic error in the AOD retrieval. For instance, the estimated negative bias in AOD difference between 306.3 and 320.1 nm is ~0.0772, while between 310.1 and 320.1 nm it is ~0.0346. If the true ¿ngstrom α was 1 and the AOD at 320.1 was 0.5, then the actual difference between 306.3 and 320.1 nm would be positive and equal to 0.022, while between 310.1 and 320.1 nm it would be 0.016. Therefore the neglected source of bias can mask this difference and result in a negative ¿ngstrom exponent (AE) value. In addition to these sources of bias, we also discuss other potential sources of uncertainty that have been previously neglected.