Significant amounts of endogenous biotin were detected by avidin-peroxidase in fixed rat kidney, liver, and brain. The staining was indistinguishable from the true signals of immunoreactivity and could not be consistently blocked by pretreatment with avidin. The finding that certain neurons in the hippocampus contain more biotin than neurons in other areas of the brain suggests that biotin might have novel functions in the brain other than its well-known role as cofactor of carboxylases. Critical examination of published immunohistochemical localization studies on rat kidney strongly suggests that many false-positive results have been considered as true signals. Interference of endogenous biotin in any study using avidin-biotin technology must be considered if biological tissues are involved. The published data obtained by this method should therefore be reevaluated. Furthermore, appropriate controls, blockers and caution in interpreting results must be exercised, not only in immunohistochemistry but also in any applications of avidin-biotin technology.