The lupus band test is widely regarded as a sensitive test in the diagnosis of lupus erythematosus, particularly as applied to cutaneous lesions. It is commonly said to be positive in 90% or more of lesions in discoid and systemic lupus. Other criteria for diagnosis are often not specified in previous studies. This and the subjective element of immunofluorescence microscopy raise some doubts about the sensitivity of the test for skin lesions. In this comparative light microscopy and immunofluorescence study, the lupus band test was positive in only 67% of discoid lesions and 73% of lesions in systemic lupus. The test was occasionally positive where light microscopy was inconclusive and was positive in 82% of systemic lupus patients' uninvolved skin. IgM was an infrequent reactant in the lupus band. The significance of these findings is discussed.