When females of Drosophila melanogaster are treated with chemical or physical mutagens, not only in one but also in both of the two homologous X chromosomes of a given oocyte, a recessive sex-linked lethal mutation may be induced. A method is described that discriminates between such "single" and "double mutations". A theory is developed to show how a comparison between the expected and the observed frequency of double mutations yields an indication of the intercellular distribution (random or nonrandom) of recessive lethal mutations induced by mutagenic agents in oocytes and, consequently, of the distribution (homogeneous or nonhomogeneous) of those agents.--Three agents were tested: FUdR (12.5, 50.5 and 81.0 micrograms/ml), mitomycin C (130.0 micrograms/ml) and X rays (2000 R, 150 kV). After FUdR feeding, no increase in the mutation frequency usually observed in D. melanogaster without mutagenic treatment was obtained (u = 0.13%, namely three single mutations among 2332 chromosomes tested). After mitomycin C feeding, 104 single and three double mutations were obtained. All of the 50 mutations observed after X irradiation were single mutations. The results obtained in the mitomycin C and radiation experiments favor the assumption of a random intercellular distribution of recessive lethal mutations induced by these two agents in oocytes of D. melanogaster. Reasons are discussed why for other types of mutagenic agents nonrandom distributions may be observed with our technique.