The spatial distribution of Mussidia nigrivenella Ragonot was studied in maize fields and on four major wild host plants in the Republic of Benin, West Africa. Maize ears were sampled at harvest in 1994 and 1995, and fruits of the wild host plants were collected monthly from January 1996 to December 1997 during the fruiting periods of the respective host species. The spatial distribution was analyzed using the variance-to-mean ratio (s²/m), an index of departure from Poisson distribution, Iwao’s patchiness regression, and Taylor’s power law. The variance-to-mean ratio and the index of departure from Poisson gave similar results of nonrandom distribution of M. nigrivenella populations on the different host plants out of Gardenia sokotensis Hutch (Rubiaceae). On this specific host plant, the variance-to-mean ratio was not different from 1, whereas there was no overlap of the observed population and the Poisson distribution. Iwao’s patchiness regression was inappropriate for our data as shown by the nonhomogeneity of variance, whereas Taylor’s power law fitted the data well. Based on Taylor’s power law, M. nigrivenella was aggregated on all host plants out of G. sokotensis. The degree of aggregation of M. nigrivenella was positively correlated with the fruit size of the host plants. The optimal number of secondary sampling units needed to estimate M. nigrivenella densities on the respective host plants in Benin, varied from four fruits on G. sokotensis to 10 pods on P. biglobosa. These values were used to calculate the time expenditure needed to sample maize fields or wild host plant trees to achieve a predefined precision level of 25%.