The relation between the number of microfilariae (mf) ingested by host-seeking vectors of human onchocerciasis and skin mf load is an important component of the population biology of Onchocerca volvulus, with implications for disease control and evaluation of the risk of transmission recrudescence. The microsimulation model ONCHOSIM has been used to assess such risk in the area of the Onchocerciasis Control Program (OCP) in West Africa, based on a strongly nonlinear relation between vector mf uptake and human mf skin density previously published. However, observed levels of recrudescence have exceeded predictions, warranting a recalibration of the model. To this end, we present the results of a series of fly-feeding experiments carried out in savanna and forest localities of West Africa. Flies belonging to Simulium damnosum s.s., S. sirbanum, S. soubrense, and S. leonense were fed on mf carriers and dissected to assess the number of ingested mf escaping imprisonment by the peritrophic matrix (the number of exo-peritrophic mf), a predictor of infective larval output. The method of instrumental variables was used to obtain (nearly) unbiased estimates of the parameters of interest, taking into account error in the measurement of skin mf density. This error is often neglected in these types of studies, making it difficult to ascertain the degree of density-dependence truly present in the relation between mf uptake and skin load. We conclude that this relation is weakly (yet significantly) nonlinear in savanna settings but indistinguishable from linearity in forest vectors. Exo-peritrophic mf uptake does not account for most of the density dependence in the transmission dynamics of the parasite as previously thought. The number of exo-mf in forest simuliids is at least five times higher than in the savanna vectors. Parasite abundance in human onchocerciasis is regulated by poorly known mechanisms operating mainly on other stages of the lifecycle.