We studied the point at which a monoclonal antibody (mAb C5) to a surface protein (Pgs25) on Plasmodium gallinaceum ookinetes blocked the infection of Aedes aegypti mosquitoes. The antibody did not block the development of zygotes to ookinetes in vitro. Development of ookinetes to oocysts in the mosquito was blocked to the same extent whether zygotes grew to ookinetes in the presence of mAb C5 or the antibody was added after the ookinetes had reached full development. When ookinetes developed in vitro in the presence of mAb C5, antibody remained on the surface of the parasite for the next 50 hr and did not block attachment to the peritrophic membrane. When ookinetes were fed to mosquitoes, two subpopulations of mosquitoes were observed (high numbers of oocysts per midgut and low numbers of oocysts per midgut). mAb C5 reduced the number of oocysts per midgut in the subpopulation that had low numbers of oocysts. The subpopulation that had high numbers of oocysts was unaffected by antibody, indicating that the antibody did not block invasion of the midgut epithelium. When mAb C5 was fed with gametocytes, the parasites invaded the epithelium at the same time (between 30 and 35 hr after the blood meal) as in controls, although at a markedly reduced rate. The ultrastructural observations were consistent with a block of parasites within the peritrophic membrane and not with a block at the epithelium, as parasites were not seen to accumulate within the space between the peritrophic membrane and the epithelium. The mechanism by which mAb C5 to Pgs25 of P. gallinaceum blocks the penetration of the peritrophic membrane remains undefined. We present evidence that the parasite modifies the peritrophic membrane during penetration, an observation first made for Babesia microti during penetration of the peritrophic membrane in Ixodes ticks. Ookinetes in the absence of antibodies appeared to disrupt the layers of the peritrophic membrane, suggesting an enzymatic mechanism for penetration.