Chimpanzees are the closest evolutionary cousins of humans, sharing >99% identity in most protein sequences. Plasmodium falciparum is the major worldwide cause of malaria mortality. Plasmodium reichenowi, a morphologically identical and genetically very similar parasite, infects chimpanzees but not humans. Conversely, experimental P. falciparum infection causes brief moderate parasitization and no severe infection in chimpanzees. This surprising host specificity remains unexplained. We modified and enhanced traditional methods for measuring sialic acid (Sia)-dependent recognition of glycophorins by merozoite erythrocyte-binding proteins, eliminating interference caused by endogenous Sias on transfected cells, and by using erythroleukemia cells to allow experimental manipulation of Sia content. We present evidence that these remarkable differences among such closely related host-parasite pairs is caused by species-specific erythrocyte-recognition profiles, apparently related to the human-specific loss of the common primate Sia N-glycolylneuraminic acid. The major merozoite-binding protein erythrocyte-binding antigen-175 of P. falciparum apparently evolved to take selective advantage of the excess of the Sia N-acetylneuraminic acid (the precursor of N-glycolylneuraminic acid) on human erythrocytes. The contrasting preference of P. reichenowi erythrocyte-binding antigen-175 for N-glycolylneuraminic acid is likely the ancestral condition. The surprising ability of P. falciparum to cause disease in New World Aotus monkeys (geographically isolated from P. falciparum until arrival of peoples from the Old World) can be explained by parallel evolution of a human-like Sia expression pattern in these distantly related primates. These results also have implications for the prehistory of hominids and for the genetic origins and recent emergence of P. falciparum as a major human pathogen.