Unlike most brood parasites, several species of cowbird (Molothrus) are generalists that parasitize multiple host species across their range and within the same communities; likewise, there is little evidence that individuals within a population specialize on host species. This situation has variously been attributed to the recency of cowbird evolution (the “evolutionary lag” hypothesis) or to hidden costs of rejection by hosts (the “equilibrium” hypothesis). Both hypotheses have some support as cowbirds are indeed a relatively young clade compared with more specialized cuckoos and cowbirds are capable of sophisticated behaviors such as retaliation against rejection (“mafia” behaviors) that would select for acceptance of cowbird eggs. Nevertheless, many species in the Americas have evolved specialized defenses against cowbird parasitism, almost all of which live in more open habitats (e. g., grasslands, shrublands, riparian strips), which indicates that coevolutionary processes can operate in ways that select for host defenses in spite of cowbird counterdefenses. We propose that the structure of landscapes in North America may explain why forest-nesting birds lack defenses against parasitism and reinforce the long-term maintenance of generalized brood parasitism in cowbirds. Because cowbirds require open habitats in which to feed, they are rare or absent in large forest tracts, which dominate much of the forest cover of the Americas. These tracts act as “source habitats” that produce surplus young that recolonize populations in smaller, fragmented forest patches in which rates of both cowbird parasitism and nest predation are very high (“sink” habitats). Evolution of anti-parasite adaptations would be very slow in this situation because most hosts are produced in areas where there is little or no cowbird parasitism. In addition, the interplay of host breeding dispersal, source-sink metapopulation dynamics, and fragmented forest habitat could further deter the evolution of host defenses against parasitism. Therefore, as long as large forest tracts remain widespread in North America, most forest birds will likely continue to lack defenses against cowbird parasitism, guaranteeing a steady supply of naïve hosts in forest habitats, even in fragmented landscapes. This situation will, in turn, favor host generalist cowbirds that actively avoid more open habitats in favor of parasitizing forest bird communities. These forest communities may also act as source populations for cowbirds, which might pump surplus generalist cowbirds into more open habitats further slowing the coevolutionary process. As long as large forest tracts are a common part of the landscape, generalist parasitism may persist indefinitely.