Genetic variation for settling responses to different types offood and for oviposition behavior exists within natural populations of Drosophila tripunctata. After flies from two isofemale strains and F2 derived from crosses between them were released in the wild and recaptured at either mushroom or tomato baits, the oviposition preferences of females for these foods were determined in the laboratory. Whereas settling behavior and oviposition-site preference were negatively correlated in the parental strains, there was no association of these behaviors in the F2 generation. Furthermore, lines produced by crossing these two strains followed by inbreeding did not manifest such a negative correlation. Together, these results indicate that two important aspects of host selection are under independent genetic control. These findings caution against using a single component of the process of host selection as indicative of the probability that a particular resource will be utilized in the field, and they bear on the question of how speciation based on genetic variation for host selection may be accomplished.Speciation is of central importance in the evolution of biological diversity, and it has, therefore, been a major focus of attention since the time of Darwin (1). However, no consensus of opinion on how new species arise has emerged; thus, recent reviews concerned with the relation between population structure and speciation have promoted radically different hypotheses (2-5). One reason for the wavering direction of speciation theory is the virtual absence of information on the nature ofintraspecific genetic variation for the types of structures, behaviors, and physiologies that would seem to be important in the development of reproductive isolation. A notable and promising exception is the work of Carson and his associates on the mode of inheritance of tibial bristles that characterize males of some populations of Drosophila silvestris and that are thought to be used during courtship (6-8).In many groups of insects, speciation is commonly accompanied by, and may, in fact, be a consequence of, shifts to new larval feeding resources (hosts) (9-13). The traits that bring about such host shifts are believed to be those governing host selection by adults, larval development on alternative resources, and mate preference (13,14). If the host also serves as a rendezvous for mating, as it does for many species (13), then the probability of mating with an individual of a particular genotype will be a pleiotropic consequence of host selection. Furthermore, recent evidence suggests that evolutionary shifts to new hosts may be due more to behavioral than to physiological changes (11). Thus, evolutionary changes in host selection may play a major role in speciation in insects (15). However, essentially nothing is known of the genetics of this process in any insects.All theoretical models of host-related speciation are based on the assumption that, within species of insects, potential hosts are ranked by an individual alo...