Recent research suggests that plant viruses, and other pathogens, frequently alter host-plant phenotypes in ways that facilitate transmission by arthropod vectors. However, many viruses infect multiple hosts, raising questions about whether these pathogens are capable of inducing transmission-facilitating phenotypes in phylogenetically divergent host plants and the extent to which evolutionary history with a given host or plant community influences such effects. To explore these issues, we worked with two newly acquired field isolates of cucumber mosaic virus (CMV)-a widespread multi-host plant pathogen transmitted in a non-persistent manner by aphids-and explored effects on the phenotypes of different host plants and on their subsequent interactions with aphid vectors. An isolate collected from cultivated squash fields (KVPG2-CMV) induced in the native squash host (Cucurbita pepo) a suite of effects on host-vector interactions suggested by previous work to be conducive to transmission (including reduced host-plant quality for aphids, rapid aphid dispersal from infected to healthy plants, and enhanced aphid attraction to the elevated emission of a volatile blend similar to that of healthy plants). A second isolate (P1-CMV) collected from cultivated pepper (Capsicum annuum) induced more neutral effects in its native host (largely exhibiting non-significant trends in the direction of effects seen for KVPG2-CMV in squash). When we attempted cross-host inoculations of these two CMV isolates (KVPG2-CMV in pepper and P1-CMV in squash), P1-CMV was only sporadically able to infect the novel host; KVPG2-CMV infected the novel pepper host with somewhat reduced success compared with its native host and reached virus titers significantly lower than those observed for either strain in its native host. Furthermore, KVPG2-CMV induced changes in the phenotype of the novel host, and consequently in host-vector interactions, dramatically different than those observed in the native host and apparently maladaptive with respect to virus transmission (e.g., host plant quality for aphids was significantly improved in this instance, and aphid dispersal was reduced). Taken together, these findings provide evidence of adaption by CMV to local hosts (including reduced infectivity and replication in novel versus native hosts) and further suggest that such adaptation may extend to effects on host-plant traits mediating interactions with aphid vectors. Thus, these results are consistent with the hypothesis that virus effects on host-vector interactions can be adaptive, and they suggest that multi-host pathogens may exhibit adaptation with respect to these and other effects on host phenotypes, perhaps especially in homogeneous monocultures.