Random amplified polymorphic DNA (RAPD) markers were used to examine population genetic structure in populations of native grape phylloxera. This research asked: (i) do RAPD markers distinguish two groups corresponding to the two host plant species; and (ii) do RAPD markers distinguish groups according to spatial location, independent of host plant association? Forty-nine phylloxera clones were collected from five pairs of adjacent individuals of two sympatric grape species in five sites along a 145 km transect in Missouri, USA. A high level of polymorphism was observed, with some evidence for structuring between host plant species and no evidence for spatial structuring. An analysis of molecular variance (amova) found that 6.52% of the variance in RAPD banding patterns was attributable to host species and 7.96% of the variance was attributable to spatial location. A cluster analysis did not result in two groups corresponding to the two hosts, or to five groups corresponding to the geographical sites sampled. A Mantel test showed a low correlation between genetic similarity and spatial location. Two of the 93 RAPD markers were nonrandomly associated between the hosts. It is suggested that there may be a small host-mediated effect on genetic variation but stochastic dispersal and a highly heterogeneous environment may be the primary influences on the observed polymorphism.