Abstract
We have previously analysed the genetic architecture of host-associated performance differences between stocks of the brown planthopper, Nilaparvata lugens (Homoptera: Delphacidae), derived from two host plants: cultivated rice variety TN1 and the semi-aquatic weed Leersia hexandra. It has been established that performance is influenced by a small effective number of loci (1 to 3), with dominance in the direction of the rice population. The results reported here show that, like performance, oviposition preference has a simple genetic basis, but in this case there is no evidence of directional dominance, sex linkage, non-allelic interaction or genotype–environment interaction. The simple genetic architecture of host-associated performance and oviposition preference are as expected if there was a sympatric host shift, most probably from Leersia to rice. However, contrary to the theoretical expectation from models of sympatric host-race formation, there was no detectable genetic association between the traits. The fact that individual females tend to distribute their eggs between plants may have helped to promote a host shift despite this lack of association.