To determine whether plant genotype can affect long-term herbivore populati
on dynamics, we used a density manipulation experiment to compare the dynam
ics of Mexican bean beetle populations living on four genotypes of soybeans
. Five densities of beetle populations were created on each plant genotype.
Each combination of density and genotype was replicated four times, yieldi
ng 80 experimental beetle populations. Populations were initiated by releas
ing adult beetles into 2.25 m(2) cages filled with soybeans and located in
a small soybean field. Beetle populations were censused weekly for one gene
ration. Recruitment curves for beetle populations on each soybean genotype
were plotted by fitting a Picker model to the data using likelihood methods
. The recruitment curves for all genotypes were significantly different fro
m the curves for all other genotypes. The equilibrium beetle density predic
ted from the curves varied among genotypes by nearly twofold, and the predi
cted dynamics were fluctuating for all genotypes, with the period of fluctu
ations differing among genotypes. Performance of beetles also differed amon
g genotypes, suggesting that these genotypes vary in some aspect of their q
uality as hosts for the beetles. These results indicate that differences in
quality among genotypes have the potential to affect long-term herbivore d
ynamics.