The parsnip webworm (Depressaria pastinacella) and the wild parsnip (Pastinaca sativa) together represent a potentially “coevolved” system in that throughout their ranges the plant has relatively few other herbivores and the insect has virtually no other hosts. Individual wild parsnip plants within a central Illinois population vary in their content and composition of furanocoumarins, secondary compounds with insecticidal properties. Half‐sib and parent‐offspring regression estimates of the heritability of furanocoumarins demonstrate that this variation is genetically based. Wild parsnip plants also vary in their resistance to damage by the parsnip webworm, which feeds on flowers and developing seeds. In an experimental garden, seed production in the primary umbel ranged from 0 to 1,664 seeds among individuals, and mean seed production of half‐sib families ranged from 3.7 seeds to 446.0 seeds. Approximately 75% of the variation in resistance among half‐sib families to D. pastinacella was attributable to four furanocoumarin characteristics—resistance is positively related to the proportion of bergapten and the amount of sphondin in seeds, and negatively related to the amount of bergapten and the proportion of sphondin in leaves. Each of the four resistance factors had significant heritability. Thus, resistance in wild parsnip to the parsnip webworm is to a large extent chemically based and genetically controlled. Genetic correlations among fitness and resistance characters, however, tend to limit coevolutionary responses between herbivore and plant. In greenhouse plants protected from herbivory, several of the resistance factors have negative genetic correlations with potential seed production. Ostensibly, highly resistant plants in the absence of herbivory would be at a competitive disadvantage in the field. The selective impact of the herbivore is also limited in this population by a negative genetic correlation among resistance factors. Selection to increase one resistance factor (e.g., the proportion of bergapten in the seed) would at the same time decrease the amount of a second resistance factor (e.g., the amount of sphondin in the seed). The wild parsnip and the parsnip webworm, then, appear to have reached an evolutionary “stalemate” in the coevolutionary arms race.