The probability of attack and patterns of constitutive and induced defense: A test of optimal defense theory

Optimal defense theory predicts that tissues that are unlikely to be attacked by herbivores should have low constitutive amounts of defense and high inducibility, while tissues that are likely to be attacked should have high levels of constitutive defense and low inducibility. We artificially damaged roots, leaves, and reproductive parts of wild parsnip (Pastinaca sativa L.) and found that these parts differ not only in constitutive levels of chemical defense but also in the degree to which these defenses are inducible. Reproductive parts contain the highest constitutive concentrations of a toxic furanocoumarin, xanthotoxin, but xanthotoxin production is not inducible in these parts. Roots contain the lowest constitutive levels of xanthotoxin and are highly inducible. Leaves are intermediate in both the constitutive amounts of xanthotoxin they contain and inducibility. To test the prediction of optimal defense theory, we surveyed three populations of wild parsnip for the presence of damage to roots, leaves, and reproductive parts in the summers of 1993 and 1994. The probabilities of attack of these organs were remarkably consistent among populations and years. Roots were least likely to be attacked, while reproductive parts and leaves had high probabilities of attack. Although reproductive parts and leaves were equally likely to suffer damage, reproductive parts were far more extensively damaged, Thus, relationships between patterns of defense and attack were consistent with predictions based on optimal defense theory.