Search

We tested whether differences in ploidy level and previous exposure to herbivory can affect plant tolerance to herbivory. We conducted a common garden experiment with 12 populations of two ploidy levels of the perennial herb Cardamine pratensis (five populations of tetraploid ssp. pratensis and seven populations of octoploid ssp. paludosa). Earlier studies have shown that attack rates by the main herbivore, the orange tip butterfly Anthocharis cardamines, are lower in populations of octoploids than in populations of tetraploids, and vary among populations. In the common garden experiment, a combination of natural and artificial damage significantly reduced seed and flower production. We measured tolerance based on four plant-performance metrics: survival, growth, seed production and clonal reproduction. For three of these measurements, tolerance of damage did not differ between ploidy levels. For clonal reproduction, the octoploids had a higher tolerance than the tetraploids, although they experience lower herbivore attack rates in natural populations. Populations from sites with high levels of herbivory had higher tolerance, measured by seed production, than populations with low levels of herbivory. We did not detect any significant costs of tolerance. We conclude that high intensity of herbivory has selected for high tolerance measured by seed production in C. pratensis.

Many plants show compensatory regrowth after herbivory and dormant buds often have an important role in compensatory responses. Theoretical models have shown that herbivore damage may select for a bud bank, i.e., a pool of dormant buds that are protected from herbivory and that are activated after herbivore damage. Earlier models assumed that undamaged plants cannot activate their dormant buds without damage, although they apparently have sufficient resources for successful seed production through the additional shoots dormant buds could produce. However, many plants are able to gradually activate buds over an extended period of time without any cue from damage. The aim of this study was to analyze how herbivory imposes selection for gradual mobilization of the bud bank. I assume that selection pressures that affect the fraction of buds active at each time point include damage by herbivores, time left to the end of season, and the opportunity costs of dormant buds. I modelled bud dynamics with gradual activation when there is a single damage event and (i) when the seed set of a shoot is not dependent on the time it is active, or (ii) when the seed set of a shoot diminishes with later activation. In addition, I analyzed how (iii) risk of repeated herbivory affects selection for gradual activation. Under these models, gradual activation is optimal over a wide range of herbivory pressures. Selection appears to favour activation of all buds at the beginning of the season only when herbivore pressure is weak and when early shoots have a higher seed set than late shoots. Alternatively, strong herbivore pressure and late damage may select for a large bud bank throughout the growing season, without gradual activation; the bud bank is only mobilized after damage. In this case, damaged plants can overcompensate, i.e. they have a higher seed set than undamaged plants with the same bud activation pattern. Selection for overcompensation demands a stronger herbivore pressure in this current model than in earlier bud bank models. The model never predicts selection for overcompensation when there is a risk of repeated herbivory.