Evolution of plant defences along an invasion chronosequence: defence is lost due to enemy release – but not forever

The success of invasive plants has often been attributed to their rapid evolution at the introduced range. In particular, release from native enemies has been suggested to select for an evolutionary shift in resource allocation patterns from herbivore defence to increased size. Such evolutionary processes can take place not only between the native and invasive ranges but also within the invasive range over time, but this premise has been very seldom studied. In this study, we examined the potential for post-invasion evolution in two traits hypothesized to facilitate plant invasion success, that is herbivore resistance and allelopathic ability. We studied these traits in the invasive plant Impatiens glandulifera by comparing plants from its native populations and from populations across its invasion chronosequence. Results of common-garden experiment and chemical analyses revealed that plants from native populations or older populations within the invasive range show greater resistance to the generalist herbivore, Deilephila elpenor, coupled with greater production of the secondary defence compound 2-methoxy-1,4-naphthoquinone glycoside. In contrast, no differences were found between populations in their allelopathic effect on the germination of the co-occurring neighbour, Urtica doica. Finally, results from a field survey suggested that older populations within the invasive range incur greater attack rates from local herbivores compared to more recently established populations. Synthesis. Our findings support the idea that the selection pressure of enemy release at the introduced range might attenuate over time, leading to the evolutionary recovery of enemy resistance. This study emphasizes the importance of incorporating the effect of time since introduction when examining evolutionary or ecological processes of plant invasions.