Plant resistance to the feeding by herbivorous insects has recently been found to be positively or negatively influenced by prior egg deposition. Here we show how crucial it is to conduct experiments on plant responses to herbivory under conditions that simulate natural insect behaviour. We used a well- studied plant – herbivore system, Arabidopsis thaliana and the cabbage white butterfly Pieris brassicae, testing the effects of naturally laid eggs (rather than egg extracts) and allowing larvae to feed gregariously as they do naturally (rather than placing single larvae on plants). Under natural conditions, newly hatched larvae start feeding on their egg shells before they consume leaf tissue, but access to egg shells had no effect on subsequent larval performance in our experiments. However, young larvae feeding gregariously on leaves previously laden with eggs caused less feeding damage, gained less weight during the first 2 days, and suffered twice as high a mortality until pupation compared to larvae feeding on plants that had never had eggs. The concentration of the major anti-herbivore defences of A. thaliana, the glucosinolates, was not significantly increased by oviposition, but the amount of the most abundant member of this class, 4-methylsulfinylbutyl glucosinolate was 1.8-fold lower in larval-damaged leaves with prior egg deposition compared to damaged leaves that had never had eggs. There were also few significant changes in the transcript levels of glucosinolate metabolic genes, except that egg deposition suppressed the feeding-induced up-regulation of FMOGS-OX2, a gene encoding a flavin monooxygenase involved in the last step of 4-methylsulfinylbutyl glucosinolate biosynthesis. Hence, our study demonstrates that oviposition does increase A. thaliana resistance to feeding by subsequently hatching larvae, but this cannot be attributed simply to changes in glucosinolate content.
