• Most plants suffer some degree of herbivore attack and many actively defend themselves against such an event. However, while such defence is generally assumed to be costly, it has sometimes proved difficult to demonstrate the costs of defensive compounds.
• Here, we present a method for analysing growth rates which allows the effects of variation in initial plant size to be properly accounted for and apply it to 30 lines
from a recombinant inbred population of Arabidopsis thaliana. We then relate different measures of relative growth rate (RGR) to damage caused by a specialist
lepidopteran insect and to levels of putative defensive compounds measured on the same lines.
• We show that seed size variation within the recombinant inbred population is large enough to generate differences in RGR, even when no other physiological differences exist. However, once size-standardized, RGR was positively correlated with herbivore damage (fast-growing lines suffered more damage) and was negatively correlated with the concentration of several glucosinolate compounds.
• We conclude that defensive compounds do have a growth cost and that the production of such compounds results in reduced herbivore damage. However, size standardization of RGR was essential to uncovering the growth costs of