To analyse induced glucosinolates (GSLs) upon root herbivory at different root sections, we conducted a pot experiment using Anomala cuprea grubs and four Brassicaceae; Brassica rapa, B. nigra, B. oleracea, and Sinapis alba. Individuals of these four plant species were grown in dedicated mesocosms. They received the following root herbivory treatments; two grubs confined either to the bottom soil, or to the middle section, or to the top soil. Plants grown in the same set-ups but without root herbivores served as controls. Glucosinolate levels of the leaf lamina, petiole, and stem as well as of the taproot, lateral roots, and fine roots were measured after eight days of herbivory.
 
Plant biomass reduction due to herbivory was the largest when herbivores were confined to the top soil. In the three Brassica species, taproot GSL levels increased upon herbivory independent where the root herbivores were feeding. Glucosinolate levels in fine roots and shoots, on the other hand, rarely responded to root herbivory. Indole GSLs, which are more effective to pathogens than to herbivores, were induced stronger than aliphatic and aromatic GSLs. Sinapis alba did not show remarkable increments in any GSL level upon herbivory.
 
These results show that locally and systemically induced defences in roots are consistent with the optimal defence theory: the taproot which is the most vulnerable and valuable to plant performance shows the highest increase in defence induction. The induced GSL profiles suggest that the response may not only target herbivores, but may also help to prevent secondary infection by microbial pathogens.