Summary 1. Root herbivores and pathogens interfere with basic below‐ground plant function, and can thereby affect plant fitness and spatial and temporal patterns in natural plant communities. However, there has been little development of concepts and theories on below‐ground plant defence, a deficit that is in contrast to the abundance of theorizing for above‐ground plant parts. 2. A review of the past 10 years of research on below‐ground plant–herbivore interactions has revealed that, similar to above‐ground tissues, root defences can be expressed constitutively or induced upon herbivore attack, and can be classified into direct and indirect traits, tolerance, and escape. Indeed, it has been shown that roots tolerate herbivory by outgrowing or re‐growing lost tissues, or resist it by producing secondary metabolites that are toxic to herbivores or attract natural enemies of herbivores. 3. We propose that, similar to above‐ground plant–herbivore theories, the partition of abiotic and biotic factors over ecological succession can serve as the basis for predicting investment in defence strategies below‐ground. 4. Investigation of herbivore pressure and root responses along primary and secondary successional gradients suggests that: (i) roots are often fast growing, thinner and softer in early compared to later succession. (ii) Insect and nematode herbivore pressure increases until mid‐succession and later decreases. (iii) Mycorrhizal abundance increases with succession, and the composition of fungal species changes through succession, often shifting from arbuscular mycorrhizae to ecto‐mycorrhizae. 5. Based on these findings, and on classical (above‐ground) plant defence theory, we suggest the following set of testable hypotheses for below‐ground plant defence: (i) During succession, early plants invest most of their resources in growth and less in defences (associated with a general lack of herbivores and pathogens, and with limited availability of resources in the system), therefore relying more on re‐growth (tolerance) strategies. (ii) During mid‐succession, a buildup of herbivore pressure facilitates replacement by plant species that exhibit greater direct and indirect defence strategies. (iii) Constitutive and inducible levels of defences may trade‐off, and early successional plants should rely more on induction of defences after herbivore attack, whereas late successional plants will increasingly rely on constitutively produced levels of physical and chemical defence. (iv) Successional changes in microbial associations have consequences for root defence by improving plant nutrition and defence expression as well as directly competing for root space; however, toxic or impenetrable root defences may also limit association with root symbionts, and so may constrain the expression of root defence.