Full text: Download
Spatial variation in plant chemical defence towards herbivores can help us understand variation in herbivore top–down control of shrubs in the Arctic and possibly also shrub responses to global warming. Less defended, non‐resinous shrubs could be more influenced by herbivores than more defended, resinous shrubs. However, sparse field measurements limit our current understanding of how much of the circum‐Arctic variation in defence compounds is explained by taxa or defence functional groups (resinous/non‐resinous). We measured circum‐Arctic chemical defence and leaf digestibility in resinous (Betula glandulosa, B. nana ssp. exilis) and non‐resinous (B. nana ssp. nana, B. pumila) shrub birches to see how they vary among and within taxa and functional groups. Using liquid chromatography–mass spectrometry (LC–MS) metabolomic analyses and in vitro leaf digestibility via incubation in cattle rumen fluid, we analysed defence composition and leaf digestibility in 128 samples from 44 tundra locations.We found biogeographical patterns in anti‐herbivore defence where mean leaf triterpene concentrations and twig resin gland density were greater in resinous taxa and mean concentrations of condensing tannins were greater in non‐resinous taxa. This indicates a biome‐wide trade‐off between triterpene‐ or tannin‐dominated defences. However, we also found variations in chemical defence composition and resin gland density both within and among functional groups (resinous/non‐resinous) and taxa, suggesting these categorisations only partly predict chemical herbivore defence. Complex tannins were the only defence compounds negatively related to in vitro digestibility, identifying this previously neglected tannin group as having a potential key role in birch anti‐herbivore defence.We conclude that circum‐Arctic variation in birch anti‐herbivore defence can be partly derived from biogeographical distributions of birch taxa, although our detailed mapping of plant defence provides more information on this variation and can be used for better predictions of herbivore effects on Arctic vegetation.