Heterogeneity in leaf tannin inducibility across woody and non-woody plants: A meta-analysis

Tannins are widespread secondary metabolites that can be induced by herbivory or mechanical damage and contribute to defense against herbivores. Reported tannin inducibility varies across taxa and study designs, making it difficult to separate biological drivers from methodological heterogeneity. Because woodiness covaries with longevity and storage and transport capacity, it may shape inducible chemical defenses, but evidence remains mixed. We synthesized 513 effect sizes from 73 studies, quantified tannin inducibility as the log response ratio (lnRR), and assessed phylogenetic signal. Using multilevel meta-analytic models, we tested the effect of woodiness and interactions between woodiness and multiple experimental contexts, and related lnRR to damage intensity, damage duration, and sampling lag. We found that tannin inducibility was positive but highly heterogeneous, and its phylogenetic signal was weak and inconsistent. Woody taxa showed consistently positive and often statistically detectable inducibility across tannin groups, sampling designs, and induction methods, whereas estimates for non-woody taxa were more variable and often overlapped zero. Moreover, lnRR increased with damage intensity in woody plants, whereas it increased with damage duration only in non-woody plants. In conclusion, our synthesis indicates woodiness as an informative axis for leaf tannin inducibility, suggesting that variation in inducible quantitative defenses is not fully explained by phylogenetic relatedness and may reflect broad life history strategy. To move from pattern to mechanism, future studies should pair inducibility estimates with measurements of storage, transport, and allocation traits under standardized damage and sampling protocols, thereby improving comparability across systems and strengthening mechanistic inference.