Pairwise trade-offs among the main dimensions of life history of plants, namely growth, reproduction and defence, have been hypothesized and repeatedly tested. Experimental evidences, however, are far from conclusive, probably as a consequence of different methodological approaches, but also of an underlying multivariate complexity that is not considered in pairwise analyses. We carried out two long-term field experiments with 81 ramets clonally replicated from 6 genotypes of maritime pine (Pinus pinaster) to assess the allocation to growth, reproduction and defence in response to increased resource availability (fertilization and control treatments) and simulated herbivory (low intensity mechanical wounding, high intensity mechanical wounding, needle clipping, jasmonate application and control treatments). Fertilization increased tree growth and N and P concentration, and altered the production of resin and polyphenolics in complex ways, contingent on time and tissue. A natural event of infestation by Pissodes castaneus was more intense on fertilized trees. Fertilization increased both reproductive effort and output in terms of quantity (number of seeds) but not quality (dry weight of seeds). Needle clipping and methyl-jasmonate, but not mechanical wounding, increased the concentration of needle resin and phenols shortly after treatment, while concentrations in the phloem remain unaltered. Damage by P. castaneus did not differ across induction treatments. Generally, needle clipping had the strongest effect on reproductive variables, and mechanical wounding the mildest, with a high variability on intensity and direction of effects. Needle clipping decreased tree growth, while the rest of herbivory treatments did increase it. Complex interactions among life-history dimensions arose, showing that pairwise approaches are too simplistic for unravelling these complexities. This three-way relationship imposes new environmental and physiological constraints to the evolution and maintenance of reproductive strategies.

The authors thank the support from the Spanish Ministry of Economy, Industry and Competitiveness (grants AGL2012-4051-C03-01 and AGL2015–68274-C3-2-R), Ministry of Science, Innovation and Universities (grant RTI2018-094691-B-C31) and Spanish Research Council (CSIC; grant AEPP-2022AEP013). The European Union also supported this work, by means of its Structural Fund FEDER.

Peer reviewed