Testing general predictions of the stress gradient hypothesis under high inter‐ and intra‐specific nurse shrub variability along a climatic gradient

Question Could inter‐ and intra‐specific shrub variability override abiotic stress in driving shrub‐grass interactions along an altitudinal gradient? Location Eleven sites along an altitudinal gradient in the mediterranean‐temperate transition of the central western Iberian Peninsula ( S alamanca province, S pain). Methods Among study sites, different shrub species account for inter‐specific variability. At each site, three patches differing in age of dominant shrubs were established to represent intra‐specific variability. We established eight 25 × 25 cm sampling quadrants within and beyond the influence of shrubs, and surveyed abundance of a mesophytic perennial grass ( A grostis castellana ) in each patch. As a surrogate of the balance in shrub‐grass interaction, we computed RII (relative intensity of interaction). Using PCA we reduced long‐term climatic data, and used the first axis as a proxy for climatic gradient. We modelled the abundance of A . castellana and RII by applying a type I likelihood ratio test (generalized linear model) to the nested design. Results PCA ordination showed a principal climatic gradient associated with altitude (increasing humid/colder towards higher altitude). Beyond the influence of shrubs, abundance of A . castellana among sites had a hump‐shaped relationship to the climatic gradient. Age of patches did not affect abundance of A . castellana in open habitats. Beneath shrubs, there were no significant effects of nurse species on abundance of A . castellana , nor any relationship to the climatic gradient. The balance of the interaction revealed significant differences among sites, reflecting net positive interactions at both climatic extremes of the gradient, and net negative interactions at intermediate climatic conditions. Shrub age had a significant effect on the balance of the interaction because of generalized neutral/positive, but not negative, effects of younger shrubs (7–10 years old) along the climatic gradient. Significant impacts of nurse shrub inter‐specific variability were associated with consistent positive effects of C istus ladanifer on A . castellana abundance. Conclusions Inter‐ and intra‐specific shrub variability does not markedly affect general predictions of the stress gradient hypothesis. Hence, abiotic stresses were pivotal in driving the balance of the interaction between A . castellana and associated shrubs along the altitudinal gradient. The interaction shifted from negative to positive where there was a departure from the climatic optimum for A . castellana, irrespective of species or age of shrubs.