Search

Aims: Elevational patterns of plant species richness may be caused by multiple underlying mechanisms, and the same pattern can be predicted by different mechanisms. Using the steep elevational gradient of Tenerife as a model system, we aimed to test if the application of a trait‐based approach can help disentangle the role of potential mechanisms behind local elevational plant species richness patterns. Location: Tenerife, Canary Islands, Spain. Methods: Based on vegetation relevés from natural vegetation and disturbed roadside habitat, along an elevational gradient of 2300 m, we observed a peak of plant species richness in the lowest third of the gradient. We considered three mechanisms potentially shaping this pattern: environmental filtering (temperature and precipitation), effects of area and disturbance. For these mechanisms, we hypothesized a distinct pattern of functional trait–elevation relationships. These were tested with in‐situ data of nine functional leaf traits, from which we calculated community‐weighted means (CWM) of traits and functional diversity (Rao's Q). Results: While species richness was significantly positively correlated with temperature, area and disturbance, filtering through temperature was the only mechanism for which we could confirm most of our mechanism‐specific hypotheses about elevational trait changes: with increasing elevation, CWMs of most traits indicated shifts from acquisitive to conservative growth strategies, and functional diversity decreased. The shift of growth strategies also supported the disturbance effect, as we found overall more acquisitive communities at roadsides compared to natural habitats. Conclusions: Our results indicate that simple correlations between species richness and abiotic variables are not necessarily causal. Additional testing of mechanism‐specific hypotheses for elevational patterns of both CWMs and functional diversity can help distinguishing between correlational and mechanistic relationships between species richness and environmental variables. The trait‐based framework presented here can be fruitfully applied to better understand species richness patterns in other regions and across other types of environmental gradients. [ABSTRACT FROM AUTHOR]

A non-native plant species spreading along an environmental gradient may need to adjust its growth to the prevailing conditions that it encounters by a combination of phenotypic plasticity and genetic adaptation. There have been several studies of how non-native species respond to changing environmental conditions along latitudinal gradients, but much less is known about elevational gradients. We conducted a climate chamber experiment to investigate plastic and genetically based growth responses of 13 herbaceous non-native plants along an elevational gradient from 100 to 2,000 m a.s.l. in Tenerife. Conditions in the field ranged from high anthropogenic disturbance but generally favourable temperatures for plant growth in the lower half of the gradient, to low disturbance but much cooler conditions in the upper half. We collected seed from low, mid and high elevations and grew them in climate chambers under the characteristic temperatures at these three elevations. Growth of all species was reduced under lower temperatures along both halves of the gradient. We found consistent genetically based differences in growth over the upper elevational gradient, with plants from high-elevation sites growing more slowly than those from mid-elevation ones, while the pattern in the lower part of the gradient was more mixed. Our data suggest that many non-native plants might respond to climate along elevational gradients by genetically based changes in key traits, especially at higher elevations where low temperatures probably impose a stronger selection pressure. At lower elevations, where anthropogenic influences are greater, higher gene flow and frequent disturbance might favour genotypes with broad ecological amplitudes. Thus the importance of evolutionary processes for invasion success is likely to be context-dependent.