Ecol. Lett. http://doi.org/f97pdq (2017)

The mechanisms through which communities respond to climate can be complex, making the forecasting of community responses to climate change challenging. Changes in community traits are commonly interpreted as resulting from environmental selection of optimum trait values. However, if the available number and type of trait combinations are constrained by environmental stress — because the cost of adopting extreme trait values is only affordable when trade-offs are made with other traits — this could constrain species diversity.

Addressing this hypothesis, John M. Dwyer from the University of Queensland, Australia, and Daniel C. Laughlin at the University of Waikato, New Zealand, investigate plant communities growing along a hydrological gradient in Western Australia. They assess how covariance among specific leaf area, maximum height, and seed mass changes with aridity, and how community diversity is related to trait covariance.

Commonly used functional diversity measures were poorly explained by aridity, and were surprisingly weak predictors of community richness. Contrastingly, the covariance between maximum height and seed mass strengthened along the aridity gradient, and was strongly associated with declining richness. These findings demonstrate that declines in species richness under environmental stress can be due to increasing constraints on multidimensional phenotypes.