Meta-analyses of reproductive changes in angiosperm populations in response to elevation reveal a lack of global patterns.

BACKGROUND AND

aimsElevation is a major factor shaping plant populations on a global scale. At the same time, reproductive traits play a major role in plant fitness. With increasing elevation and increasingly harsh conditions, decreases in pollinator visitation rates, sexual investment, seed set and heterozygosity (owing to increased selfing) are expected. In response, selection and/or phenotypic plasticity could lead to an increase in floral displays by plants to increase their attractiveness to pollinators and compensate for the negative fitness impacts of reduced pollinator activity. A large body of literature tests these hypotheses at the among-species level, but empirical evidence at the population level (i.e. wihin species), where adaptive change might occur, is still limited to species-specific studies. Unravelling the global patterns of change in the reproductive traits, flower visitation rates and heterozygosity of plant populations across variable environmental conditions, especially climate, can help us to understand how species are able to cope with shifting conditions associated with global change, particularly in mountains. Here, we adopted meta-analytical approaches to assess the reproductive changes of plant populations in response to elevation on a global scale.

methodsWe used a data set with 243 paired populations of plants at 'lower' and 'higher' elevations, spanning an elevation range of 0-4380 m a.s.l. and taken from 121 angiosperm species and 115 published studies. We analysed changes in the number of flowers, their size and longevity, pollen production, flower visitation rate, seed set and expected heterozygosity. We then tested whether the observed patterns for each trait were dependent upon plant phylogeny and various ecogeographical factors and species traits. KEY

resultsWe found no evidence of elevation having a global effect on the reproductive traits of angiosperm populations. This null global pattern was not affected by geography or phylogenetics.

conclusionsOur results suggest that changes in reproductive traits, flower visitation rates and heterozygosity in plant populations across elevations are specific to each species and ecosystem. Hence, macroevolutionary (across species) and macroecological patterns of elevation of plant reproduction reported previously are apparently not simply the outcome of microevolutionary changes (within species). This apparent specificity of response across plant species poses difficulties in predicting the effects of global changes and, specifically, climatic changes, on the fate of plant species, populations and communities.