Trait divergence and trade-offs among Brassicaceae species differing in elevational distribution

Abstract

Species have restricted geographic distributions and the causes are still largely unknown. Temperature has long been associated with distribution limits, suggesting that there are ubiquitous constraints to the evolution of the climate niche. Here, we investigated the traits involved in such constraints by macroevolutionary comparisons involving 100 Brassicaceae species differing in elevational distribution. Plants were grown under three temperature treatments (regular frost, mild, regular heat) and phenotyped for phenological, morphological, and thermal resistance traits. Trait values were analyzed by assessing the effect of temperature and elevational distribution, by comparing models of evolutionary trajectories, and by correlative approaches to identify trade-offs. Analyses pointed to size, leaf morphology, and growth under heat as among the most discriminating traits between low- and high-elevation species, with high-elevation species growing faster under the occurrence of regular heat bouts, at the cost of reduced size. Mixed models and evolutionary models supported adaptive divergence for these traits, and correlation analysis indicated their involvement in moderate trade-offs. Finally, we found asymmetry in trait evolution, with evolvability across traits being 50% less constrained under regular frost. Overall, results suggest that trade-offs between traits under adaptive divergence contribute to the disparate distribution of species along the elevational gradient.

Keywords: Heat and frost stress; Ornstein-Uhlenbeck; macroevolution; phylogenetic signal; range limits; thermal niche.

Figure 1

Boxplot showing the distribution of species‐mean trait values for which species differed depending on their median elevation (low vs. high elevation), either across growth treatments or in a particular growth treatment (regular frost, mild, regular heat). For simplicity, only data of the second round of sowing are included and traits for which mixed‐effects models and evolutionary models produced concordant results (data of both rounds of sowing and all traits shown in Fig. S2). Colors inside boxes represent the treatments (blue for Frost, grayscale for Mild, and red for Heat), whereas the intensity represents median elevation of species occurrence (dark colors for low elevation, light colors for high elevation). The thick horizontal line is the median, and the lower and upper hinges are the 25th and 75th percentiles; whiskers extend from the hinges to the most extreme data points within 1.5 × IQR, and dots are values beyond those ranges.

Figure 2

Trait differentiation between low‐ and high‐elevation species, as revealed by discriminant analyses and multitrait correlations. Each point represents a species. The median elevation of origin is represented by a color scale ranging from green (low elevation) to brown (high elevation). The black line reflects the relationship between pairs of traits, and the associated correlation coefficient is reported (significance: ^* P < 0.05, ^** P < 0.01, ^*** P < 0.001; full details in Table S4). Trait values are centered and scaled to unit variance.