Functional traits
Leaf traits were measured on at least five leaves per species. Fresh mass was measured on fully hydrated leaves. Leaf length and width were measured with digital calipers or rulers, and leaf lamina thickness was recorded on at least three positions from the apical, median and basal section for each leaf using a digital micrometer. We measured leaf area with a LI-3100C area meter (LI-COR, Lincoln, NE). Dry mass was determined after drying the leaves at 70°C. From these measurements we calculated leaf mass per area (LMA, kg m–2), leaf dry matter content (LDMC, g g–1), and leaf density (leaf dry mass per unit leaf volume). To address hypothesis 3 we estimated leaf thermal time constant τ (s) from first principles following Michaletz et al . (2015), using species-level mean values of LMA, LDMC, leaf area, and leaf width; mean annual temperature derived from WorldClim2; and wind speed set to 2 m s–1, as:
\(\tau=\varphi LMA\left[\frac{c_{p,w}}{LDMC\ \bullet\ h}+\frac{c_{p,d}-c_{p,w}}{h}\right]\)Eqn 2
where φ (dimensionless) is the ratio of projected-to-total leaf area, cp,d (J kg–1K–1) is the specific heat capacity of dry leaf matter, and cp,w (J kg–1K–1) is the specific heat capacity of water. The heat transfer coefficient h (W m–2K–1) was calculated as \(h=\rho_{a}c_{p,a}g_{h}\), where ρa (kg m–3) is air density, cp,a (J kg–1K–1) is the specific heat capacity of air at a constant pressure, and gh is the heat conductance. The effects of radiation and transpiration were not included as these data were not available for all species. This permits us, however, to quantify how variation in the measured morphological traits alone causes variation in thermal time constants. Functional trait data were obtained for 132 species as insufficient healthy leaves were available for all species. For species with compound leaves we used traits (LMA, size, τ, etc) collected on leaflets for further analyses. Where necessary, traits were transformed to improve normality of the data.