Figure legends

Fig. 1 : A conceptional image for “direct effects” and “indirect effects” of leaf‐trichome resistance on gas exchanges. Inb and c , blue and red arrows represent the direct effects (i.e., increasing diffusion resistance for CO2 and H2O fluxes) and the indirect effects (i.e., enhancing gas diffusion due to increased leaf temperature via increasing resistance for heat fluxes) of leaf-trichome resistance, respectively. Wide grey arrows on the y-axis represent the “combined effects” of leaf-trichomes resistance on the photosynthetic and transpiration rate.
Fig. 2 : Representative diurnal variations in temperature of leaves either trichome-shaved or non- shaved (intact) along with diurnal variations in air temperature and relative humidity at four elevational sites (a , c , e , g ). Difference in the temperature between intact and trichome‐shaved leaves are shown where the positive values indicate higher temperature in the intact leaves (b , d , f , h ). These three days included typical sunny days and cloudy or rainy days (8/9/2017–8/11/2017 in the 2400‐m site, 8/10/2017–8/12/2017 in the 1800‐m site, 8/13/2017–8/15/2017 in the 1200‐m site, and 8/14/2017–8/16/2017 in the 700‐m site). The error bars represent standard error.
Fig. 3 : Relationship between trichome thickness and the temperature differences between intact and trichome‐shaved leaves at the 1800 m site (air temperature 22.0 ± 1.0 °C). Positive values represent that leaf temperature was higher in leaves with leaf trichomes than in their compared leaves without leaf trichomes. The points and error bars represent mean values and standard errors for three days.
Fig. 4 : Environmental variables, simulated leaf temperature, and gas-exchange characteristics at five elevational sites. Field measured air temperature, relative humidity, and PPFD (photosynthetic photon flux density) for the simulation analyses (a ,b ), calculated diurnal variations in leaf temperature, difference in temperature between leaf and air, the photosynthetic rate, and the transpiration rate of leaves with the mean traits values for each site (c , d , e, f) , and the effects of leaf‐trichome resistance on leaf temperature, the photosynthetic rate, the transpiration rate which were calculated as differences between leaves with and without leaf‐trichome resistance (g , h, i ). Positive values mean that leaves with trichomes have higher values than those without trichomes.
Fig. 5 : The calculated values of the H2O-flux resistances (a ), the heat-flux resistances (b ), the daily photosynthetic rate (d ), the daily transpiration rate (e ), and the daily water-use efficiency (f ), and the effects of leaf‐trichome resistance on each physiological characteristic (c, g ) in the simulation analyses. In c andg , positive values represent that values are greater in leaves with trichomes than those without trichomes (see Eq. 1).
Fig. 6 : Changes in effects of leaf‐trichome resistance on the resistance on the H2O-flux resistances (a ), the heat-flux resistances (b ), mean and maximum leaf temperature at daytime (c , d ), the daily photosynthetic rate (e ), the daily transpiration rate (f ), and the daily WUE (g ) at each elevational site with varying trichome thickness in the simulation analyses. Positive values represent that values are greater in leaves with trichomes than those without trichomes (see Eq. 1). Circle cross points represent the values when the average trichome thickness for each elevational site was used. The error bar represents the standard deviation of trichome thickness for each elevational site (Table 3).
Fig. 7 : Simulated photosynthetic rates (A ) as a function of trichome thickness and characteristic leaf dimension (an indicator of leaf width) for conditions of stomatal coefficient (a ) = 2.5 and relative humidity (h ) = 0.8 in the sensitivity analyses.a : a heatmap of A calculated for air temperature (Ta ) = 20°C and PPFD (Q ) = 2400 µmol m-2 s-1, and plotted against trichome thickness (x -axis; 0.00-1.00 mm) and characteristic leaf dimension (y -axis; 1.0-5.0 cm). b : a heatmap represents to what extent the leaf-trichome resistance enhances (orange) or suppresses (navy) A in the same condition as a .c : heatmaps of A calculated for variousTa and PPFD and each heatmap in c is drawn for the same axis as in a . d : heatmaps of the effects of leaf-trichome resistance on A calculated for variousTa and Q and each heatmap in d is drawn for the same axis as in b . In c and d , boxed sections represent a and b , respectively, and two-way arrows represent temperature ranges at each site.
Fig. 8: Simulated photosynthetic rates (A ), transpiration rates (E ), and water-use efficiency (WUE) with relative humidity (h ) = 0.8 as a function of leaf traits (trichome thickness, characteristic leaf dimension, stomatal coefficient) and environmental conditions (air temperature, light intensity) in the sensitivity analyses. Similar to Figure 7, each heatmap on the left side represents A , E , and WUE in relation to various trichome thickness (x-axis) and characteristic leaf dimension (y-axis), and each heatmap on the right side represents to what extent the leaf-trichome resistance enhances (orange) or suppresses (navy) A , E , and WUE in the same set of x- and y-axis. These simulations are done for different stomatal coefficient (a ). The simulation for different relative humidity (h ) was shown in Figure S9 (h = 0.2) and S10 (a = 0.5).