Coordination of traits associated with gmand Kleaf
Our findings on the coordination of traits associated with photosynthetic C-gain and transpirational water-loss in C4 grasses contrast some of those reported previously for C3 species. For instance, C4 grasses adapted to drier habitats exhibit traits associated with greater gm and lower Kleaf (Table 1). Also, gm, Anet and traits associated with gm like SDada, SR and Smes (Pathare et al ., 2020) scaled positively with traits like IVD, leaf thickness and VED (Fig. 1) which are known to be important determinants of Kleaf (Sack et al. , 2013; Buckley et al. , 2015) These results suggest that Kleaf may be negatively related to gmand hence Anet for the C4 grasses belonging to habitats with diverse MAP. Indeed, Kleafestimated using anatomical traits scaled negatively with gm and Anet. This finding contrasts the previous reports of a positive relationship of Kleafwith gm and Anet observed in C3 species (Sack & Holbrook, 2006; Flexas et al. , 2013; Xiong et al. , 2017; Drake et al. , 2019) and could be partly explained by the carbon concentrating mechanism of C4 species that maintains high Anet at relatively low gsw compared to C3species (Ocheltree et al. , 2016) and the relationship of gm and Kleaf with leaf thickness. Specifically, greater leaf thickness in C4 grasses from drier habitats was associated with greater VED and lower total VLA (Fig. 1) which may imply a lower Kleaf and also an increase in space available for photosynthetic tissue (Brodribb et al. , 2007; McKown et al. , 2014; Zwieniecki & Boyce, 2014; Buckley et al. , 2015). Also, in these C4 grasses, greater leaf thickness was associated with a greater SDada, SR and Smes implying a greater gm and Anet (Muir, 2018). Consequently, we observed a negative relationship of Kleaf with gm and hence Anet in these C4 grasses. The negative relationship of SDada with total VLA and Kleaf (Fig. 1) also contrasts the previous reports for C3 species (Drake et al ., 2019) and suggests that, for the C4 grasses the presence of a greater number of stomata may not be associated with greater investment in leaf water transport tissue and hence Kleaf , though it is associated with a greater gm and Anet(Pathare et al ., 2020).
Species adapted to drier habitats are known to employ a safer xylem strategy, wherein, resistance to cavitation is achieved by maintaining lower Kleaf, which however comes at the cost of Anet thus leading to the safety versus efficiency trade-off (Zimmermann, 1983; Meinzer et al. , 2010). However, C4 grasses adapted to low MAP exhibited traits associated with lower Kleaf, but greater gm and Anet (Table 1, Fig. 1). Previous studies have also observed a decoupling between Kleafand Anet for the C4 grasses (Kocacinar & Sage, 2003; Ocheltree et al. , 2016). Our results along with these previous findings suggest that maintaining a greater Kleaf in order to achieve greater Anetmay not be a necessity for C4 grasses and that the safety-versus efficiency trade-off may not apply to the C4 grasses which can achieve greater gmand Anet in drier habitats whilst maintaining a lower Kleaf.
The Kleaf values estimated here (10 to 27 mmol m-2 s-1 MPa-1) using anatomical traits, are within the range measured previously (5 to 30 mmol m-2 s-1MPa-1) for diverse C4 grasses (Liu & Osborne, 2015; Liu et al. , 2019). Because Kleafwas not measured directly, but estimated using a semi-empirical model based on diverse plant groups (Brodribb et al. , 2007; de Boeret al. , 2016), there are some uncertainties associated with using this model for C4 grasses. For instance, the presence of Kranz anatomy, presence or absence of bundle-sheath suberisation and the relationship of stomatal density with VLA (Fig.1) can all influence the estimates of Kleaf. There is a need to address these uncertainties in future efforts for estimating Kleaf in C4 grasses. Though we investigated the relationship of Kleaf with gm and important mesophyll and BS traits, studies suggest that these traits may correlate strongly with the Kox component of Kleaf, then Kx (Sack & Scoffoni, 2013; Buckley et al. , 2015). Investigating the relationship of gm with Kox, once a direct method for estimating Kox is developed (Song & Barbour, 2016; Barbour, 2017; Barbour et al. , 2017), will provide greater ability to scale up from physiological processes to whole-leaf functions.