RESULTS
There was a significant reduction on liana performance toward higher latitudes (F (2,13) 2.37 p =0.015, fig. 1), reaching differences up to 5 times on apical growth rate between liana species in the northern and southern sites (Nahuelbuta 10.6 cm*month-1 v/s Aysén 2.1 cm*month-1, Tukey HSD p= 0.017). At the same time, the loss of liana performance was correlated with the increase in the percentage of loss conductivity towards higher latitudes (Fig. 1) evidencing that cold temperatures have negative effect over liana performance at higher latitudes.
We observed marked differences in hydraulic traits strategies associated to efficiency in water transport capacity along the latitudinal gradient. The maximum specific hydraulic conductivity (Ks_max) significantly differ between sites (p =0.006), with a decreased 26.1 times in lianas species at the northern site “Nahuelbuta” compared to the southern sites of the gradient “Aysen” (Table 2, Tukey HSD p =0.001) and the middle site “Puyehue” (Tukey HSD, p =0.027). Nahuelbuta lianas species have a Ks_max of 28.7 Kg s-1m-1 MPa-1 while in Aysén they barely reached 1.1 Kg s-1 m-1MPa-1 (Table 2). It must be added that the percentage of loss conductivity (PLC) along the sites differ significantly (p =0.01), increasing from 24.5% at the northern site to 37.5% in the southern site (Table 2, Tukey HSD, Nahuelbuta differ from Puyehue p=0.05 and Aysen p =0.04) intensified the loss of hydraulic efficiency in liana species in the colder extreme of the gradient (Table 2), thus, the joint reduction in hydraulic transport capacity between the liana species in Nahuelbuta and Aysén was 42 times.
We also detected differences in traits associated to hydraulic transport safety. The lower water transport capacity of lianas at higher latitudes was also correlated with the decrease in xylem vessel diameter, which decreased 30% toward the southern site (Table 2). This also vessel density differs between sites (p =0.04) increasing by 80%, and wood density also differ between sites (p =0.04) and increase 19% towards the south end of the gradient (Table 2). As mentioned before water transport capacity is strongly modulated by freezing-thaw embolism, which would have a strong impact on the relative growth rate.
Contrary to what we expected, root pressure did not increase with latitude (Table 2, Tukey HSD only “Nahuelbuta” differ from “Puyehue” p=0.02). Liana species in Puyehue showed the highest root pressure of the gradient (32.7 kPa), followed by liana species in Aysén (13.7 kPa) and Nahuelbuta (9.9 kPa). Based on this, lianas would be able to revert embolism up to a maximum height of 1 meter on species in Nahuelbuta, 3.2 meter on species in Puyehue and 1.4 meter on species in Aysén, so root pressure would have a significant effect only on the basal part of the individual. However, we could have underestimated the root pressure for the species in Aysén, since the measurements had to be repeated due to manometers freezing on a couple of occasions.
We observed a decrease in the proportion of wide vessels toward the cold extreme of the gradient (Fig. 2, exemplifying two species along the latitudinal gradient). When considering that vessels with diameter higher than 100 µm are responsible for most of Ks, we estimated that 14.7% of the vessels are responsible for 69% of theoretical hydraulic conductivity in Nahuelbuta. Meanwhile 7.6% of vessels are responsible for 47% of theoretical hydraulic conductivity in Puyehue. No species showed vessels with diameter higher to 100 µm in Aysén, and all hydraulic conductivity was generated by narrow vessels. The frequency distribution is significant different for diameter of xylem vessel and relative contribution to total hydraulic conductivity, as shown by the Kolmogorov-Smirnov test (Xylem vessel for the specieMitraria coccinea sp., in Nahuelbuta p =0.002, in Puyehuep =0.01, and Aysén p =0.001; Relative contribution to total hydraulic conductivity for the specie Mitraria coccinea sp for Nahuelbuta p =0.01, Puyehue p= 0.013, Aysén p =0.001; Xylem vessel for the specie Hydrangea serratifolia sp in Nahuelbuta p =0.003, Puyehue p =0.027, and Aysénp =0.001; Relative contribution to total hydraulic conductivity for the specie Hydrangea serratifolia sp in Nahuelbutap =0.04, Puyehue p =0.022, and Aysén p =0.001. All these results show that the relative contribution of large vessel diameter to water transport was reduced along the latitudinal gradient, evidencing a tradeoff between efficiency and vascular security in the cold extreme.
The Principal Component Analysis based on all functional traits studied shows that the first component explained 45.9% of the variation and the second component explain 18.6% of the variation. This PCA analysis show that high performance is related to high efficiency in water transport (high specific hydraulic conductivity and wide vessels) (Fig.3). On the other side, low performance was related to low efficiency in water transport (high wood density, narrow vessels, and high percentage of loss conductivity). This relationship between performance, efficiency or safety in water transport is robust enough to functionally differentiate liana communities between sites along the latitudinal gradient (Fig. 3). The linear regression of the first and second component and the above growth rate (AGR) for each site was significant in Nahuelbuta (Factor 1 R2= 0.71, p =0.017; Factor 2 R2=0.57 p =0.06; fig. 4), where traits that are positively associated with growth rate were the maximum specific hydraulic conductivity (Ks_max) and xylem vessel diameter and density, on the contrary the percentage of loss conductivity (PLC) and root pressure was negative related to growth rate. In Puyehue data set, none of the components were significantly related to above growth rate (AGR; Factor 1 R2= 0.007, p=0.66; Factor 2 R2=0.039, p= 0.75, Factor 3 R2=0.14 , p=0.1; fig. 4). For Aysen, a low and nearly significant linear regression (Factor 1 R2= 0.21,p = 0.06; Factor 2 R2= 0.34, p = 0.07; fig. 4) was observed between the first and second component, the traits that positively relate to growth rate was xylem vessel diameter, root pressure and wood density, and negatively the percentage of loss conductivity.