Stem and leaf hydraulics of congeneric tree species from adjacent tropical savanna and forest ecosystems

Leaf and stem functional traits related to plant water relations were studied for six congeneric species pairs, each composed of one tree species typical of savanna habitats and another typical of adjacent forest habitats, to determine whether there were intrinsic differences in plant hydraulics between these two functional types. Only individuals growing in savanna habitats were studied. Most stem traits, including wood density, the xylem water potential at 50% loss of hydraulic conductivity, sapwood area specific conductivity, and leaf area specific conductivity did not differ significantly between savanna and forest species. However, maximum leaf hydraulic conductance (K _leaf) and leaf capacitance tended to be higher in savanna species. Predawn leaf water potential and leaf mass per area were also higher in savanna species in all congeneric pairs. Hydraulic vulnerability curves of stems and leaves indicated that leaves were more vulnerable to drought-induced cavitation than terminal branches regardless of genus. The midday K _leaf values estimated from leaf vulnerability curves were very low implying that daily embolism repair may occur in leaves. An electric circuit analog model predicted that, compared to forest species, savanna species took longer for their leaf water potentials to drop from predawn values to values corresponding to 50% loss of K _leaf or to the turgor loss points, suggesting that savanna species were more buffered from changes in leaf water potential. The results of this study suggest that the relative success of savanna over forest species in savanna is related in part to their ability to cope with drought, which is determined more by leaf than by stem hydraulic traits. Variation among genera accounted for a large proportion of the total variance in most traits, which indicates that, despite different selective pressures in savanna and forest habitats, phylogeny has a stronger effect than habitat in determining most hydraulic traits.