Nội dung

Original article Root and shoot hydraulic conductance of seven Quercus species Andrea Nardini b Melvin T. Tyree a Dipartimento di Biologia, Università di Trieste, Via L. Giorgieri 10, 34127 Trieste, Italy USDA Forest Service, Northeastern Forest Experiment Station, 705 Spear Street, Burlington, VT 05402-0968, USA (Received 13 November 1998; accepted 22 February 1999) Abstract - The root (K ) and shoot (K R ) hydraulic conductances of seven different Quercus species, as well as the leaf blade S hydraulic resistance (R ), were measured in potted plants with the aim of understanding whether a relationship exists between the LL hydraulic architecture and the general ecological behaviour of different species of this genus. The K Rvalues were scaled by dividing by root surface area (K ) and by leaf surface area (K RR ) and the K RL S values were scaled by dividing by leaf surface area (K ). The SL likely drought-adapted species (Quercus suber, Q. pubescens, Q. petraea) showed lower K RL and , RR lower K K SL and higher R LL with respect to the known water-demanding species (Q. alba, Q. cerris, Q. robur, Q. rubra). The possible physiological and ecological significance of such differences are discussed. (© Inra/Elsevier, Paris.) root hydraulic conductance / shoot hydraulic conductance / leaf blade resistance / Quercus / high pressure flow meter Résumé - Les conductivités hydrauliques de la racine et de la tige de sept espèces de Quercus. Les conductivités hydrauliques de la racine (K ) et de la tige (K R ) des sept espèces de Quercus ont été mesurées avec LL ) et la résistance hydraulique des feuilles (R S pour objectif la compréhension de la relation qui existe entre l’écologie de l’espèce et son architecture hydraulique. Les valeurs des Ront été divisées par les surfaces des feuilles (K K ) et des racines (K RL ), celles des K RR S par les surfaces des feuilles (K ). Les , SL RR K RL et K K SL des espèces adaptées aux environnements arides (Q. suber, Q. pubescens, Q. petraea) sont inférieures et leurs R LL supérieures par rapport aux valeurs de celles adaptées aux environnements humides (Q. alba, Q. cerris, Q. robur, Q. rubra). Cet article se propose d’illustere ces différentces au plan physiologique et écologique. conductivité hydraulique de la racine / conductivité hydraulique de la tige / Quercus / HPFM conductance in xylem is expected low leaf water potential, because leaf water potential at a given transpiration rate is determined by soil water potential as well as by root and shoot hydraulic conductance [16]. This means that the higher the root and/or shoot hydraulic conductance, the less negative would be the leaf water potential and the less severe would be the water stress suffered by the plant in terms of reduced cell expansion, protein synthesis, stomatal conductance and photosynthesis [15]. A low 1. Introduction to cause a Many recent studies have reported the water relations of Quercus species [1, 3, 6, 18] with the aim of better understanding their different levels of adaptation to drought. A good correlation was found between vulnerability to cavitation in stems and drought tolerance [4, 8, 22]. Other studies show that hydraulic architectures of trees might be related to drought adaptation [2, 3, 23, 28]. * Correspondence and salleo@univ.trieste.it reprints hydraulic On the other hand, a high shoot hydraulic conduc- tance (due to wide conduits) might increase vulnerability to cavitation, as suggested by some authors [10, 11] others [21, 24]. As a consequence, it is still unclear whether a high hydraulic conductance of shoot and root can be of advantage to plants under water stress conditions. although questioned by To the best of our knowledge, only a few studies have appeared in the literature reporting measurements of the hydraulic conductance of whole root systems of Quercus species [12, 13]. Even less data have been reported from parallel measurements of root and shoot hydraulic conductances of different Quercus species. relation (if any) between the conductances and the general hydraulic ecological behaviour of different species of the genus Quercus, root and shoot hydraulic conductances were measured for seven oak species. In an attempt to find a root and shoot 2. Materials and methods Quercus species used in this study were Q. suber L., Q. pubescens Willd, Q. petraea (Matt) Liebl, Q. alba L., Q. cerris L., Q. robur L. and Q. rubra L. These Quercus species were selected because they are representative of different levels of adaptation to drought, ranging from species well adapted to drought such as Q. suber to water-demanding species such as Q. rubra. In particular, Q. suber is a Mediterranean evergreen sclerophyll growing from the sea level up to 700 m in altitude [17]. Q. pubescens is a semi-deciduous species growing The in calcareous soils between sea level and 1 200 m in altitude within the sub-Mediterranean climatic area (southeastern Europe [17]). Q. petraea is a European species growing in sub-acid soils between sea level and 1 000 m in altitude in Atlantic climate zones [17]. Q. cerris is a euro-Mediterranean species growing in acid soils with good water availability [17]. Finally, Q. robur is a European species growing on nutrient-rich soils, with high water availability [17]. During a visit to the United States Department of Agriculture (USDA) Northeastern Forest Experiment Station (Burlington, VT, USA), preliminary measurements of root and shoot hydraulic conductance were performed in Q. rubra and Q. alba. Although both Quercus species have an American distribution area, they were added to the present study because they represent two cases of adaptation to different water availability. Experiments were replicated on five to ten 3-year-old seedlings of each species. The seedlings were grown in pots. Dimensions of the seedlings are reported in table I in terms of height (h), trunk diameter (Ø ), total leaf surT face area ) L (A and root surface area ). R (A Pots were cylindrical in shape with a diameter of 150 mm and a height of 250 mm. Seedlings of Q. rubra and Q. alba had been grown in pots since seed germination in the greenhouse of the USDA Forest Service, (Northeastern Forest Experiment Station, Burlington, VT, USA). Experiments on these two species were performed at the Northeastern Forest Experiment Station in July 1996. Seedlings of the other species, i.e. Q. suber, Q. pubesQ. petraea, Q. cerris and Q. robur were grown in the Botanical Garden of the University of Trieste (northeastern Italy). Experiments on these species were carried out in June 1997. All the seedlings were well irrigated with about 200 g of water supplied every 2 d. cens, Root (K ) and shoot (K R ) hydraulic conductances of S five seedlings per species were measured using a high pressure flow meter (HPFM) recently described by Tyree et al. [25, 26]. The HPFM is an apparatus designed to perfuse water into the base of a root system or a shoot while rapidly changing the applied pressure (P) and simultaneously measuring the corresponding flow (F) (transient mode [26]). The HPFM can also be used to perform steady-state measurements of shoot hydraulic conductance. In this case, the pressure applied to the stem is maintained constant at P 0.3 MPa until a stable flow is recorded. In practice, it is never possible to keep = flow and pressure perfectly constant, so it is best to refer to such measurements as quasi-steady state. The HPFM technique was used in the transient mode for measuring root and shoot conductances, and in the quasi-steady-state mode for measuring leaf blade resistance (see later). The quasi-steady-state mode was not used on the roots because the continuous perfusion could cause accumulation of solutes in the stele by reverse osmosis, causing a continual decrease in driving force on water movement [25]. The pots were enclosed in plastic bags and immersed in water. The shoots were excised under water at about 70 mm above the soil, thus preventing xylem embolism. The HPFM was connected first to the base of the excised root system. The pressure was increased continually from 0.03 to 0.50 MPa within 90 s. The HPFM was equipped to record F and the corresponding P every 3 s. From the slope of the linear region of the relation of F to P it was possible to calculate root hydraulic conductance ). R (K During K Rmeasurements, the shoots remained with the cut surface immersed in distilled water while enclosed in plastic bags to prevent evaporation. The base of the stem was connected to the HPFM and the stem was perfused with distilled water filtered to 0.1 μm at a pressure of 0.3 MPa for 1-2 h. After, leaf air spaces were infiltrated with water so that water dripped from the stomata of most leaves. The pressure was then released to 0.03 MPa and maintained constant for 10 min. Three to five transient measurements per seedlings were performed. From the slope of the linear relation of F to P, the stem hydraulic conductance (K ) was calculated by S linear regression of data. The pressure was then increased again to 0.3 MPa, and the hydraulic conductance of the shoot was measured in the quasi-steady-state mode. of leaf blade (i.e. the inverse also measured in the quasi-steadystate mode by measuring shoot hydraulic resistance after removal of leaf blades. Leaf blade resistance (R ) was L calculated from: The of hydraulic resistance conductance) was elastic expansion of some components of the instrument such as tubing and connections [26]. Therefore, additional measurements of the relation of F to P were performed with the connection to solid metal rods. A linear relation of F to P with a minimal slope due to the intrinsic elasticity of the instrument was obtained. This slope was subtracted from the slope of the straight line relating F to P measured on the root or the shoot connected to the HPFM. After each experiment, the A L of the seedlings was measured using a leaf area meter (Li-Cor model 3000-A equipped with Li-Cor Belt Conveyor 3050-A). The total Rof the seedlings was also estimated as follows: the A soil was carefully removed from the root system under a gentle jet of water. The fine roots (< 2 mm in diameter) were then excised into segments 50 mm in length. The Rof ten subsamples per species was calculated by placA ing the root segments (which were brown) into a glass box and covering them with a white plastic sheet to keep them in a fixed position while improving the contrast of the root images. The box was placed on a scanner (Epson model GT-9000 Epson Europe, The Netherland) connected to a computer. A program (developed by Dr P. Ganis, Department of Biology, University of Trieste, Italy) read the bit-map images and calculated the A . R The root images were processed by the software and the Rwas obtained by multiplying the calculated area by π A assuming the root segments as cylindrical in shape. Root subsamples were then put in an oven for 3 days at 70 °C to obtain their dry weights. A conversion factor between root dry weight and surface area was obtained. The whole root system was then oven-dried and the total A R of each seedling was calculated. The A Rfor Q. alba and Q. rubra seedlings was not measured. R K and K S were both scaled by A L so that root (K ) RL conductances and shoot (K per leaf unit sur) hydraulic SL face area were obtained. K Rwas also divided by A , thus R obtaining the root hydraulic conductance per root unit surface area (K ). Finally, R RR L was multiplied by A , L thus obtaining the leaf blade hydraulic resistance normalised by leaf surface area (R ). LL 3. Results where R S is the resistance of the leafy shoot and R S-L is the resistance of the shoot after removal of the leaves. made in Burlington During preliminary (VT, USA), the agreement of transient versus quasisteady-state measurements of shoot hydraulic conductance was tested on Q. rubra shoots of different basal diameter, using the same procedure described earlier. measurements A spurious component of the hydraulic conductance when using the HPFM could be due to the measurements The relation of F to P as measured in the transient mode in roots and shoots was non-linear up to an applied pressure of 0.15 MPa, then became distinctly linear. The initial non-linearity was probably due to intrinsic elasticity of plant organs. The root and shoot hydraulic conductances measured in the different Quercus species are reported in figure 1. Root hydraulic conductance per leaf unit surface area increase from the drought-adapted to the water-demandA Student’s t-test (P ≤ 0.05) indicated that the group of drought-adapted species (Q. suber, Q. pubescens, Q. petraea) showed significantly lower values than the water-demanding species (Q. cerris, Q. robur, Q. rubra). Generally, root and shoot hydraulic conductance were approximately equal in all species except in Q. petraea and Q. robur, whose s RLwere 57 K and 59 % of the corresponding s. SL K ing species. (K figure 1, dashed columns) ranged between 4.23 x , RL -5 -5 kg·s 10 ·MPafor Q. petraea up to 11.29 x 10 -2 ·m -1 rubra. The for Q. drought-adapted ·MPa -2 ·m -1 kg·s had lower x and 4.23 -5 10 5.41 . -2 ·m -1 kg·s RL (4.98, K ,respectively) than the mesophilous species (Q. -1 MPa alba, Q. cerris, Q. robur and Q. rubra; K RL = 7.51, 8.83, -5 kg·s 6.34 and 11.29 x 10 , respectively). ·MPa -2 ·m -1 Student’s t-test (P ≤ 0.05) revealed that Q. suber, Q. pubescens and Q. petraea were not significantly different from each other, but they were all significantly different from Q. alba, Q. cerris, Q. robur and Q. rubra. Q. rubra was significantly different from all the other species (Q. suber, Q. pubescens, Q. petraea) values of species. hydraulic conductance per root unit surface area 1, white columns) was approximately the figure , RR (K same as root hydraulic conductance per leaf unit surface area ) in Q. suber, Q. pubescens and Q. cerris RL (K because root surface area approximately equalled leaf surface area. K RR of Q. petraea and Q. robur were 46 and 50 % of , R of both RL respectively, because the A K the The twice was approximately species . L A R of Q. A Root alba and Q. rubra were not measured, so it was not possible to calculate the K RR of these two species. hydraulic conductance per leaf unit surface area , figure 1, black columns) ranged between 5.32 x SL (K Shoot -5 kg·s for Q. suber and 12.2 x 10 ·MPa -2 ·m -1 ·MPa for Q. rubra. The K -2 ·m -1 kg·s SL was found to -5 10 hydraulic conductance as measured in the quasi-steady-state mode was lower than the values Shoot recorded in the transient mode. The mean values of transient to quasi-steady-state ratio were 2.53 for Q. suber, 1.11for Q. pubescens, 1.18 for Q. petraea, 1.60 for Q. alba, 1.83 for Q. cerris, 2.51 for Q. robur and 1.91 for Q. rubra. In Q. rubra, a good correlation was found between shoot basal diameter and transient to steadystate ratio; the transient to quasi-steady-state shoot hydraulic conductance ratio increased with basal diame- 2 0.787, figure 2). (r The R LL (figure 3) was found to range between 0.89 x 4 MPa· 4 MPa -1 10 ·kg in Q. rubra and 3.68 x 10 2 s·m tended to be in the higher ·kg in Q. robur. R 2 s·m -1 LL ter = drought-adapted species than in the water-demanding species, although the Student’s t-test revealed that the differences were only slightly significant (P between 0.05 and 0.1). The only exception was Q. robur, which was significantly different from all the other species. An interesting relationship was found between the general ecology of some of the species studied and the ratio of root dry weight to root surface area (RDW/A , R The two species better adapted to drought (Q. suber and Q. pubescens) showed significantly higher values of this ratio (2.51 and 2.63 x 10 -2 kg·m , respec-2 tively) than Q. petraea, Q. cerris and Q. robur, in which , respectively. -2 R was 1.71, 1.44 and 1.31 kg·m RDW/A Q. suber and Q. pubescens were not significantly different from each other, but they were significantly different from all the other species; Q. petraea was significantly different from all the other species; Q. cerris and Q. robur were not significantly different from each other (Student’s t-test, P ≤ 0.05). figure 4). 4. Discussion The RL and K K SL were of similar order of magnitude reported for other tree species [23, 26, 27]. We found a general trend of RL and K K SL showing higher values in oak species typically growing in humid areas with respect to those adapted to aridity (figure 1). Species success in mesic sites may depend on rapid growth. Rapidly growing plants are better competitors for light and soil resources. Rapid growth is promoted when growing meristems are less water stressed. A high K SL value will ensure rapid equilibration of shoots with SOIL water potential at night which will promote rapid Ψ growth. A high K SL value will also promote maximal values of Ψ MERISTEM water potential during the day. In arid environments where growth is usually slow because as of limited availability, the ability to tolerate drought important than the ability to transport water rapidly. Hence, arid zone plants need to invest less is water more carbon into shoot conductance and thus have lower K SL values. Our data suggest that high root and shoot conductances are not physiological features conferring drought resistance to plants, at least in the genus Quercus. On the contrary, it seems that high K RL and SL are important features allowing some species to K compete more successfully in regions of high water availability, thus forcing low K RL and/or K SL species to migrate to habitats were water is less abundant and growth rate is less critical to survival. study, two alternative methods of scalhydraulic conductance were compared. R K was In the present ing root normalised per leaf unit surface area as well as per root unit surface area. While in Q. suber, Q. pubescens and Q. cerris K RL equalled , RR in Q. petraea and Q. robur, K they did not. Scaling K R by A R is a more correct procedure when root physiology is under investigation. Scaling K R by A L seems to be more appropriate in an ecological context. In fact, K RL is the expression of the of the root ’sufficiency’ system to provide water to leaves [27]. Normalisation by A L is sometimes more accurate than Because of the difficulty in digging out whole root systems from the soil, the error that can be made when scaling K R by A R is intrinsically important and by A . R would underestimate A . Moreover, the use of roots less R than 2 mm in diameter for calculating A Ris rather arbibecause it is still unclear what fraction of the root trary surface area is involved in water absorption. Therefore, we feel that scaling up Rby A K L is much less subject to error when studying the hydraulic behaviour of whole root systems growing in the soil. The observed difference between transient and quasimeasurements of shoot hydraulic conducbe explained in terms of intrinsic elasticity of the stem as due to air bubbles in the xylem vessels. During transient measurements, air bubbles initially present in the xylem are continuously compressed as the pressure applied increases. This causes an additional flow that is recorded by the instrument, thus overestimating . SL During steady-state measurements the bubbles K steady-state tance might are completely compressed (and eventually dissolved) and the flow due to bubble compression does not affect the measurement. This seems to be confirmed by experiments performed on Q. rubra, showing that the discrepancies between transient and quasi-steady-state measurements are much more evident in larger and older stems. Older stems have more embolised vessels than younger stems. Our data would suggest that quasi-steady-state measurements of hydraulic conductance are more correct than transient measurements, at least in larger stems. However, it has been convincingly demonstrated that quasi-steady-state measurements of K RL are affected by a number of problems (e.g. solute accumulation in the stele [25]); therefore, in roots it is preferable to measure RL in the transient mode. Roots contain less embolised K tissue than shoots, thus transient measures of K Rare probably more accurate. Tyree et al. [26] discussed the effect of elasticity and air bubbles on conductance measurements in shoots. The effect of air bubbles can be distinguished from the effect of elasticity, when the air bubbles are separated from the HPFM by a low hydraulic resistance, i.e. when the bubbles are present at the base of a shoot or in the connector between the HPFM and the shoot. Elastic effects cause an offset in the y-intercept of the plot of flow versus pressure, but elasticity has only a minor effect on slope (= hydraulic conductance). Air bubbles in the HPFM connector affect the slope at low pressure (0-0.2 MPa), but has a rapidly diminished contribution to the slope at higher pressure. The air-bubble effect reported here is a newly recognised phenomenon. When the hydraulic resistance for water flow from the base of the shoot to the air bubbles is sufficiently high, the effect of the air bubbles increases the slope (= conductance) over the whole range of applied pressure. LL R ’s 3) were measured in the seven Quercus species (figure similar to those reported by Tyree et al. [23] for Q. robur, Q. petraea, Q. pubescens and Q. rubra. R LL includes vascular as well as non-vascular water pathways from the leaf base to mesophyll air spaces, but it is generally thought that the main hydraulic resistance is located in the non-vascular component of the path [20]. The higher the resistance to water flow, the larger should be the water potential drop in the guard cells of stomata during transpiration. This might cause stomatal closure under water stress conditions. A rapid and substantial drop in leaf water potential is advantageous in that it allows stomata to close before xylem water potential reaches the cavitation threshold [9]. Thus, differences in LL could account for the different capabilities of stomR atal control of embolism observed in Quercus species [5]. The higher s LL have been reported in the more R drought-adapted species, with the exception of Q. robur. Field studies by Nardini et al. [14] show that Q. suber (with a high ) LL had good stomatal control of water loss R under drought stress conditions while Q. cerris (with a low ) LL was unable to prevent water loss by stomatal R closure. The ratio of RDW/A R (figure 4) was higher in the drought-adapted species than in the water-demanding species. It is very likely that high values of this ratio are mainly due to roots with many small and very densely packed cells in the cortex. When the RDW/A R ratio was or no significant correlation was plotted versus K RL , RR K found between the two parameters for the different species. It is generally thought that the main resistance to water flow in plant roots is located in the non-vascular pathway [7]. According to the ’root composite model’ proposed by Steudle and Heydt [19], water migrates in the root across the apoplastic pathway at high transpiration rates. In this case, the resistance to water flow is mainly dependent on the overall length of the path, which does not change much when many densely packed cells are compared to somehow looser cortex cells. This could explain why a significant correlation could not be found between root conductance and root mass per unit surface area. An alternative explanation for the higher R ratio measured in drought-adapted species RDW/A could be that these species might accumulate more starch in their roots. In conclusion, our results indicate that significant differences in the stem hydraulic architecture of Quercus species can account for their different ecological requirements, although further studies are needed to compare the physiological indices with species ecology. In particular, the case of Q. robur deserves further investigation, because this species showed somewhat peculiar features when compared with other water-demanding Quercus trees. References [1]Abrams M.D., Adaptations and responses to drought in Quercus species of North America, Tree Physiol. 7 (1990) 227-238. [2] Acherar M., Rambal S., Comparative water relations of four Mediterranean oak species, Vegetatio 99-100 (1992) 177-184. [3] Cochard H., Tyree M.T., Xylem dysfunction in Quercus: vessel sizes, tyloses, cavitation and seasonal changes in embolism, Tree Physiol. 6 (1990) 393-407. [4] Cochard H., Bréda N., Granier A., Aussenac G., Vulnerability to air embolism of three European oak species (Quercus petraea (Matt) Liebl, Q. pubescens Willd, Q. L.), Ann. Sci. For. 49 (1992) 225-233. robur [5] Cochard H., Bréda N., Granier A., Whole tree hydraulic conductance and water loss regulation in Quercus during drought: evidence for stomatal control of embolism?, Ann. Sci. For. 53 (1996) 197-206. [6] Epron D., Dreyer E., Long-term effects of drought on photosynthesis of adult oak trees [Quercus petraea (Matt) Liebl. and Quercus robur L.] in a natural stand, New Phytol. 125 (1993) 381-389. [7] Frensch J., Steudle E., Axial and radial hydraulic resisof maize (Zea mays L.), Plant Physiol. 91 (1989) tance to roots 719-726. [8] Higgs K.H., Wood V., Drought susceptibility and xylem in seedlings of 4 For. 52 (1995) 507-513. dysfunction European oak species, Ann. Sci. [14] Nardini A., Lo Gullo M.A., Salleo S., Competitive strategies for water availability in two Mediterranean Quercus species, Plant Cell Environ. 22(1999) 109-116. [15] Nilsen E.T., Orcutt D.M., Physiology of Plants under Stress, John Wiley & Sons, Inc., New York, 1996. [16] Pallardy S.G., Hydraulic architecture and conductivity: an overview, in: Kreeb K.H., Richter H., Hynckley T.M. (Eds.), Structural and Functional Responses to Environmental Stresses, SPB Academic Publishing, The Hague, the Netherlands, 1989, pp. 3-19. [17] Pignatti S., Flora d’Italia, Edagricole, Bologna, 1982. [18] Salleo S., Lo Gullo M.A., Sclerophylly and plant water relations in three Mediterranean Quercus species, Ann. Bot. 65 (1990) 259-270. [19] Steudle E., Heydt H., Water transport across tree roots, in: Rennenberg H., Eschrich W., Ziegler H. (Eds.), Trees Contributions to Modern Tree Physiology, Backhuys Publishers, Leiden, the Netherlands, 1997, pp. 239-255. [20] Tyree M.T., Cheung Y.N.S., Resistance to water flow in Fagus grandifolia leaves, Can. J. Bot. 55 (1977) 2591-2599. [21] Tyree M.T., Dixon M.A., Water stress induced cavitation and embolism in some woody plants, Physiol. Plant. 66 (1986) 397-405. [22] Tyree M.T., Cochard H., Summer and winter embolism in oak: impact on water relations, Ann. Sci. For. 53 (1996) 173-180. [23] Tyree M.T., Sinclair B., Lu P., Granier A., Whole hydraulic resistance in Quercus species measured with a high-pressure flowmeter, Ann. Sci. For. 50 (1993) 417-423. [24] Tyree M.T., Davis S.D., Cochard H., Biophysical pershoot xylem spectives of xylem evolution: is there a tradeoff of hydraulic efficiency for vulnerability to dysfunction?, IAWA Bull. 15 [10] Lo Gullo M.A., Salleo S., Wood anatomy of some trees (1994) 335-360. [25] Tyree M.T., Yang S., Cruiziat P., Sinclair B., Novel Jones H.G., Sutherland R.A., Stomatal control of embolism, Plant Cell Environ.14 (1991) 607-612. [9] with diffuse- and ring-porous wood: some functional and ecological interpretations, Giorn. Bot. Ital. 124 (1990) 601-613. [11] Lo Gullo M.A., Salleo S., Different vulnerabilities of Quercus ilex L. to freeze- and summer drought-induced xylem embolism: an ecological interpretation, Plant Cell Environ. 16 (1993) 511-519. [12] Nardini A., Ghirardelli L., Salleo S., Vulnerability to freeze-stress of seedlings of Quercus ilex L.: an ecological interpretation, Ann. Sci. For. 55 (1998) 553-565. [13] Nardini A., Lo Gullo M.A., Salleo S., Seasonal changes trees: an 81-90. of root hydraulic conductance (K ) in four forest RL ecological interpretation, Plant Ecol. 139 (1998) methods of measuring hydraulic conductivity of tree root systems and interpretation using AMAIZED, Plant Physiol. 104 (1994) 189-199. [26] Tyree M.T., Patiño S., Bennink J., Alexander J., Dynamic measurements of root hydraulic conductance using a high-pressure flowmeter in the laboratory and field, J. Exp. Bot. 46 (1995) 83-94. [27] Tyree M.T., Velez V., Dalling J.W., Growth dynamics of root and shoot hydraulic conductance in seedlings of five neotropical tree species: scaling to show possible adaptation to differing light regimes, Oecologia114(1998) 293-298. [28] Zimmermann M.H., Xylem Structure and the Ascent of Sap, Springer-Verlag, Berlin, 1983.