• B. J. Bond,

  • M. G. Ryan

Comment on ‘Hydraulic limitation of tree height: a critique’ by Becker, Meinzer & Wullschleger

In our paper in BioScience ( Ryan & Yoder 1997), we noted that there are striking similarities in growth patterns of woody plants. For example, height growth, maximum height and stem growth tend to be closely correlated for even-age stands or open-grown trees, and all are highly predictable for a particular species on a particular site. We reviewed several hypotheses to explain these patterns and concluded that one of them, which we termed the ‘hydraulic limitation hypothesis’, was especially attractive. The hydraulic limitation hypothesis predicts that the hydraulic resistance of the soil-to-leaf pathway will increase as trees grow in size and that this increased resistance will reduce stomatal opening, photosynthesis and growth as trees approach maximum height. The hypothesis provides a functional linkage among height growth, photosynthesis and productivity, and there is experimental evidence supporting underlying mechanisms. For the alternative hypotheses we considered, either there is experimental disproof (for the respiration hypothesis) or limited information available (for the maturation hypothesis).

In their critique, Becker, Meinzer & Wullschleger (2000) offer interpretations that they believe contradict the hydraulic limitation hypothesis. They discuss ways that woody plants may ‘compensate’ for the effects of size on hydraulic resistance and they offer several alternate mechanisms for height limitation: competition, genetics, reproduction and environmental damage. We welcome their criticism and the discussion it provokes, for, as Wilson (1998) stated:

‘there (are) two kinds of original thinkers, those who upon viewing disorder try to create order, and those who upon encountering order try to protest it by creating disorder. The tension between the two is what drives learning forward.’

We will limit our response to three points. First, Becker et al. (2000) misinterpret the hypothesis and most of the counter examples they offer are not valid tests of the hydraulic limitation hypothesis. We also discuss studies not completed when the original article was published. Second, Becker et al. (2000) suggest that there are compensatory mechanisms that allow trees to maintain ‘homeostatic’ balance of water supply and demand as they grow. We agree that compensation occurs [Mencuccini & Magnani (2000) have written an excellent summary of these adjustments] but offer counter-examples that suggest that these compensations do not entirely offset hydraulic limitation. Third, we argue that the alternatives offered by Becker et al. (2000) for height limitation either are not satisfactory or do not exclude the role of hydraulics in limiting tree height and tree growth.

Becker et al. (2000) use two inappropriate comparisons when attempting to examine the hydraulic limitation hypothesis. First, they compare hydraulic characteristics for trees not only of different size but also of different species and, second, they compare trees of the same species and different size, but growing in different environments. Neither of these is a valid test of the hydraulic limitation hypothesis. The hydraulic limitation hypothesis predicts that whole-tree hydraulic conductance, normalized by leaf area (which we term leaf-specific conductance, or KL) of genetically identical trees in identical environments should decrease as they approach their maximum height. The hypothesis does not predict that KL should always decrease with increasing height. Species differ in maximum height, hydraulic architecture, xylem anatomy and vulnerability to xylem cavitation, and it would be quite surprising if KL increased with height for any combination of species.

Even within a species, there are many situations in which KL may be unrelated or even positively correlated to tree height without contradicting the hypothesis of hydraulic limitation to maximum height. As we explained in Ryan & Yoder (1997), trees growing on nutrient-poor sites are expected to have low KL compared to larger trees of the same species on better sites because of changes in carbon allocation and wood structure, such as an increase in the proportion of latewood to earlywood. This may explain why maximum height varies with site fertility for genetically identical trees. Likewise, a small tree in a forest understorey with suppressed growth rates owing to shading or limited space for root exploration will not necessarily have greater KL than a non-suppressed, larger tree of the same species.

Becker et al. (2000) cited several studies that found little or no relationship between KL and tree height. None of these studies is a reasonable test of the hypothesis that height and productivity are ultimately constrained by tree hydraulics. In three of the studies ( Andrade et al. 1998 ; Goldstein et al. 1998 ; Becker, Tyree & Tsuda 1999), different species were compared, with one case ( Becker et al. 1999 ) even comparing tropical and temperate species. Dawson (1996) reported lower sapflow (per unit leaf area) for small Acer saccarum compared with large trees, but the roots of the large trees grew through cracks in a soil hard pan, allowing them access to ground water that was not available to the smaller trees. The critical question in reference to hydraulic limitation is whether KL and transpiration decline after the trees gain access to ground water and approach their maximum height for that site. In another study ( Saliendra, Sperry & Comstock 1995) cited by Becker et al. (2000) , KL decreased with increasing size of Betula occidentalis (juveniles > saplings > adults). However, the trees were growing in a riparian community and it is not clear whether small and large trees were growing in equivalent environments; small trees may have been in the understorey. So even though this study conforms to expectations of hydraulic limitation to height, it is not a good test of the hydraulic limitation hypothesis.

To avoid these problems in interpretation, we have limited most of our comparisons to individual, open-grown trees of different sizes in the same environment or to chronosequences of even-age stands in similar environments. Furthermore, we have sought experimental situations where we know that height growth and volume growth efficiency decreased with size. Since our 1997 paper, several new studies have demonstrated a decrease in KL with increasing size in such situations. KL was lower in 12–15 m Nothofagus solandrii compared to adjacent 3–4 m trees (M. G. Ryan and D. Whitehead, unpublished data). Also, for five paired old and young N. solandrii stands along an elevation gradient, δ13C was 1·5% greater in the older stands, indicating a consistent pattern of lower canopy stomatal conductance in the older trees (M. G. Ryan and D. Whitehead, unpublished data). In wet soil conditions in Wind River, WA, KL of 400+ year-old Douglas-fir trees was significantly lower than that of 20 and 40 year-old trees (Phillips et al. 1999). When gravitational resistance was subtracted, the 40 year-old trees had higher KL than either of the other two age classes. Growth and yield data for the area show that productivity peaks at about 35–45 years. In another recent study, mean stomatal conductance of Fagus sylvatica was 60% less in trees 30 m taller when soil moisture was high ( Schafer, Tenhunen & Oren 1999).

We proposed that hydraulic limitations to height and productivity are universal, so it is important to test the hypothesis in a variety of ecosystems and situations. In measurements of two tropical hardwood species (Tapirira guianensis and Simarouba amara) in Panama, we found no evidence of lower KL in larger trees compared with smaller trees (N. Phillips et al., unpublished). We also discovered that it is very difficult to design a good test of the hydraulic limitation hypothesis in mixed-age, mixed-size tropical forests. For small trees in gaps, the light and humidity environment is distinctly different from the overstorey trees, and their growth may be suppressed by competition from neighbouring trees. In this mixed-age and mixed environment stand, we were not comparing trees of different size growing under identical conditions. In fact, it is quite possible that the smaller trees were growing less per unit leaf area that the larger trees.

We realized that to test the hydraulic limitation hypothesis, experiments must be designed carefully to avoid misinterpretations owing to possible change in environmental conditions as trees grow. One approach might be to remove competing vegetation around trees of different sizes (but with the same species) and examine their growth and water relations over several years. Alternatively, a very long-term study could follow growth and water relations of individual trees as they emerge from the subcanopy to the overstorey and achieve their maximum height. We suspect that the choice of experimental locations may underlie the different interpretations between us and Becker et al. (2000) . Most of the studies cited in Becker et al. (2000) are mixed-age stands, where environment almost certainly differs with tree size, and where tall trees are in an environment with greater evaporative demand.

Becker et al. (2000) discussed three mechanisms that could compensate for hydraulic resistance owing to increased size: decreased leaf area:sapwood area ratio (AL/AS), increased wood permeability and increased water storage (capacitance) in larger trees. There is good experimental evidence for all three mechanisms. Magnani, Mencuccini & Grace (1999) proposed a fourth possibility: increased allocation to fine roots may compensate for increased hydraulic resistance above ground (thus decreasing the relative importance of below-ground resistance in larger trees). But do these mechanisms allow trees to maintain a homeostatic balance between water supply and demand as they approach maximum height? We know that they do not in the Pinus ponderosa system we have studied intensively. Three independent measurements, including leaf-level gas exchange ( Yoder et al. 1994 ; Hubbard, Bond & Ryan 1999), stable carbon isotopes in foliage and stemwood ( Yoder et al. 1994 ) and whole-tree sapflow (M. G. Ryan et al. unpubl. data) all indicate that larger trees have lower stomatal or canopy conductance than small trees in the same environment. Mencuccini & Magnani’s (2000) analysis of published data showed that changes in sapwood permeability and AL/AS were not sufficient to compensate for total above-ground resistance with increasing tree size.

If trees adjust their carbon allocation or xylem anatomy to compensate for potential increases in hydraulic resistance owing to size, then this resistance must be an important factor in both evolutionary adaptation and acclimation through the life of the individual. In our view, the fact that these compensations occur is evidence that hydraulic limitations must be important to overall fitness of woody plants. As Mencuccini & Magnani (2000) point out, these compensations are likely to incur ‘costs’. For example, increased vulnerability to xylem cavitation is associated with increases in permeability, and allocation to fine roots would reduce above-ground growth. A more sophisticated and comprehensive view of hydraulic limitations to tree height and tree growth should take into consideration these costs in addition to the direct effects of decreased KL on leaf gas exchange.

We conclude with a brief discussion of genetic vs physical limitations of height growth. The form and functions of all living things are constrained by genetic factors, operating interactively with environmental factors. This is not incompatible with the idea of hydraulic limitation. Genetic factors could determine differences among species in hydraulic architecture, vulnerability to xylem cavitation, and stomatal sensitivity to water stress. According to the hydraulic limitation hypothesis, these genetic differences should result in different maximum heights for different species growing in the same conditions. The hypothesis also accounts for why an individual might achieve a different maximum height in different environmental conditions. There is no doubt that reproductive efforts also have a large impact on productivity and probably height growth. But again, this is not incompatible with the idea that maximum height is ultimately constrained by hydraulics.

It is true, as Becker et al. (2000) point out, that grafting experiments have demonstrated a ‘memory’ effect. Tissues from older plants grafted onto younger rootstock maintain growth characteristics of the older plant. We discussed this in our BioScience article and will not repeat the arguments here except to point out that very little is known about internal, or genetic, regulation of maturation and senescence processes in perennial plants. However, if a genetic change has occurred with age, there would need to be some measurable characteristic (such as a decrease in photosynthetic capacity) to cause a reduced growth at the same leaf area. Currently, the evidence for such a characteristic is mixed ( Ryan, Binkley & Fownes 1997). For genetic differences to explain observed patterns of height and volume growth as woody plants age, it will be necessary to demonstrate how endogenous controls interact with environmental cues such that genetically identical individuals achieve different maximum heights and sizes in different environmental conditions.