• acclimation;
  • adaptation;
  • drought;
  • hydraulic architecture;
  • phenotypic plasticity;
  • xylem embolism

The interest in the hydraulic architecture of trees, and the interplay between xylem anatomical characteristics and dimensions, its hydraulic conductance and water transport through the plant, stretches back to Leonardo da Vinci’s observations and Huber’s (1928) first systematic measurements of a constant area of conductive sapwood across the entire length of the tree and of a close correspondence with the amount of leaf area supported. Apart from the pure interest in the inner workings of plant self-organization, the study of hydraulic architecture has since demonstrated its role in key tree and ecosystem processes as a result of (1) the close association between gaseous-phase and liquid-phase conductances (i.e. between stomatal and hydraulic conductances, responsible for leaf water loss and replenishment, respectively), and important effects on CO2 availability and photosynthetic rates (Brodribb et al., 2005), (2) the requirements imposed by hydraulic constraints on growth allocation between tree parts, which affects both primary production (through allocation to transpiring foliage) and net ecosystem production (through allocation to short-lived, easily decomposed fine roots; Litton et al., 2007) and (3) xylem and foliage vulnerability to extreme events, whenever the limits imposed by plant hydraulic architecture and stomatal behaviour are exceeded, resulting in extreme tissue dehydration and foliage dieback (Martínez-Vilalta & Piñol, 2002). Several studies have demonstrated the variability of plant hydraulic architecture both between and within species, which is reflected in the huge variability in tree form and function across scales. It has been suggested that the observed differences could mirror the variability in environmental conditions experienced by different species and individuals, resulting in an optimal behaviour under the pressure of evolutionary processes (Magnani et al., 2002).

‘A more fundamental question is whether the pattern observed at a regional level is the result of species adaptation to site-specific conditions through the emergence of local ecotypes, or of structural and functional acclimation and phenotypic plasticity.’

In this issue of New Phytologist, Martínez-Vilalta et al., (pp. 353–364) explore the geographic pattern of hydraulic architecture in Pinus sylvestris, which, because of its wide natural range as well as its ecological and productive relevance, has been the subject of a number of related studies and can be therefore viewed as a de facto model species in forest tree functional and evolutionary ecology. Seen in the context of this large body of studies on the geographic and genetic variability of functional traits, the article provides not only new valuable information (in terms of parameters explored and their correlation and trade-offs), but also important hints of how evolutionary strategies could differ at the intraspecific and interspecific levels, and for different traits and processes.

The authors did not find evidence at the intraspecific level of some of the associations and trade-offs between hydraulic traits that have been commonly reported across species. An interesting case in point is the observation of a lack of variability across the entire latitudinal range explored (from Finland to Spain and southern Italy) in xylem vulnerability to embolism, which is determined by tracheid fine anatomical features (Hacke & Jansen, 2009) and appears to determine the minimum water potential the plant can tolerate (Jacobsen et al., 2007). By contrast, large interspecific differences in vulnerability to xylem cavitation have been found among coniferous and other evergreen species (Maherali et al., 2004), which were related to mean annual precipitation. A similar pattern has been recently observed among a range of shrub species (Bhaskar et al., 2007), after accounting for phylogenetic effects. Can the limited variability reported here for P. sylvestris populations (and for P. ponderosa, described by Maherali & DeLucia, 2000) be extended to the entire genus, as suggested by Martínez-Vilalta et al. (2004)? A similar homeostatic canalization in the face of environmental variability has been reported for Quercus wislizenii adult trees by Matzner et al. (2001), but further research is needed on other species before any general conclusions can be drawn.

Two other crucial questions are raised by the analysis of Martínez-Vilalta and colleagues, both of which are related to the interpretation of the observed pattern at a regional level and to the design of future studies.

In order to avoid the confounding effects of the correlation commonly observed between climatic variables, the authors bundle several of them together, through a principal component analysis, into an index of climate dryness. This was found to be closely related to a number of functional parameters, such as the ratio between transpiring foliage and branch sapwood area (AL : AS). Indeed, the geographic variability across Europe in AL : AS has been variously attributed in the past to the effects of evaporative demand (Berninger et al., 1995) or temperature (Palmroth et al., 1999), both contributing to the new dryness index. It would be important, however, to disentangle the effects of individual variables, as their trajectories in response to future climate change could differ. This is no easy task, as truly manipulative studies are hardly feasible in slow-growing forest trees. The only way forward lies in the careful design of future regional studies, which should locate study plots on the basis not of their geographic distribution, but of a homogeneous exploration of the climate space (e.g. temperature × drought combinations).

A more fundamental question is whether the pattern observed at the regional level is the result of species adaptation to site-specific conditions through the emergence of local ecotypes, or of structural and functional acclimation and phenotypic plasticity. This has great ecological and evolutionary significance. On the one hand, if the variability currently observed was the result of continuous acclimation to variable conditions, we could expect local individuals to keep changing in response to future climate change and for pine trees from Germany to resemble their Spanish kin in a few decades’ time without a reduction in population fitness or an increase in vulnerability. The opposite would be true if observed differences were the result of selection and therefore genetically encoded, as they could provide optimal fitness under present, but not future, conditions (Magnani et al., 2002). Under an evolutionary perspective, a large phenotypic plasticity could be explained by the variety of environmental conditions experienced by trees over their lifetime, which would prevent stabilizing selection from prevailing (Bradshaw, 1965).

Which process does prevail? As an example, Palmroth et al. (1999) also reported a close relationship between the leaf-to-sapwood area ratio of different P. sylvestris provenances and annual temperature. Interestingly, the same relationship was observed in situ and in a common garden experiment, considering the temperature at the site of seed origin. This would suggest a strong genetic control and a general lack of phenotypic plasticity for this hydraulic trait (see Fig. 1). This contrasts with other key functional traits reported by Oleksyn et al. (2003). In an analysis of foliage nutrient concentration in P. sylvestris from different geographic origins, the geographic pattern recorded in situ (determined by the interaction between ecotypic adaptation and acclimation to local conditions) was found to differ substantially from that observed under constant environmental conditions in a common garden experiment (determined only by ecotypic adaptation).


Figure 1.  Schematic representation of the possible response of plant functional traits to environmental variables across regional transects (thin line) or to prevailing conditions at the site of seed origin in common garden experiments (thick line). Four possible cases can be distinguished: (a) a parallel response at the two scales is indicative of directional selection and long-term adaptation to local conditions; (b) by contrast, a lack of variability in common garden experiments suggests that the pattern observed at the regional level is the result of acclimation to site conditions; (c) contrasting responses at the two scales demonstrate a strong genetic × environment interaction, eliciting further studies on the genetic control of phenotypic plasticity; (d) finally, a limited variability of the trait across both scales (e.g. for minimum water potential homeostasis, constant vulnerability to embolism) demonstrates canalization as a result of homeostatic processes.

Download figure to PowerPoint

How can future experiments try and disentangle the role of adaptation and acclimation, and help us predict the impact of future conditions on forest tree function and vulnerability? Only a combination of in situ and ex situ measurements can provide the answer to this question, exploiting the network of provenance trials established with a different purpose by forest services over the years. The answer could differ depending on the trait and scale considered, as hinted by Martínez-Vilalta and colleagues.


  1. Top of page
  2. References