A relationship between humidity response, growth form and photosynthetic operating point in C 3 plants

Transcription

1 Plant, Cell and Environment (1999) 22, A relationship between humidity response, growth form and photosynthetic operating point in C 3 plants P. J. FRANKS 1,2 & G. D. FARQUHAR 1 1 Environmental Biology Group, Research School of Biological Sciences, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia and 2 Department of Tropical Plant Sciences, James Cook University, PO Box 6811, Cairns 4870, Australia ABSTRACT Gas exchange experiments were performed with 13 plant species that differ from each other in growth-form and natural habitat. These comprised three herbaceous species, two ferns, two temperate deciduous trees, five rainforest trees and one liana from wet tropical forest. The aims were to investigate whether plants of similar growth-form and from similar habitats tended to respond similarly to a change in leaf-to-air vapour pressure difference (VPD), and to compare their ratio of intercellular to ambient partial pressures of CO 2 for given conditions. Leaves were subjected to a step change in VPD and the initial and final steady rates of transpiration were used to calculate an index of sensitivity, f, which enabled comparison of species. The results suggest that species of similar growth-form and habitat respond similarly to increasing VPD, with the temperate deciduous trees undergoing a greater reduction in stomatal conductance than the herbaceous plants in wellwatered soil. Also, for these experimental conditions, the ratio of leaf internal to ambient CO 2 partial pressure (p i /p a ) was positively correlated with both CO 2 assimilation rate and stomatal insensitivity to VPD, across the 13 species. The results are discussed in terms of growth strategies and possible advantages and limitations of hydraulic systems in different plants. Key-words: Argyrodendron peralatum; Austrobaileya scandens; Dysoxylum gaudichaudianum; Ginkgo biloba; Idiospermum australiense; Neolitsea dealbata; Osmunda regalis; Prunus armeniaca; Pteridium aquilinum; Toona australis; Tradescantia virginiana; Triticum aestivum; Vicia faba; humidity response, leaf gas exchange, stomata. INTRODUCTION The success of higher plants in their colonization of land under a wide range of climatic extremes is attributable largely to their ability to compromise between adequate carbon uptake and adequate hydration. With the evolution of novel carbon fixing strategies like C 4 and crassulacean Correspondence: P. J. Franks; peter.franks@jcu.edu.au 2 Present address: Department of Tropical Plant Sciences, James Cook University, PO Box 6811, Cairns 4870, Australia. acid metabolism (CAM) photosynthesis, some plants are able to vastly improve their chances of survival in dry environments. However, for the bulk of the world s flora, having C 3 photosynthesis, the need for comparatively high stomatal conductance and exposure to daily fluctuations in leaf-to-air vapour pressure difference (VPD) mean an ever present threat of dehydration through transpirational water loss. It has been shown in many plants that there is a reduction in steady-state stomatal conductance with an increase in VPD (see review by Monteith 1995) and this is widely interpreted as a means by which plants can minimize water loss for a given carbon gain (Cowan 1977; Cowan & Farquhar 1977; Farquhar 1978; Farquhar, Schulze & Küppers 1980; Guehl & Aussenac 1987; Makela, Berninger & Hari 1996). The underlying mechanism of this apparently beneficial adaptation is poorly understood (Grantz 1990; Cowan 1994; Franks, Cowan & Farquhar 1997; Jones 1998). Furthermore, despite the many published gas exchange data showing the responses of different species to VPD, it is not clear whether or not certain modes of response are characteristic of certain types of plants. It is therefore difficult to gain an overall perspective of the prominence of the VPD response across the broader plant community. We performed gas exchange measurements on four anatomically and ecologically distinct groups of plants to investigate the possibility that plants of similar growth form and from similar natural habitats tended to respond similarly to a change in VPD. The four groups were (1) herbaceous crop plants (2) ferns/herbaceous plants from moist habitats (3) rainforest trees, and (4) deciduous trees from temperate environments. MATERIALS AND METHODS Plant material The study was restricted to plants with C 3 photosynthesis and comprised 13 species. These included three herbaceous species (Triticum aestivum L. var. Matong, Tradescantia virginiana L. and Vicia faba L.), two ferns (Pteridium aquilinum (L.) Kuhn and Osmunda regalis L.), two temperate deciduous trees (Prunus armeniaca L. and Ginkgo biloba L.), five rainforest trees (Toona australis (F. Muell.) Harms, Argyrodendron peralatum (Bailey) Edlin, Idiosper Blackwell Science Ltd 1337

2 1338 P. J. Franks and G. D. Farquhar mum australiense (Diels) S. T. Blake, Dysoxylum gaudichaudianum (Adr. Juss.) Miq. and Neolitsea dealbata (R. Br) Merr.) and one liana from wet tropical forest (Austrobaileya scandens White). All plants, except the deciduous trees, were grown in a glasshouse (day/night air temperature 30/25 C, relative humidity 80%) under 50% shade cloth. Ginkgo biloba and P. armeniaca were grown under 50% shade cloth in the open, experiencing daytime temperatures similar to those for glasshouse plants during their summer growth period. Plants were well watered and fertilized with a controlled release fertilizer (Osmocote Plus; Grace-Sierra Pty Ltd, Castle Hill, Australia). Because of their small size, T. virginiana and Vicia faba were grown in 3 L pots, and T. aestivum was grown in 1 L pots.all other plants were grown in 10 L pots. Soil used was 5 : 2:2:1, compost :sand:peat: perlite. Tradescantia virginiana, A. scandens and the ferns were cloned from root cuttings, and the remaining species were grown from seed. Only new, fully expanded leaves/leaflets/fronds were used in gas exchange measurements. The third leaf was used in gas exchange experiments with T. aestivum and V. faba. Trees were all about 1 m in height when used in gas exchange experiments. Gas exchange measurements Experiments were performed with an open-flow gas exchange system. Details of this system have been given by Brugnoli et al. (1988). All experiments were performed during the natural photoperiod, with measurements limited to one leaf per day. Measurements were made on a single leaf for n = 3 4 individuals within a species. On the evening prior to an experiment the plant was brought to the laboratory and kept well watered in darkness or very low light (less than 5 mmol m -2 s -1 ) until the next morning when a leaf (or leaflet in the case of large compound leaves) was sealed into the gas exchange chamber in which the desired initial conditions (temperature, air vapour pressure deficit, CO 2 concentration) had been pre-set. Light was then increased to the desired level and steady-state values of transpiration rate (E), CO 2 assimilation rate (A), leaf conductance to water vapour (g) and ratio of intercellular to ambient CO 2 partial pressure (p i /p a ) were obtained. Leaf boundary layer conductance (5 mol m -2 s -1 for wet filter paper) was assumed to be sufficiently large to be unimportant in this analysis. The VPD was kept at the desired level (see below). In all experiments photosynthetic photon flux density (PPFD) was 400 mmol m -2 s -1 and ambient CO 2 concentration was 340 mmol mol -1. To ensure that bulk leaf water potentials were similarly high for each experiment, the soil was kept well watered, and transpiration from surrounding leaves/fronds was minimized by shielding from direct light. By manipulating the environment of a single leaf, rather than the whole canopy, information is gained about the leaf as a hydraulic unit, rather than the whole plant. Although it is possible that stomatal closure will be more pronounced following an increase in canopy transpiration, compared with the response to an increase in transpiration rate from a single leaf, we attempted to maintain a relatively high and constant water potential in leaf xylem to provide a common baseline for comparison of different species. Leaf temperature was 30 C (±1 C). Although the species in this study experience a variety of temperature ranges in their natural habitats, the optimum temperature for photosynthesis is thought to be close to 30 C for most of them (Larcher 1995). It is for this reason that, for the purpose of comparing gas exchange in such a diverse group of species, mean daily glasshouse temperature and leaf temperature during gas exchange measurements were both maintained at approximately 30 C. Normalizing the response to change in VPD The group of plants studied here, being of wide ranging growth form and natural habitat, operate over widely differing ranges of CO 2 assimilation rate and stomatal conductance for given ambient conditions. This presents some difficulty when attempting to compare the responses of these plants to perturbations. Our solution to this was to normalize the VPD response for a standard increment in VPD from 1 to 2 kpa.the normalized response is expressed as the final steady-state transpiration rate, for a given increase in VPD, relative to what it would have been if no change in stomatal conductance had occurred. Referring to Fig. 1, this ratio f is given by f = b/(a + b), where b is the actual transpiration rate at 2 kpa and (a + b) the rate of transpiration that would have occurred had there been no change in steady-state stomatal conductance following the VPD change. We will refer to f as an index of stomatal sensitivity to VPD, although, as f is negatively related to the magnitude of the VPD response, it could be referred to as an index of insensitivity. In theory the lower limit of f is zero, occurring if stomata were to close completely following the VPD change, reducing transpiration rate to zero. This would also require a cuticular conductance of zero. The upper limit of f is 1, and occurs if there is no change in stomatal conductance following the increase in VPD. In practice f will lie somewhere between 0 and 1, and in almost all cases between 0 5 and 1 (see Appendix). In most cases the increment in VPD was applied in one single step to minimize the confounding effects of diurnal rhythms and photosynthetic time elapsed (Franks et al. 1997). Where VPD could not be controlled at exactly 1 or 2 kpa, E at exactly 1 or 2 kpa was obtained by extrapolation. RESULTS The results of the VPD experiments are presented in Fig. 2 (see also Table 1). By normalizing the data in terms of the index of stomatal sensitivity to VPD, f, it was possible to make some semi-quantitative comparisons between this diverse group of plants. The 13 species are divided into four groups on the basis of growth form and habitat: (1) herbaceous crop plants (2) ferns/herbaceous plants from moist areas (3) rainforest trees (including the tropical liana A scandens) and (4) temperate deciduous trees. A one-way

3 Humidity response and growth form 1339 Figure 1. An example (using data for Argyrodendron peralatum) of how the response to a change in VPD was normalized. Data points indicate steady-state transpiration rates E 1 and E 2 at different VPD. The normalized response, f, is given by the ratio b/(a + b), where a + b is the steady-state transpiration rate with no change in conductance, and a is the reduction in transpiration due to stomatal closure. E 2 = fe 1. Alternatively, f = g 2 /g 1, where g 2 and g 1 are leaf conductances to water vapour at E 2 and E 1, respectively. Because this ratio changes with the magnitude of the increment in VPD, as well as the magnitude of initial and final VPDs, f was calculated for the same VPD increment for each species (1 2 kpa). Leaf irradiance 400 mmol m -2 s -1 ; leaf temperature 31 C; plant well-watered. analysis of variance reveals mean f between the four groups differs significantly (F = 34 2, P < 0 001). Mean f within each group did not differ significantly at the 0 1% level. The temperate deciduous trees were the most sensitive to an increase in VPD. Reduction in stomatal conductance was so dramatic in these plants following the increase in VPD that transpiration rate was held almost constant. For these plants transpiration rate remained almost constant following a further increase in VPD (data not shown). The next most responsive to an increase in VPD were the rainforest trees, of which D. gaudichaudianum was least sensitive. The least responsive to a step increase in VPD, under these conditions, were the herbaceous crop plants. The average f for this group was 0 82, indicating little reduction in stomatal conductance following an increase in VPD. Instantaneous water-use efficiency, defined as the ratio A/E, may be represented by a complementary measure p i /p a, where p i and p a are the partial pressures of CO 2 in air inside and outside the leaf, respectively (Farquhar, Ehleringer & Hubick 1989). Created from gas exchange data collected during the VPD experiments, Fig. 3 shows the relationship between mean species p i /p a and assimilation rate. Viewing each point as the mean operating point for that species under the given conditions, it can be seen that Figure 2. A comparison of stomatal sensitivity to leaf-to-air VPD, indicated by f (see Fig. 1), for a range of species differing in growth-form and habitat. The lower the value of f, the more sensitive the plant is to VPD. For all experiments, leaf irradiance was 400 mmol m -2 s -1, leaf temperature was 30 C (± 1 C) and ambient CO 2 mole fraction was 340 mmol mol -1. Plants were all grown under similar conditions. n = 3 4; for Idiospermum n = 2. Table 1. Calculated mean leaf conductance to water vapour at 1 kpa VPD (g 1, mol m -2 s -1 ) and 2 kpa VPD (g 2, mol m -2 s -1 ) for the species studied. Conditions as for Fig. 2 Species (n) g 1 (±SE) g 2 (±SE) Argyrodendron peralatum (3) ± ± Austrobaileya scandens (3) ± ± Dysoxylum ± ± gaudichaudianum (3) Ginkgo biloba (3) ± ± Idiospermum australiense (2) ± ± Neolitsea dealbata (3) ± ± Osmunda regalis (3) ± ± Prunus armeniaca (4) ± ± Pteridium aquilinum (4) ± ± Toona australis (3) ± ± Tradescantia virginiana (4) ± ± Triticum aestivum (4) ± ± Vicia faba (4) ± ± A is positively correlated with p i /p a (correlation coefficient 0 85). It was also found that across the entire group of plants, f is positively correlated with A (correlation coefficient 0 76; Fig. 4) and p i /p a (correlation coefficient 0 94; Fig. 5). DISCUSSION Transpiration rate and VPD The results presented in Fig. 2 show varied modes of plant response to VPD. Within certain groups of plants the responses were very similar. We caution, however, that sample sizes are small and it is possible that there is a difference between the means obtained for growth forms in

4 1340 P. J. Franks and G. D. Farquhar Figure 3. Plot of mean species CO 2 assimilation rate (A) against mean species ratio of leaf intercellular CO 2 partial pressure to ambient CO 2 partial pressure (p i /p a ), for a range of C 3 species differing in growth-form and habitat. Mean ± SE; n = 3 4; correlation coefficient, r, is 0 85; leaf irradiance 400 mmol m -2 s -1 ; leaf-to-air VPD 1 ± 0 2 kpa, leaf temperature 30 C (± 1 C) ambient CO 2 partial pressure 34 Pa, plants well-watered and grown under similar conditions. For clarity, some species are represented by numbers. 1: Tradescantia virginiana; 2: Dysoxylum gaudichaudianum; 3: Prunus armeniaca; 4: Idiospermum. australianse. Straight line is an error-weighted least squares fit (y = 0 015x , r 2 = 0 72). Figure 4. Plot of mean species CO 2 assimilation rate (A), at 1 ± 0 2 kpa VPD, against mean species index of sensitivity to VPD (f). Conditions as for Fig. 2 and 3. Straight line is an error-weighted least squares fit (y = 25x 5 8, r 2 = 0 58).

5 Humidity response and growth form 1341 Figure 5. Plot of mean species ratio of leaf intercellular CO 2 partial pressure to ambient CO 2 partial pressure ( p i /p a ), at 1 ± 0 2 kpa VPD, against mean species index of sensitivity to VPD (f). Conditions as for Fig. 2 and 3. Straight line is an error-weighted least squares fit (y = 0 57x + 3 9, r 2 = 0 89). this study and the true means for these growth forms. Our study includes some species for which there already exists a considerable body of gas exchange literature, particularly in relation to VPD experiments. Examples are T. virginiana, V. faba and P. armeniaca. However, there are almost no gas exchange data of this type for ferns (even the very common ferns used in this study) and almost as few data of this type for Australian tropical rainforest trees. These results therefore help to fill some of the gaps in the literature and offer a broader insight into the nature of the VPD response across a variety of plant families. For the species that have in the past been the subject of VPD experiments, there is good qualitative agreement between the results in our study and those in the literature. Many of the earlier VPD experiments which employed modern gas exchange equipment and theory were performed on herbaceous crop plants, and one of the more comprehensive of these was the study by Rawson, Begg & Woodward (1977). In their study they compared the VPD responses of a number of C 3 species, including wheat, soybean, sunflower and sorghum, to step changes in VPD over the range kpa.they found little or no response to VPD in these species, which were under high light and were well watered. In our experiments the herbaceous crop plants wheat (T. aestivum) and broad bean (V. faba) were the least sensitive to changes in VPD. Yong, Wong & Farquhar (1997) show a similarly small response in soybean and cocklebur. Aston (1976) found sunflower to be considerably more sensitive than this, although in his experiments the change in VPD was imposed on the whole canopy. This added perturbation to leaf water status by canopy transpiration, perhaps more indicative of natural conditions, probably contributed to the differences between results of Aston (1976) and Rawson et al. (1977) for sunflower. However, Aston (1976) also observed little change in leaf water status, as measured with a b-gauge. Presumably, in the interest of maximizing productivity, crop plants will have inadvertently been selected for high stomatal conductances. This at least seems to be the case with semi-dwarf bread wheat varieties (Sayre 1996). These high stomatal conductances may contribute to very high rates of transpiration under natural conditions. It could be argued that while water availability is non-limiting, these plants achieve their relatively high rates of productivity with correspondingly high rates of water loss. In fact it is generally found that herbaceous crop plants display the lowest water-use efficiencies (defined here as unit dry matter production per unit water transpired) amongst vascular C 3 plants (Maximov 1929; Larcher 1995). Anatomical and physiological observations of herbaceous crop plants would support the theory that the drop in water potential at high transpiration rates, associated with high VPD, is minimized in these plants by an efficient water conducting system. Indeed, a study by Turner, Schulze & Gollan (1984) showed that leaf water potential declined more in woody than herbaceous species as a result of increasing VPD. The largest measured water potential gradients along wheat and lupin plants, estimated from the

6 1342 P. J. Franks and G. D. Farquhar data of Gallardo et al. (1996), are between 0 5 and 1 5 MPa m -1. Begg & Turner (1970) measured up to 0 8 MPa m -1 in tobacco. In contrast to this, water potential gradients in several species of fern, having similar stature but with their xylem lacking efficient water conducting vessels, range from 2 to 24 MPa m -1 (Woodhouse & Nobel 1982). Therefore it is possible that crop plants such as wheat and broad bean will maintain high transpiration rates at high VPDs until external factors such as local short-term or wide-spread longer term soil water deficits lead to reduced stomatal conductances, through turgor loss and/or the generation of abscisic acid. In experiments with castor bean (Ricinus communis) Macklon & Weatherley (1965) found that the leaf water potential of transpiring plants rooted in water changed little when transpirational flux was more than doubled. They also showed that transpiration rates at similar VPD were twice as high in plants rooted in water, compared with those rooted in soil, and that, unlike plants rooted in water, plants rooted in soil experienced a significant decrease in leaf water potential following an increase in transpiration rate. Macklon & Weatherley (1965) concluded from this that soil hydraulic resistance was the primary cause, in these plants, of water deficits in the leaf at high VPD. This could explain why herbaceous crop plants grown at high VPD exhibit reduced yields, even when well supplied with water (Woodward & Begg 1976; Sinclair, Tanner & Bennett 1984). Recently, Tardieu & Simonneau (1998) reviewed the variability of the VPD response and highlighted the potential influence of both soil moisture and soil hydraulic conductance. It would appear therefore that water-use efficiency in the field, at least for certain crop plants, has much to do with soil properties. In this sense, it could be expected that water-use efficiency in plants with relatively high hydraulic conductance, such as wheat, would be increased by higher soil hydraulic resistances. This topic has undergone little investigation, but the principle is supported by the results of Masle & Farquhar (1988), who observed increased wateruse efficiency in wheat as a result of increased soil penetration resistance. Because of the strong negative correlation between cereal dry matter production and water-use efficiency in well-watered crops (Condon, Richards & Farquhar 1987; Ehdaie et al. 1991), crop breeders are presented with quite a challenge in attempting to produce high yielding, water-use efficient crops. With the help of rapid screening methods, such as the measurement of carbon isotope discrimination (Farquhar et al. 1989), plant physiologists are helping to make some progress towards this goal (Hubick & Farquhar 1987; Dingkuhn et al. 1991; Wright, Rao & Farquhar 1994; Hall et al. 1994). However, more knowledge about stomatal functioning in response to VPD might provide an additional source of guidance in crop breeding programs. The ferns P. aquilinum and O. regalis, together with the herbaceous T. virginiana, were remarkably similar in terms of their stomatal sensitivity to a change in VPD.As a group, they were slightly more sensitive than the crop plants, but less sensitive than the temperate deciduous trees, or most of the rainforest trees. These results are consistent with those of Roberts, Wallace & Pitman (1984), who found that stomata of P. aquilinum forming a forest understorey were less sensitive to VPD than those of Pinus sylvestris forming the forest canopy. It should be noted that our survey does not include representatives of the small number of fern species (many with CAM photosynthesis) adapted to very dry environments. Although few studies investigating stomatal response to VPD have been carried out with ferns, there have been numerous VPD experiments conducted with T. virginiana. The behaviour of T. virginiana is best summarized in the results of Nonami, Schulze & Ziegler (1990).The value of f we calculated from their data was the same as that in Fig. 2 for identical conditions. Furthermore, the data of Nonami et al. (1990) reveal that f would decrease dramatically under conditions of moderate water stress (y xylem ª 0 25 MPa). This effect can be seen in the results of Turner, Schulze & Gollan (1985) for sunflower. It could also be the case with the ferns in this study, perhaps to an even greater extent considering the relatively high water potential gradients in ferns. It is possible that for ferns the inability to substantially reduce stomatal conductance at high VPD, in conjunction with poor xylem hydraulic conductivity, leads to low water potentials and reduced growth in drier environments. There are many observations that suggest a positive correlation between mature frond height and moisture availability for P. aquilinum and Osmunda regalis, as well as other species of ferns (Moore 1860; Tryon & Tryon 1982). Further quantitative physiological studies are necessary in order to determine to what extent limited stomatal closure at high VPD and an inefficient hydraulic system confine many fern species to moist environments. Results obtained for the young rainforest trees showed a varied response to VPD. This could reflect the anatomical and taxonomic diversity of the group. Despite this, these rainforest trees seem to operate between the temperate deciduous trees and the ferns in terms of f. The most sensitive to VPD was N. dealbata and the least sensitive was D. gaudichaudianum. Both these species are found in their mature form in the subcanopy or canopy of wet tropical forest, but both are also highly competitive early successional species and therefore adapted to conditions of high evaporative demand. The marked difference between N. dealbata and D. gaudichaudianum in response to VPD suggests therefore that the degree of stomatal closure in dry air is not necessarily indicative of the environment in which a plant occurs naturally. Rather, it may indicate something about the inherent hydraulic limitations for which a plant must somehow compensate. Prunus armeniaca and G. biloba were both highly sensitive to VPD and this is consistent with studies on other temperate deciduous trees. In a field study, Schulze et al. (1972) measured substantial reductions in stomatal conductance of irrigated P. armeniaca trees in response to increasing

7 Humidity response and growth form 1343 VPD. Similarly dramatic responses have been observed in Malus pumila (Fanjul & Jones 1982) and Acer campestre (Küppers 1984). It is apparent that under the conditions in which our data were gathered, f was constrained between approximately 0 5 and 1. This result awaits a full theoretical explanation. If the VPD response could be attributed to a purely hydraulic feedback system then a simple feedback model, such as that described in the Appendix, can explain the constraint on f. However, if the VPD response were due solely to hydraulic feedback, some plants (for example the temperate deciduous trees in this study) would appear to exhibit feedback loop gains approaching negative infinity. It is unlikely that systems with such high loop gain could exhibit the stability observed during these gas exchange measurements. This observation, together with data from other studies suggesting that f may in some cases be less than 0 5, supports the hypothesis that a hydraulic feedback mechanism may control only part of the VPD response, in conjunction with VPD-stimulated metabolic modulation of ion fluxes (Cowan 1977; Lösch & Schenk 1978; Grantz & Zeiger 1986; Franks et al. 1997). Trends in leaf hydraulic parameters The hydraulic feedback loop affecting stomata, easily simplified to the form shown in the appendix, is in fact rather complex and involves components with non-linear dynamic characteristics. For example, the relationship between stomatal aperture and guard cell pressure has long been recognized as non-linear and confounded by the mechanical influence of epidermal and subsidiary cell turgor (Raschke & Dickerson 1972; Meidner & Edwards 1975; Meidner & Bannister 1979; Franks, Cowan & Farquhar 1998). Recent work suggests that these characteristics could differ fundamentally between plant taxa (Franks et al. 1998). Although measurement of many of the feedback loop components is technically difficult, the characteristics of some can be inferred from gas exchange data. The following analysis serves to briefly illustrate how differences in guard cell wall properties and local hydraulic conductances might contribute to the trends observed in this study. With reference to the appendix, the open loop gain H of the hydraulic feedback loop may be summarized as E g H = (1) g E The change in transpiration rate de associated with a change in stomatal conductance dg is given by E w = (2) g, where w is the difference between the mole fractions of water vapour inside and outside the leaf (proportional to VPD). Let Y w equal the water potential at the guard cell wall. Water potential within guard cells Y g will tend to equilibrate with Y w. A change in Y w, dy w, as a result, for example, of a change in transpiration rate, will tend to change guard cell water potential. The final change in guard cell water potential will depend also on any change in guard cell osmotic pressure. We simplify this analysis by assuming (1) guard cell osmotic pressure P g and epidermal cell osmotic pressure P e are maintained constant, and (2) hydraulic conductances and sites of evaporation in the epidermis are such that the draw-down in water potential is sufficiently greater in guard cells than in epidermal cells so that passive opening as a result of epidermal turgor loss is counteracted. Arguments in support of these assumptions have been put forward by Cowan (1994). In an alternative model, Haefner, Buckley & Mott (1997) assume P g to be a function of epidermal turgor. Following assumptions (1) and (2), a change in Y w will cause an equal change in guard cell hydrostatic pressure P g, which manifests itself as a change in guard cell volume dv and stomatal aperture or conductance, dg. For a small change in Y w, and hence P g, the accompanying change in guard cell volume dv is related to guard cell elasticity e and initial guard cell volume V: d e = V P g (3) dv Taking, in this instance, guard cell volume as proportional to stomatal conductance, Eqn 3 may be rewritten as dy w e* = g (4) dg For a given change in transpiration rate de,dy w is determined by plant hydraulic conductance L: L =- E (5) y w Equation 1 can be expanded to show the influence of Y w : E g H = y w (6) g y w E Equations 2, 4 and 5 can be rearranged and combined to replace the terms in Eqn 6, expressing H in the form H w g 1 =- (7) e * L Taking e* and L in Eqn 7 as being representative of bulk average guard cell elasticity and xylem-to-guard-cell hydraulic conductance, respectively, it can be seen how differences in these physical properties might affect the hydraulic feedback loop gain and hence leaf gas exchange at different VPD. The more elastic the guard cells (small e*) and/or the lower the hydraulic conductance L, the higher the loop gain H and, potentially, the higher the stomatal sensitivity to VPD. Using Eqn 7, we estimated values for the combined term e*l from our gas exchange data. When f is plotted against these estimates of e*l, the trend across the whole range of species shows f to be related exponentially to e*l (Fig. 6). The plant groupings for f defined in Fig. 2 can be transferred directly to f in Fig. 6. Hence those with highest f (the two herbaceous crop plants) have the highest e*l. It is unclear which of the two components, e*

8 1344 P. J. Franks and G. D. Farquhar potassium content following stomatal closure in response to dry air, implying some form of active regulation of guard cell ion content in response to VPD. However, it is still not known how P g is affected by a change in VPD. Photosynthetic operating point Figure 6. Plot of stomatal index of sensitivity to VPD, f, against Le* for all the species combined. The component L represents the hydraulic conductance in the vicinity of guard cells. The component e* represents guard cell elasticity. The combined term e*l was calculated on the basis of a hydraulic model (see Discussion). Exponential line is a least squares fit (y = e (-x/ ), r 2 = 0 90). or L, is driving the trend represented in Figs 2 and 6 as little is known about the exact value or natural variability of both these terms. It is also unclear whether, on the other hand, the trend is driven by a loss of the possible metabolic component of the VPD response in less sensitive plants. It is possible that both e* and L could vary during the development of an individual, as a result of environmental stresses. However, the similarities observed here between plants of similar growth form imply that e* and L are constrained genetically, in a similar manner to xylem hydraulic conductance. Closer examination of Fig. 3 raises further questions about the mechanism of the VPD response. The purely hydraulic model applied here suggests that those plants with highest stomatal sensitivity to VPD require extremely elastic guard cells and/or extremely low hydraulic conductances in the vicinity of the guard cells. In fact it would require L to be perhaps orders of magnitude smaller than values measured in other regions of the plant hydraulic system. To verify this, some method for measuring or estimating Y w will have to be developed, perhaps along the lines of work by Shackel (1987) and Nonami et al. (1990) who used a pressure probe technique to measure epidermal cell turgor in response to changing VPD. Insights will also be gained from further research into the nature of e and e*, through measurement of guard cell inflation characteristics (e.g. Franks et al. 1998). Furthermore, the above analysis assumes as a first approximation constant P g. Implicit with this assumption is that some solute efflux from guard cells must occur to maintain P g constant as guard cell volume decreases. Using a staining technique, Lösch & Schenk (1978) observed a decrease in guard cell Figure 3 is a composite picture of how all the plants used in this study relate to one another in terms of their operating point in the given conditions. The data suggest a negative correlation between productivity (represented by A) and water-use efficiency (represented by, and negatively related to, p i /p a ) for this diverse group of plants.the operating point for each species may represent the condition of optimal carbon gain with respect to water loss (von Caemmerer & Farquhar 1981). In any given conditions of light, ambient CO 2 concentration and stomatal conductance, leaf CO 2 assimilation rate is a function of both the capacity for the regeneration of the substrate for the enzyme ribulose bisphosphate (RuP 2 ) carboxylase-oxygenase (Rubisco) (which is often dominated by the thylakoid membrane electron transport capacity), together with Rubisco activity itself (Farquhar & von Caemmerer 1982).These processes colimit the rate of CO 2 assimilation.through varying stomatal conductance, the plant can operate anywhere on the A versus p i /p a curve for which p i /p a 1. Photosynthetic capacity may be defined as CO 2 assimilation rate at a given p i, for given ambient conditions. Alternatively, allowing for different stomatal sensitivities, photosynthetic capacity may be defined as the CO 2 assimilation rate at which the plant operates, for given ambient conditions. If it is to optimize carbon gain, A, with respect to water loss, E (von Caemmerer & Farquhar 1981; Cowan 1986), then the plant will assume a conductance, g, to keep the ratio of the sensitivities E/ g and A/ g constant at some value (Cowan 1977; Cowan & Farquhar 1977). There is often a large change in E/ A at the point of transition from Rubisco-limited to Rubisco-substrate-regenerationlimited CO 2 assimilation rate, so in many conditions the optimal solution is to be near this point of transition (von Caemmerer & Farquhar 1981). This point, C, is determined primarily by the ratio of potential electron transport rate (J) to Rubisco activity with saturating CO 2 (V max ). It has been shown that C is fairly conservative, that is, for a given growth irradiance, J and V max will tend to vary in proportion to one another, thereby maintaining the same p i /p a (von Caemmerer & Farquhar 1981). However, the results in Fig. 3 suggest that for a broad group of plants, varying widely in growth form and natural habitat, C may vary also (i.e. C 3 plants with inherently higher photosynthetic capacities may invest relatively more in thylakoid proteins). This is assuming that the operating points plotted in Fig. 3 correspond with the transition point C for these species. Theoretical studies based on the optimization of water use have predicted that such a correlation might be the case for a single plant (Cowan 1977; Cowan 1986). However, there is no reason to believe, a priori, that such a correlation should exist across a number of different

9 Humidity response and growth form 1345 species, as in Fig. 3. A shift in the ratio J/V max with increasing photosynthetic capacity has not been shown before with a broad range of species, although Watanabe, Evans & Chow (1994) reported a tendency for J/V max to be lower for modern Australian wheat cultivars when compared with those released at the turn of the century. Although Fig. 3 in itself is not proof of there being a natural tendency for a shift towards increasing J/V max with increasing photosynthetic capacity, there is additional evidence to suggest this might be the case. The ratio J/V max depends on the relative investment of nitrogen in the thylakoid proteins versus that allocated to Rubisco. Evans (1987), 1989) has discussed the nitrogen cost of thylakoids in detail. Terashima & Evans (1988) calculated the total nitrogen cost of thylakoids in spinach to be about 64 mol N mol -1 Chl.Evans & Seemann (1989) gave the same figure as an estimate of the nitrogen cost of the Calvin cycle enzymes (of which 76% is allocated to Rubisco). Provided irradiance is sufficiently high in relation to the light saturated potential rate of electron transport, J max, then J will increase almost in proportion to J max. Therefore, if chloroplasts maintain constant J/V max for increased photosynthetic capacity, the benefit of extra nitrogen investment will be shared about equally between thylakoid and soluble proteins.however,in terms of efficiency this may not be the best strategy. The same increase in photosynthetic capacity can be obtained with less nitrogen if the ratio J/V max is increased. This is possible because the thylakoid proteins having the strongest influence on J max (ATPase and cytochrome b/f complexes) make up only a small percentage of the total nitrogen allocated to thylakoids. For instance, Evans & Seemann (1989) described how a doubling of the cytochrome f content from 1 to 2 mmol cyt f mol -1 Chl could lead to a doubling of the electron transport capacity while increasing the nitrogen allocated to the thylakoids by only 18%. Regardless of the precise costing it would nonetheless make sense for a plant to favour an increase in J/V max when increasing its photosynthetic capacity. There is, however, a catch. Increasing photosynthetic capacity via a proportional increase in J and V max demands higher stomatal conductances, and if higher photosynthetic capacity is associated with an increase in J/V max, then even higher stomatal conductances are necessary. This would in turn place greater demand on the plant hydraulic system, for a given VPD, and therefore the benefits of an increase in J/V max will be enhanced by increased hydraulic conductance. The results presented in Fig. 3 are suggestive of an hypothesis that, across taxa, there is a shift towards increasing J/V max with increasing photosynthetic capacity. This trend might, in turn, be constrained by the capabilities of the plant hydraulic system. In this sense T. aestivum and V. faba, which are known to have highly conductive hydraulic systems, demonstrate the highest CO 2 assimilation rate and p i /p a under these experimental conditions. It remains to be seen how the slope of the regression line in Fig. 3 changes for different conditions. It is likely to be steeper for high light and shallower for low light conditions, but the ranking of species should stay relatively unchanged. The trends shown in Figs 3, 4 and 5 could indicate a general association between plant photosynthetic capacity and hydraulic capacity. It is likely that high stomatal conductance associated with high CO 2 assimilation rates are supported by comparatively high plant hydraulic conductances. Further experimental work is required in order to determine whether or not a strong VPD response is the result of an inferior plant hydraulic system. It is possible that because properties of the plant hydraulic system are highly conserved, plant gas exchange characteristics will be constrained by inherent plant hydraulic properties. A number of studies have already alluded to this. For example, Meinzer & Grantz (1991) and Meinzer, Saliendra & Crisoto (1992) discussed a close association between gas exchange and hydraulic conductance in sugar cane and coffee, respectively. Ehleringer (1994), referring particularly to evidence from studies on desert vegetation, argued that life-form characters, coupled with hydraulic properties, should impose constraints on actual photosynthetic rates. Lösch & Schulze (1994) proposed that stomatal function and plant structure are tightly coordinated, with each species developing a unique balance between leaf conductance and shoot hydraulic conductivity. At a larger scale Schulze et al. (1994) identified a correlation between maximum stomatal conductance and maximum assimilation rate inferred from leaf nitrogen concentration, within which distinct global vegetation types can be grouped. Our study reinforces these observations and provides a clearer picture of the relationship between gas exchange, plant growth form and habitat. Further studies are necessary to establish the exact nature of the relationship between photosynthetic operating point and the plant hydraulic system. ACKNOWLEDGMENTS We thank Professor I. R. Cowan and an anonymous referee for helpful comments on the manuscript. S. C. Wong, W. Coupland and P. Groeneveld provided excellent technical assistance. REFERENCES Aston M.J. (1976) Variation of stomatal diffusive resistance with ambient humidity in sunflower (Helianthus annus). Australian Journal of Plant Physiology 3, Begg J.E. & Turner N.C. (1970) Water potential gradients in field tobacco. Plant Physiology 46, Brugnoli E., Hubick K.T., von Caemmerer S., Wong S.C. & Farquhar G.D. (1988) Correlation between the carbon isotope discrimination in leaf starch and sugars of C 3 plants and the ratio of intercellular and atmospheric partial pressures of carbon dioxide. Plant Physiology 88, von Caemmerer S. & Farquhar G.D. (1981) Some relationships between the biochemistry of photosynthesis and the gas exchange of leaves. Planta 153, Condon A.G., Richards R.A. & Farquhar G.D. (1987) Carbon isotope discrimination is positively correlated with grain yield and dry matter production in field-grown wheat. Crop Science 27,

10 1346 P. J. Franks and G. D. Farquhar Cowan I.R. (1977) Stomatal behaviour and environment. Advances in Botanical Research 4, Cowan I.R. (1986) Economics of carbon fixation in higher plants. In On the Economy of Plant Form and Function (ed. T.J. Givnish), pp Cambridge University Press, Cambridge. Cowan I.R. (1994) As to the mode of action of the guard cells in dry air. In Ecophysiology of Photosynthesis (eds E.-D. Schulze & M.M. Caldwell), pp Springer-Verlag, New York. Cowan I.R. & Farquhar G.D. (1977) Stomatal function in relation to leaf metabolism and environment. Symposia of the Society for Experimental Biology 31, Dingkuhn M., Farquhar G.D., De Datta S.K. & O Toole J.C. (1991) Discrimination of 13 C among upland rices having different water use efficiencies. Australian Journal of Agricultural Research 42, Ehdaie B., Hall A.E., Farquhar G.D., Nguyen H.T. & Waines J.G. (1991) Water-use efficiency and carbon isotope discrimination in wheat. Crop Science 31, Ehleringer J.R. (1994) Variation in gas exchange characters among desert plants. In Ecophysiology of Photosynthesis (eds E.-D. Schulze & M.M. Caldwell). pp Springer-Verlag, New York. Evans J.R. (1987) The relationships between electron transport components and photosynthetic capacity in pea leaves grown at different irradiances. Australian Journal of Plant Physiology 14, Evans J.R. (1989) Photosynthesis and nitrogen relationships in leaves of C 3 plants. Oecologia 78, Evans J.R. & Seemann J.R. (1989) The allocation of protein nitrogen in the photosynthetic apparatus: costs, consequences, and control. In Photosynthesis (ed. W.R. Briggs), pp Alan R. Liss, Inc., New York. Fanjul L. & Jones H.G. (1982) Rapid stomatal responses to humidity. Planta 154, Farquhar G.D. & von Caemmerer S. (1982) Modelling of photosynthetic response to environmental conditions. In Encyclopedia of Plant Physiology. New Series,Vol. 12b (eds O.L. Lange, P.S. Nobel, C.B. Osmond & H. Ziegler), pp Springer-Verlag, Berlin. Farquhar G.D. (1978) Feedforward responses of stomata to humidity. Australian Journal of Plant Physiology 5, Farquhar G.D., Ehleringer J.R. & Hubick K.T. (1989) Carbon isotope discrimination and photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology 40, Farquhar G.D., Schulze E.-D. & Küppers M. (1980) Responses to humidity by stomata of Nicotina glauca L. & Corylus avellana L. are consistent with the optimisation of carbon dioxide uptake with respect to water loss. Australian Journal of Plant Physiology 7, Franks P.J., Cowan I.R. & Farquhar G.D. (1997) The apparent feedforward response of stomata to air vapour pressure deficit: Information revealed by different experimental techniques with two rainforest trees. Plant, Cell and Environment 20, Franks P.J., Cowan I.R. & Farquhar G.D. (1998) A study of stomatal mechanics using the cell pressure probe. Plant, Cell and Environment 21, Gallardo M., Eastham J., Gregory P.J. & Turner N.C. (1996) A comparison of plant hydraulic conductances in wheat and lupin. Journal of Experimental Botany 47, Grantz D.A. (1990) Plant response to atmospheric humidity. Plant, Cell and Environment 13, Grantz D.A. & Zeiger E. (1986) Stomatal responses to light and leaf-air vapor pressure difference show similar kinetics in sugarcane and soybean. Plant Physiology 81, Guehl J.M. & Aussenac G. (1987) Photosynthesis decrease and stomatal control of gas exchange in Abies alba Mill. In response to vapour pressure difference. Plant Physiology 83, Haefner I.W., Buckley T.N. & Mott K.A. (1997) A spatially explicit model of patchy stomatal responses to humidity. Plant, Cell and Environment 20, Hall A.E., Richards R.A., Condon A.G., Wright G.C. & Farquhar G.D. (1994) Carbon isotope discrimination and plant breeding. Plant Breeding Reviews 4, Hubick K.T. & Farquhar G.D. (1987) Carbon isotope discrimination: Selecting for water use efficiency. The Australian Cotton Grower August-October, Jones H.G. (1998) Stomatal control of photosynthesis and transpiration. Journal of Experimental Botany 49, Küppers M. (1984) Carbon relations and competition between woody species in a central European hedgerow. II. Stomatal responses, water use and hydraulic conductivity in the root/leaf pathway. Oecologia 64, Larcher W. (1995) Physiological Plant Ecology. Springer-Verlag, New York. Lösch R. & Schenk B. (1978) Humidity response of stomata and the potassium content of guard cells. Journal of Experimental Botany 29, Lösch R. & Schulze E.-D. (1994) Internal coordination of plant responses to drought and evaporative demand. In Ecophysiology of Photosynthesis (eds E.-D. Schulze & M.M. Caldwell), pp Springer-Verlag, New York. Macklon A.E.S. & Weatherley P.E. (1965) Controlled environment studies of the nature and origins of water deficits in plants. New Phytologist 64, Makela A., Berninger F. & Hari P. (1996) Optimal control of gas exchange during drought: theoretical analysis. Annals of Botany 77, Masle J. & Farquhar G.D. (1988) Effect of soil strength on the relation of water-use efficiency and growth to carbon isotope discrimination in wheat seedlings. Plant Physiology 86, Maximov N.A. (1929) The Plant in Relation to Water: a Study of the Physiological Basis of Drought Resistance. George Allen & Unwin Pty Ltd, London. (English translation). Meidner H. & Bannister P. (1979) Pressure and solute potentials in stomatal cells of Tradescantia virginiana. Journal of Experimental Botany 30, Meidner H. & Edwards M. (1975) Direct measurement of turgor pressure potentials of guard cells, I. Journal of Experimental Botany 26, Meinzer F.C. & Grantz D.A. (1991) Coordination of stomatal, hydraulic, and canopy boundary layer properties: Do stomata balance conductance by measuring transpiration? Physiologia Plantarum 83, Meinzer F.C., Saliendra N.Z. & Crisoto C.H. (1992) Carbon isotope discrimination and gas exchange in Coffea arabica during adjustment to different soil moisture regimes. Australian Journal of Plant Physiology 19, Monteith J.L. (1995) A reinterpretation of stomatal response to humidity. Plant, Cell and Environment 18, Moore T. (1860) The Octavo Nature-Printed British Ferns: Being Figures and Descriptions of the Species and Varieties of Ferns Found in the United Kingdom, II. Athyrium To Ophioglossum. Bradbury and Evans, London. Nonami H., Schulze E.-D. & Ziegler H. (1990) Mechanism of stomatal movement in response to air humidity, irradiance and xylem water potential. Planta 183, Rawson H.M., Begg J.E. & Woodward R.G. (1977) The effect of atmospheric humidity on photosynthesis, transpiration and water use efficiency of leaves of several plant species. Planta 134, Roberts J., Wallace J.S. & Pitman R.M. (1984) Factors affecting stomatal conductance of bracken below a forest canopy. Journal of Applied Ecology 21,