HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Wall, G. W. Articles by Idso, S. B. Search for Related Content PubMed Articles by Wall, G. W. Articles by Idso, S. B. Agricola Articles by Wall, G. W. Articles by Idso, S. B. Related Collections Carbon Sequestration Global Change Agroclimatology Crop Ecology Crop Physiology & Metabolism Wheat

Agroclimatology

Interactive Effects of Elevated Carbon Dioxide and Drought on Wheat

^a G.W. Wall, B.A. Kimball, D.J. Hunsaker, P.J. Pinter, Jr., R.L. LaMorte, and S.B. Idso (retired), USDA-ARS, U.S. Water Conservation Lab., 4331 E. Broadway Rd., Phoenix, AZ 85040
^b LI-COR, P.O. Box 4425, Lincoln, NE 68504
^c Univ. of Illinois, Dep. of Crop Science and Plant Biology, 1201 W. Gregory Dr., Urbana-Champaign, IL 61801
^d Dep. of Animal and Plant Sciences, Univ. of Sheffield, Sheffield, S10 2TN, UK
^e D.L. Hendrix (retired), USDA-ARS, Western Cotton Research Lab., 4135 E. Broadway Rd., Phoenix, AZ 85040
^f Potsdam Institute for Climate Impact Research, P.O. Box 601203, D-14412 Potsdam, Germany
^g Dep. of Soil Science, Humboldt Univ. of Berlin, Invalidenstrasse 42, 10115 Berlin, Germany
^h Lab. of Tree-Ring Research, Univ. of Arizona, Tucson, AZ, 85721

^* Corresponding author (gwall{at}uswcl.ars.ag.gov)

Received for publication March 30, 2004.

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Atmospheric CO₂ concentration (C_a) continues to rise. An imperative exists, therefore, to elucidate the interactive effects of elevated C_a and drought on plant water relations of wheat (Triticum aestivum L.). A spring wheat (cv. Yecora Rojo) crop was exposed to ambient (Control: 370 µmol mol^–1) and free-air CO₂ enrichment (FACE: ambient + 180 µmol mol^–1) under ample (Wet), and reduced (Dry), water supplies (100 and 50% replacement of evapotranspiration, respectively) over a 2-yr study. Our objective was to characterize and quantify the responses of 26 edaphic, gas exchange, water relations, carbohydrate pool dynamics, growth, and development parameters to rising C_a and drought. Increasing C_a minimized the deleterious effects of soil–water depletion by increasing drought avoidance (i.e., lower stomatal conductance and transpiration rate, and growth and development of a more robust root system) and drought tolerance (i.e., enhanced osmoregulation and adaptation of tissue) mechanisms, resulting in a 30% reduction in water stress–induced midafternoon depressions in net assimilation rate. An elevated C_a–based increase in daily and seasonal carbon gain resulted in a positive feedback between source capacity (shoots) and sink demand (roots). Devoid of a concomitant rise in global temperature resulting from the rise in C_a, improved water relations for a herbaceous, cool-season, annual, C₃ cereal monocot grass (i.e., wheat) are anticipated in a future high-CO₂ world. These findings are applicable to other graminaceous species of a similar function-type as wheat common to temperate zone grassland prairies and savannas, especially under dryland conditions.

Abbreviations: A, instantaneous leaf net assimilation rate [µmol (CO₂) m^–2 s^–1] • A_max, daily maximum leaf net assimilation rate [µmol (CO₂) m^–2 s^–1] • A_N, season-long relative mean leaf net assimilation rate normalized between 0 to1 scale (i.e., A_N = A/A_max; dimensionless) • A', daily integral of net leaf carbon accumulation [g (C) m^–2 d^–1] • A'', seasonal integral of net leaf carbon accumulation [g (C) m^–2 yr^–1] • ANOVA, analysis of variance • B_S, season-long average total shoot biomass (g m^–2 ground area) • B_S(max) maximum season-long total shoot biomass (g m^–2 ground area) • B_R, season-long average living root biomass (g m^–2 ground area) • B_R(max), maximum season-long living root biomass (g m^–2 ground area) • B_R/B_S, season-long living root to total shoot biomass ratio (dimensionless) • CD, Control-Dry • CW, Control-Wet • C, carbon dioxide effect in ANOVA • C_a, atmospheric CO₂ concentration [µmol (CO₂) mol^–1] • C_i, intercellular CO₂ concentration [µmol (CO₂) mol^–1] • C_i/C_a, ratio of C_i to C_a (dimensionless) • CHO, concentration of simple sugars (sucrose, glucose, fructose) in leaf tissue (g kg^–1) • D, soil dehydration cycle effect in ANOVA • DAE, day after 50% emergence • E, effect in the ANOVA (i.e., Y, D, T, C, I, Z) • ET, season-long average daily evapotranspiration rate (mm d^–1) • ET_c, season-long cumulative evapotranspiration (mm) • e_a, atmospheric water vapor pressure at T_a (Pa) • e_s, atmospheric saturation water vapor pressure at T_a (Pa) • e*, atmospheric water vapor pressure deficit (i.e., e* = e_s – e_a) at T_a (Pa) • F, F statistic • FACE, free-air-CO₂-enrichment • FD, FACE-Dry • FW, FACE-Wet • Fru, concentration of fructose in leaf tissue (g kg^–1) • g_s, stomatal conductance to water vapor [mol (H₂O) m^–2 s^–1] • Glu, concentration of glucose in leaf tissue (g kg^–1) • HMW, concentration of high molecular weight fructans in leaf tissue (g kg^–1) • I, irrigation effect in ANOVA • IWUE (A/g _s) intrinsic water use efficiency [µmol (CO₂) mol (H₂O)^–1] • PPFD, photosynthetic photon flux density [µmol (photons) m^–2 s^–1] • LMW, concentration of low molecular weight fructans in leaf tissue (g kg^–1) • Suc, concentration of sucrose in leaf tissue (g kg^–1) • P, probability of a greater F statistic by chance • S, concentration of starch in leaf tissue (g kg^–1) • T, time of day [mid-morning (MM: 2.5 h prior to solar noon), midday (MD: solar noon), and midafternoon (MA: 2.5 h after solar noon)] effect in ANOVA • TR, leaf transpiration rate [mmol (H₂O) m^–2 s^–1] • T_a, ambient air temperature inside cuvette (°C) • T_l, leaf temperature inside cuvette (°C) • T_l, leaf – air temperature inside cuvette (°C) • TNC, concentration of total nonstructural carbohydrates in leaf tissue (g kg^–1) • WUE, (A/TR) water use efficiency [µmol (CO₂) mmol (H₂O)^–1] • Y, year effect in ANOVA • Z, canopy height effect in ANOVA • _M, soil matric potential (MPa) • _W, total leaf water potential (MPa) • _W(PB), total leaf water potential measured with a pressure chamber (MPa) • _W(PSY), total leaf water potential measured with thermocouple psychrometers (MPa) • , leaf osmotic potential (MPa) • _P, leaf turgor potential (MPa) • , rate of osmotic adjustment (dimensionless) • _P, rate of loss of turgor (dimensionless) • _S, volumetric soil–water content (m³ m^–3)

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
THE atmospheric CO₂ concentration (C_a) is rising about 0.5% per annum from present day levels of approximately 370 µmol mol^–1 and is projected to reach 500 to 1000 µmol mol^–1 by the 21st century, depending on future emission rates (IPCC, 1996, 2001). The increase in C_a, as well as of other radiatively active gases, is predicted to cause global warming and to alter precipitation patterns (McCarthy et al., 2001). Higher levels of C_a directly affects the physiology, water relations, growth, and development of plants (Long et al., 2004; Ainsworth and Long, 2005), as do warmer temperatures and altered water supplies. In this article we focus on the effects of elevated C_a on the water relations of wheat at both ample and stress levels of soil–water supply.

Wheat is the world's foremost grain source (FAO, 1996). As such, it has been identified as a high priority for global change research (IPCC, 1996, 2001) by the Global Change and Terrestrial Ecosystems–International Geosphere–Biosphere Program (Steffen et al., 1992). Wheat, a cool-season annual monocot C₃ grass, is representative of graminaceous species common to temperate zone grassland prairies and savannas that cover a significant proportion of Earth's landmass and that make a substantial contribution to the above- and belowground global carbon balance. An imperative exists, therefore, to elucidate the effects that a future high-CO₂ world and potential changes in soil–water supply could have on edaphic, gas exchange, water relations, carbohydrate pool dynamics, growth, and development characteristics of wheat. Dissemination of such information could be of use to those individuals responsible for formulating strategies to mitigate the adverse effects of global change, while optimizing those that are beneficial (i.e., development of new cultivars with higher yields or improved grain end-use quality).

Our objective was to characterize and quantify a wheat plant's response to global change. Pursuant to this aim we elucidated the interdependency of four edaphic, nine gas exchange, four leaf tissue carbohydrate content, four water relations, and five growth parameter responses of spring wheat grown under ample and reduced soil–water content and ambient and elevated C_a levels. Specifically, we hypothesized the following:

1. An elevated C_a–based decrease in leaf stomatal conductance (g_s) will reduce depletion of volumetric soil–water content (_S), thereby causing less negative soil matric potential (_M).
   
2. Conservation of water via an elevated C_a–based reduction in g_s will enable stomata to remain open for a longer period into a drought.
   
3. During the diurnal period, when transpiration rates (TR) are at their maximum, elevated C_a will improve (less negative) midday total plant water potential (_W), and during the nocturnal period greater recovery in plant water status will be observed at predawn and sunset.
   
4. Under more severe drought conditions, elevated C_a will enhance the rate of osmotic adjustment (), thereby resulting in a decrease in the rate in loss of turgor (_P), as evidenced by leaves that maintain higher values of turgor potential (_P).
   
5. Under ample water supply any reduction in carbon gain will occur mostly because of nonstomatal limitations, whereas under drought conditions stomatal limitations will reduce carbon gain.
   
6. On a relative basis, the net gain in carbon uptake in response to elevated C_a will be proportionately greater under dry than wet.
   
7. Improved water relations that increase carbon gain will be evidenced by an increase in the concentration of total nonstructural carbohydrates (TNC) in source leaves.
   
8. Elevated C_a will foster a positive feedback relationship between aboveground source capacity (shoots) and belowground sink demand (roots), and this effect will be proportionately greater under dry than wet conditions.
   

To test the validity of these eight hypotheses, we characterized and quantified the effect of tissue dessication (during soil dehydration) and subsequent recovery (after soil rehydration) on a total of 26 parameters of open-field spring wheat grown in ambient air and air enriched with C_a, and under four soil dehydration–rehydration cycles over a 2-yr study.

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Site and Experimental Design
During the 1992–1993 and 1993–1994 growing seasons, two experiments were conducted to investigate the interactive effects of C_a and _S on spring wheat. The experiments were conducted in a 12-ha open field at the University of Arizona's Maricopa Agricultural Center, Maricopa, AZ (33.07° N lat; 111.97° W long; 361 m above sea level), located 50 km south of Phoenix, AZ. The soil was a Trix clay loam [fine-loamy, mixed (calcareous) hyperthermic Typic Torrifluvents]. Details of the field site and biological activity sampling areas (Wall and Kimball, 1993), as well as the irrigation system (Hunsaker et al., 1996), have been described previously. Except for the use of drip rather than flood irrigation, all agronomic practices were in accordance with local cultural practices.

Carbon Dioxide and Irrigation Treatment Description
The FACE approach was employed to create an elevated C_a treatment (C) (Hendrey, 1993). The FACE plots were fumigated with CO₂ 24 h d^–1 at rates required to maintain a concentration of approximately 550 µmol mol^–1 at the center point of the plot from 50% emergence until physiological maturity (1 Jan.–16 May 1993, and 28 Dec. 1993–31 May 1994). Control plots contained a plenum and vertical vent pipes, but had no airflow and were at ambient C_a levels of approximately 370 µmol mol^–1. Seasonal average values of C_a were within 0.5 µmol mol^–1 of the set point and 93% of the 1-min averages were within 10% of the set point (Hendrey, 1993) (i.e., 548 µmol mol^–1 during the day and 598 µmol mol^–1 at night for FACE and 363 µmol mol^–1 during the day and 412 µmol mol^–1 at night for Control).

Irrigation water was supplied with a subsurface drip-tape irrigation system (0.5-m tube spacing, 0.3-m emitter spacing, 0.2-m depth) as described by Hunsaker et al. (1996). Each of the main circular C plots were split into semicircular halves, with each half receiving an irrigation (I) treatment of either 100% (Wet, well-watered) or limited 50% (Dry, water-stressed) replacement of potential evapotranspiration (ET). A water balance procedure (Fox et al., 1992) was used to determine both the timing and amount of irrigated water applied to the Wet plots. An irrigation event occurred when the soil–water depletion of plant-available water in the crop root zone for Wet plots was between 30 and 40% of available. The amount of irrigated water applied to the Wet plots was then determined to replenish the root zone to field capacity (0% depletion). Because the Dry plots received only 50% as much irrigation water as the Wet plots, the depletion of water for the Dry plots was between 60 and 80% available. A preexisting crop coefficient curve developed for wheat was used to estimated crop water use that expressed the ratio of daily wheat ET to daily potential ET for a grass-reference crop. The grass-reference ET was calculated with the modified Penman equation as implemented by Doorenbos and Pruitt (1977). All meteorological data required for the modified Penman calculation was obtained from a weather station about 2 km from the field site.

For ease in applying the I treatments, the same irrigation regime was applied to the same split-plot within a C level for each replication: strip-split-plot experimental design. These C and I levels gave four treatment combinations: Control-Dry (CD); FACE-Dry (FD); Control-Wet (CW); FACE-Wet (FW). All treatments were replicated four times for a total of eight rings—four of them were active FACE rings with blowers and four of them were inactive Control rings without blowers (Pinter et al., 2000).

Crop Culture
Certified wheat seeds (cv. Yecora Rojo) were sown (John Deere Flex-Planter 71, Moline, IL) into flat beds at 0.25-m row spacings on 15 Dec. 1992 at a plant population of 130 plants m^–2; 50% emergence occurred on 1 Jan. 1993, and the crop was harvested 25–27 May 1993. During the second year seeds were sown at a higher plant population of 180 plants m^–2 on 7–8 Dec. 1993; 50% emergence occurred on 28 Dec. 1993, and the crop was harvested on 1 June 1994. To germinate the seeds a drip irrigation system was used to apply 317 mm of water from 9 to 13 d after they were sown during 1993, whereas a portable sprinkler system was installed temporarily to apply a total of 30 mm of water from 6 to10 d after seeds were sown during 1994 (Fig. 1b , d). Because day after 50% emergence (DAE) corresponded with the first day of the year during this 2-yr study, hereafter DAE is equivalent to day of year.

View larger version (57K): [in this window] [in a new window]   Fig. 1. Soil matric potential (M) for a 0.9-m soil profile of a Trix clay loam vs. day after 50% emergence (DAE) (i.e., calendar date same as DAE) for the (a) 1993 and (c) 1994 growing seasons. Symbols in legend for panels (a) and (c) refer to Control-Dry (CD), FACE-Dry (FD), Control-Wet, and FACE-Wet (FW) treatments (Control at 370 µmol CO2 mol–1 and FACE at 550 µmol CO2 mol–1; Dry at 50% and Wet at 100% replacement of evapotranspiration). Vertical dashed lines in panels (a) and (c) separate four individual soil dehydration cycles by development stages as follows: (1) seedling development, tillering, and stem elongation (DAE 1–71 during 1993; DAE 1–76 during 1994); (2) inflorescence emergence (DAE 72–85 during 1993; DAE 77–88 during 1994); (3) anthesis (DAE 86–103 during 1993; DAE 89–111 during 1994); and (4) grain filling through physiological maturity (DAE 104–136 during 1993; DAE 112–137 during 1994). Symbols in legend given in panel (b) refer to drip-tape irrigation for Dry and Wet treatments, rainfall amounts, and N fertilizer application dates during (b) 1993 and (d) 1994. Amount of fertilizer (kg N ha–1) applied is given in parenthesis above irrigation histograms at tillering, inflorescence emergence, and anthesis for (b) 1993 and (d) 1994. Cumulative distribution curves (Cum.) denoting the amount of water (drip-tape irrigation + rainfall) given by right-most y intercept for Dry and Wet treatments in panels (b) and (d). Onset of the differential irrigation denoted by separation in Cum. curves on 1 Mar. 1993 (DAE 60) in panel (b) and 20 Jan. 1994 (DAE 20) in panel (d). Up-pointing arrows in panels (b) and (d) denote DAE when gas exchange, water relation, and leaf tissue carbohydrate concentration measurements were measured, whereas parenthesis below up-pointing arrows denote number of days between measurement dates and the last irrigated for (Dry:Wet) treatments. Letters e through j in legend of panel (c) denote DAE and number of days between measurement dates and last irrigated for (Dry:Wet) treatments during the third (e, f) and fourth (g, h, i, and j) soil dehydration cycles when standard thermocouple psychometric techniques were employed to measure leaf total water and osmotic potentials during 1994. Given on the right-most y axis are the sources of variance in ANOVA, which include carbon dioxide (C) for main plot, irrigation (I) for split-plot, and C x I interaction effects in a strip-split-plot experimental design. Significance effects given within soil dehydration cycle as *, **, ***, and ns for P 0.05, P 0.01, P 0.001, and not significant, respectively. (ne for effect not estimated); actual probability of a greater F statistic by chance reported if P 0.10 and P 0.25 P 0.10. Each mean datum was derived from four replications (i.e., means based on n = 4, SE of replication means based on n = 4).

Fertilizer amounts were applied so that nutrients were nonlimiting. A preplant application of granular (16–20–0) fertilizer supplied 54 kg N ha^–1 and 29 kg P ha^–1 in both years. Nitrogen, as urea–ammonium nitrate solution (0.32 kg N kg fertilizer^–1, Uran-32), and P, as super-phosphoric acid, fertilizer were applied through the irrigation tubes at tillering, inflorescence emergence, and anthesis (Fig. 1b, d). Total amounts of N applied were 277 and 261 kg N ha^–1, and of P were 44 and 29 kg P ha^–1 during 1993 (Fig. 1b) and 1994 (Fig. 1d), respectively.

Volumetric Soil–Water Content and Evapotranspiration Rate
Measurements of _S were obtained in each subplot using time domain reflectometry (TDR) and neutron scattering equipment (Hydroprobe Model 503 DR, Campbell Pacific Co., Martinez, CA). Access tubes (50 mm diam. and 2.0 m long) were installed in the no-traffic section (Wall and Kimball, 1993) of each of 16 plots, before seeds were sown during 1992. The TDR measurements were taken over the 0- to 0.3-m soil depth, whereas neutron scattering measurements were taken from the 0.4- to 2.0-m depth in 0.2-m increments (Hunsaker et al., 1996). The soil-water retention characteristics for the site were estimated from data obtained from Post et al. (1988). These data were used to develop individual moisture release curves for Control and FACE plots, which were used to derive _M over a 0.9-m soil profile for each date _S was measured (Fig. 1a, c).

After expected root penetration had occurred to a given depth (Robertson et al., 1993a, 1993b), a stair-step model incorporated values of _S at consecutively deeper depths in 0.3-m intervals. By assuming negligible deep percolation, average daily ET was essentially set equal to the temporal changes in _S for the active root zone. Whenever irrigation or rain occurred, however, it was estimated by interpolation by using the average ET before and after the wetting event for the active root zone. Hence, _S and the total amount of irrigation plus rain were used to calculate ET, and therefore season-long cumulative evapotranspiration (ET_C), based on the soil–water balance method (Jensen et al., 1990) as implemented by Hunsaker et al. (1996).

Measurements of Leaf Stomatal Conductance and Net Assimilation and Transpiration Rates
Gas exchange rates were measured on randomly selected leaves with three portable closed-exchange (transient) systems with a 0.25-L transparent Plexiglas cuvette (Model LI-6200, LI-COR, Lincoln, NE). Each infrared gas analyzer was calibrated against a gravimetrically prepared mixture of CO₂ in air (±1% Primary Standard, Matheson Gas Products, Cucamonga, CA), and cuvette humidity sensors were calibrated with a dew-point generator (LI-610, LI-COR, Lincoln, NE) immediately before use. Before each measurement, the cuvette was allowed to come into equilibrium with prevailing air temperature and relative humidity. As described previously by Wall et al. (2000), dawn to dusk measurements of g_s, and net assimilation (A) and transpiration (TR) rates were made on the central portion of fully expanded (ligule emerged) upper-canopy sunlit leaves, begun at a cuvette CO₂ concentration of 370 ± 35 or 550 ± 35 µmol mol^–1 for Control and FACE, respectively. One leaf per row from each of five different rows (repeated measures) was randomly selected for measurement. The leaf cuvette was held in the horizontal position and caution was used not to shade any portion of the leaf. Because three separate portable closed-exchange systems were used, three replications could be measured simultaneously within 60 min at 2-h intervals, thereby minimizing variation in gas exchange measurements due to diurnal changes in meteorological conditions, particularly incident photon flux density (PPFD: µmol photons m^–2 s^–1), air temperature (T_a), and water vapor pressure deficit (e*) (i.e., e* = e_s – e_a). Three observations were recorded, but the first 10-s measurement interval was discarded from the statistical analysis because C_a in the assimilation chamber was unstable during that period.

Each gas exchange system made direct measurements of C_a, atmospheric water vapor pressure (e_a), leaf temperature (T_l), T_a, and incident PPFD normal to the leaf surface. We calculated leaf A, g_s, TR, saturation water vapor pressure (e_s), e*, internal CO₂ concentration (C_i) and the ratio of C_i to C_a (C_i/C_a) as suggested by LI-COR (1990) following equations of von Caemmerer and Farquhar (1981). A value for water use efficiency (WUE) and intrinsic water use efficiency (IWUE) was derived as the ratio of A to TR and A to g_s, respectively,whereas the difference between T_l and T_a was used to derive the leaf minus air temperature (T_l). Phenological development (average numerical decimal code across all treatments) was used to group values of the 26 response parameters by distinctive development stages (Zadoks et al., 1974) as follows; inflorescence emergence until hard dough (DAE 75, 89, 99, 105, and 118; Zadoks 50–88) during 1993; tillering until soft dough (DAE 41, 69, 81, 95, and 123; Zadoks 24–84) during 1994. To investigate the effect of the lack of blowers in the Control plots (Pinter et al., 2000), during 1994 additional midday measurements of gas exchange parameters (i.e., g_s, A, TR, T_l, T, C_i, C_i/C_a) were taken on DAE 116, 119, 126, 131, 133, and 137 from grain filling until physiological maturity (Zadoks 84.1–87.8).

The daily accumulation of carbon (A') was derived by integrating the area under each dawn to dusk net carbon uptake curve. Dawn and dusk times in Mountain Standard Time (MST) were obtained from standard astronomical tables for Maricopa, AZ, and served as a zero response point for the purpose of integration. The seasonal accumulation of net carbon uptake (A'') was derived by integrating the area under each daily carbon uptake curve from 50% emergence until canopy senescence reduced fractional absorption of PPFD to 25% using a trapezoidal integration routine (Area.xfm, Sigma Plot, v. 4.01, SPSS, Chicago, IL).

Analysis of Leaf Total Nonstructural Carbohydrate Concentration
Ten uppermost fully expanded (i.e., ligule emerged) sunlit leaves were randomly selected every 2 h (as many as eight sample intervals) from dawn to dusk. Sampling dates corresponded with the development of leaf number five through eight and the flag leaf. At anthesis on DAE 81 during 1994, after the canopy had reached its maximum height, a whole-canopy leaf profile was sampled from the flag (top of the canopy) through the flag-4 leaf (lower photosynthetically active leaf at the bottom of the canopy) on an individual culm for each of 10 culms. All leaf samples for each treatment were pooled and kept on dry ice for up to 10 h until they were transported to the laboratory. Leaves were kept overnight on dry ice in a cold storage facility maintained at 10°C. The following day the leaf samples were freeze-dried, ground through a 0.4-mm mesh in a Wiley Mill, and quickly transferred to tightly stoppered glass vials that were stored at room temperature.

The TNC, comprised of simple sugars such as glucose (Glu), fructose (Fru), and sucrose (Suc), and complex carbohydrates such as low (LMW) and high (HMW) molecular weight fructans and starch (S), were determined using a microplate assay methodology as described elsewhere (Hendrix and Peelen, 1987; Hendrix, 1993).

Measurements of Total Leaf Water Potential with a Pressure Chamber
Blades of uppermost fully expanded sunlit leaves were excised approximately 5 mm apical to the leaf collar and sealed in a plastic bag containing a damp paper towel. Plastic bags containing excised leaves were then stored in an insulated container containing ice. No direct contact occurred between the ice and leaf samples. Values of _W were measured with a pressure chamber (bomb) (_W(PB)) (Scholander et al., 1965) (Model 3000, Soil Moisture Equipment Corp, Santa Barbara, CA). This protocol has given reliable results as long as particular precautions are taken to minimize error due to rapid water loss (Turner and Long, 1990).

Measurements of Total Leaf Water and Osmotic Potentials with Thermocouple Psychrometers
Measurements of thermocouple psychrometers-based total leaf water (_W(PSY)) and osmotic potentials were made using standard psychrometric techniques (Brown and Collins, 1980; Walker et al., 1983; Briscoe, 1984). A leaf sample approximately 50 mm long was rapidly excised from the central section of the uppermost fully expanded sunlit leaf (excluding the midrib), rolled, and inserted into the sample chamber and sealed within 15 s. Surgical gloves were worn to minimize contamination of the sample. Psychrometer chambers containing an excised leaf were placed in an insulated isothermal cooler at the field site and transported to a temperature-controlled room maintained at 25°C within 2 to 4 h. The _W(PSY) was measured when thermal and vapor pressure equilibrium had been established (within 4 h). The sample chambers were then frozen in liquid N₂ for 3 min and thawed. Measurements of were obtained after re-establishment of thermal and vapor pressure equilibrium (within 2 h). Measurements of _W(PSY) and were taken at midday from anthesis through grain filling, up to soft dough, which corresponds with the third (e:96 and f:106) and fourth (g:111, h:113, i:120, and j:124) soil dehydration cycles during 1994 (letters in circles correspond with DAE illustrated for FD treatment in Fig. 1c). Both midday and midafternoon psychrometric measurements were taken on DAE 124. Values of turgor potential (_P) were derived from _W(PSY) and (i.e., _P = _W(PSY) – ; Kirkham, 1990). No attempt was made to correct the psychrometric obtained measurements of for apoplastic dilution of the symplast (Campbell et al., 1979) or starch hydrolysis (Bennett et al., 1986).

Two types of leaf in situ thermocouple psychrometers were used in combination with their associated stainless steel sampling chambers; 48 of them were Model 84-3vC psychrometers (J.R.D. Merrill Specialty Equipment, Logan, UT), and 36 of them were Model C-30 psychrometers (Wescor, Logan, UT).

Root and Shoot Growth
The seasonal average (B_R) and maximum (B_R(max)) living root biomass (excluding crown tissue) in a 1-m soil profile was obtained from two in-row and one inter-row root cores (86 mm, i.d.) using a gas-driven soil core device (Eijkelkamp Agrisearch Equipment, Cobra Model 248, Sweden). Cores were extracted with a portable lever system to minimize mechanical damage to adjacent sampling areas. Following extraction, all individual cores were placed in a –14°C freezer within 4 h of sampling. Root and organic debris material from each core was elutriated (Smucker et al., 1982) from the soil with a modified hydropneumatic elutriation system (Wall, 2000). Live roots were separated from organic debris material manually (intact, white-colored roots). The total shoot biomass (B_S) above each root core (i.e., aboveground green and brown leaves, sheaths, culms, and ears and belowground crown tissue) was also collected. From these data the seasonal maximum shoot biomass (B_S(max)) was obtained. Both B_R and B_S material were oven-dried at 68°C for 48 h, desiccator cooled, and weighed. Lastly, the root/shoot ratio (B_R/B_S) was derived from values of B_R and B_S.

Statistical Analysis
Data were analyzed as a strip-split-plot design using the SAS Mixed Procedure (Littell et al., 1996) for the ANOVAs with C as the main plot and I as the strip-split plot (Wall and Kimball, 1993). A third factor in the ANOVA was soil dehydration cycles (D). As many as nine and six soil dehydration cycles occurred in the Dry plots during 1993 and 1994, respectively (Fig. 1a, c), whereas the Wet plots remained hydrated throughout the experiment. Nevertheless, to synchronize development stages and soil dehydration cycles between years, both Dry and Wet plots were reclassified into two separate soil dehydration cycles during vegetative and reproductive development, respectively. The D effect was treated as a repeated measure specifying first order, autoregression correlation for the covariance structure. Time of day (T) was treated as a fourth factor in the ANOVA [MM: midmorning (2.5 h before solar noon); MD: midday (solar noon); MA: midafternoon (2.5 h after solar noon)]. A fifth factor in the ANOVA was year (Y). To determine the effect of leaf position (Z) on TNC levels, a three-way ANOVA (i.e., Z, C, I) was performed.

To avoid pseudo-replication, all ANOVAs were performed on replication means. Effects (E) (i.e., Y, D, T, C, I, Z) declared "significant" in the text have P values 0.05 unless stated otherwise. Tests for homogeneity of variance and subsequent data transformation were performed whenever appropriate before any ANOVA (Box et al., 1978). A covariant analysis (_W(PSY) as the covariant) was used to determine the C effect on and _P.

	RESULTS AND DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Atmospheric Conditions
Despite a slight difference in rainfall events between years (Fig. 1b, d), climatic conditions (i.e., T_a, PPFD, and e*) in this semiarid desert region were similar between 1993 and 1994. Briefly, all measurement days had predominantly clear skies (Kimball et al., 1995; Garcia et al., 1998). During both growing seasons, maximum hourly and total solar radiation increased from 2.2 to 3.6 MJ m^–2 h^–1 and 14 to 28 MJ m^–2 d^–1, respectively, from January through May. Consequently, a corresponding increase in maximum T_a from 18 to 34°C occurred. An increase in the evaporative demand imposed on the crop was evident from the increase in midday e*, which ranged from 1.0 kPa at tillering to just under 5 kPa at hard dough for both years. Although peak PPFD, in excess of 2000 µmol (photons) m^–2 s^–1, occurred at solar noon, diurnal peaks in T_a and e* occurred at mid- to late afternoon, a period when any water stress symptoms are usually at their peak. Nevertheless, temporal trends in accumulated thermal time and ontogeny of the crop were similar between years (Pinter et al., 1996).

Analysis of Variance
Overall, C, I, and C x I interactive effects were consistent across Y, D, and T effects (Table 1). The I effect predominated over the C effect. Significance C x I interaction effects increased in both frequency and level of significance (lower P value) as _S became depleted. This occurred mostly because of a proportionately greater disparity in the C effect under Dry compared with that under Wet. Significant Y effects occurred primarily because the low-volume, high-frequency irrigation regime caused only mild water stress during 1993, whereas the high-volume, low-frequency irrigation regime caused more severe water stress during 1994 (Fig. 1a, c). The D effect was insignificant during vegetative development (first and second soil dehydration cycles), but both the frequency and level of significance of the D effect increased during reproductive development (third and fourth soil dehydration cycles) (Fig. 1a, c). Significant T effects occurred primarily because of a greater responsiveness of parameters at MA (a period when drought stress is usually at its diurnal peak) compared with either MM or MD.

View larger version (46K): [in this window] [in a new window]   Fig. 2. Dawn to dusk trends in mean stomatal conductance to water vapor (gs) of fully expanded sunlit spring wheat leaves for day after 50% emergence (DAE) and development stages given for 5 d during the (a–e) 1993 and (f–j) 1994 growing seasons. Typical values of photosynthetic photon flux density, air temperature, and vapor pressure deficit for this semiarid desert region, for DAE 75–118 during (a–e) 1993, have been reported elsewhere (Garcia et al., 1998). Symbols in legend and sources of variance and results from ANOVA same as described in Fig. 1. Each mean datum was derived from five leaves across two observations (repeated measures), made within a 60- to 90-min period (U.S. Mountain Standard Time), across two replications for Dry or four replications for Wet during 1993 and three replications for both Dry and Wet during 1994 (i.e., means based on n = 20 for Dry and n = 40 for Wet during 1993, and means based on n = 30 for both Dry and Wet during 1994). Vertical bars are 1 SE of replication means (i.e., based on n = 2 for Dry and n = 4 for Wet during 1993 and n = 3 for both Dry and Wet during 1994). The above illustration was derived from measurements of as many as 3540 leaves.

View larger version (47K): [in this window] [in a new window]   Fig. 7. Mean maximum midday (MD: solar noon) net assimilation rate (A) (a, b) and stomatal conductance to water vapor (gs) (g, h) for fully expanded sunlit spring wheat leaves for day after 50% emergence (DAE) and development stages given during the (a, g) 1993 and (b, h) 1994 growing seasons, respectively. Histograms shown during 1994 in panel b (inserts c, d, e, and f) illustrate values of A from inflorescence emergence (DAE 81) until physiological maturity for (c, d) Wet (DAE 137) and (e, f) Dry (DAE 123) treatments, a period of accelerated senescence. Treatment inversion in the carbon dioxide (C) effect on values of A are denoted by up-pointing arrows in panel (b), which intersect approximately on DAE 124 (histogram panels c and d) for Wet, and on DAE 103 (histogram panels e and f) for Dry treatments. Histograms shown during 1994 in panel (h) (inserts i and j) illustrate values of gs for (i) Wet and (j) Dry treatments. Unlike the treatment inversion observed in the C effect for A (inserts c, d, e, f), no such effect was observed in gs (i, j). All measurements were simultaneous with, and sampled as described for gs in Fig. 2 and for A in Fig. 6. Symbols in legend, source of variance, and results from ANOVA same as described in Fig. 1. The above illustration was derived from measurements of as many as 960 leaves.

View larger version (45K): [in this window] [in a new window]   Fig. 3. Predawn to sunset trends in mean total leaf water potential measured with a pressure chamber (W(PB)) of expanded sunlit spring wheat leaves for 50% day after emergence (DAE) and development stages given for 5 d during the (a–e) 1993 and (f–j) 1994 growing seasons. Symbols in legend, sources of variance, and results from ANOVA same as described in Fig. 1. Each mean datum was derived from three to four subsamples for four replications (i.e., means based on n = 12 or n = 16). Vertical bars are 1 SE of replication means (i.e., n = 4). The above illustration was derived from measurements from as many as 3392 leaves.

View larger version (39K): [in this window] [in a new window]   Fig. 4. Dawn to dusk trends in mean total leaf water potential measured with a pressure chamber (W(PB)) in fully expanded sunlit spring wheat leaves at (a, b) predawn, (c, d) midday (solar noon), and (e, f) sunset for day after 50% emergence (DAE) 43 to 118 during the (a, c, e) 1993 and DAE 41 to123 during the (b, d, f) 1994 growing seasons. Measurement dates correspond with up-pointing arrows shown in Fig. 1b, d. Each mean datum is the pooled average of three to four leaves collected within a 15-min period in four replicate blocks (i.e., means based on n = 12 or n = 16). Vertical bars are 1 SE of replication means (i.e., n = 4). Symbols in legend, source of variance, and results from ANOVA are the same as described in Fig. 1. Note that the y axis scale for (a, b) predawn and (e, f) sunset are the same, but the scale for (c, d) midday is double that of predawn and sunset. The above illustration was derived from measurements of as many as 2304 leaves.

View larger version (52K): [in this window] [in a new window]   Fig. 5. Mean (b) total leaf water (W(PSY)), and (b) osmotic () potentials determined using standard thermocouple psychometric techniques, and the (a) resultant pressure (P) potential (i.e., P = W(PSY) – ) for fully expanded sunlit flag leaves of spring wheat at midday (MD) (solar noon) and at midafternoon (MA) (2.5 h after solar noon) on day after 50% emergence (DAE) and development stage given during the 1994 growing season. Letters e through j in legend of panel (b) denote DAE and number of days between measurement dates and last irrigated for (Dry:Wet) treatments during the third (e, f) and fourth (g, h, i, and j) soil dehydration cycles (Fig. 1c). Vertical dashed lines in panels (a) and (b) delineate the third and fourth soil dehydration cycles illustrated in Fig. 1c. Horizontal arrows in panel (a) denote overall trend in values of P across treatments. Each mean datum for W(PSY), , and the resultant P was derived from three repeated measures across four treatments and three replications (3 x 4 x 3) when Model C-30s psychrometers were used at MD on DAE 120 and MA on DAE 124, and four repeated measurements across four treatments and three replications (4 x 4 x 3) when Model 84-3vC J.D.J. Merrill Specialty Equipment, Longan, UT, were used at MD (DAE 96, 106, 111, 113, 124) (i.e., means based on n = 36 or n = 48, respectively). Vertical bars are 1 SE of replication means (i.e., n = 3). Symbols in legend, source of variance, and results from ANOVA same as described in Fig. 1. Regression of on W(PSY) (i.e., rate of change in osmotic potential; ) from anthesis until soft dough for fully expanded sunlit flag leaves of (c) spring wheat. Summary statistics for a linear regression are as follows: N, number of observations; R2, coefficient of determination; Sy/x, SE of y given a value of x. The above illustration was derived from measurements from as many as 312 leaves.

View larger version (45K): [in this window] [in a new window]   Fig. 6. Dawn to dusk trends in mean leaf net assimilation rate (A) of fully expanded sunlit spring wheat leaves for day after 50% emergence (DAE) and development stages given for 5 d during the (a–e) 1993 and (f–j) 1994 growing seasons. Measurements were simultaneous with and sampled as described for gs in Fig. 2. Symbols in legend, source of variance, and results from ANOVA same as described in Fig. 1. The above illustration was derived from measurements of as many as 3540 leaves.

View larger version (42K): [in this window] [in a new window]   Fig. 8. Dawn to dusk trends in mean leaf total nonstructural carbohydrates (TNC), consisting of water-soluble sugars (glucose, fructose, sucrose), low and high molecular weight fructans, and starch. Uppermost fully expanded sunlit spring wheat leaves were sampled for day after 50% emergence (DAE) and development stages given for 6 d during the 1993 (a–f) and 5 d during the 1994 growing seasons (g–k) (note that during 1993 panels (e) and (f) are denoted by the right-most y axis, which is scaled half that of the left-most y axis). Symbols in legend, source of variance, and results from ANOVA are the same as described in Fig. 1. Each mean datum was derived from four individual assays (repeated measures) performed on a 30-mg sample (i.e., pooled sample taken from the thoroughly mixed powder of 10 freeze-dried leaves) across four replications at 2-h sample intervals (i.e., means based on n = 16). Vertical bars are 1 SE of replication means (i.e., n = 4). The above illustration was derived from measurements of as many as 9120 leaves.

View larger version (58K): [in this window] [in a new window]   Fig. 9. Midday (MD: solar noon) trends in leaf tissue total nonstructural carbohydrate concentration (TNC) (i.e., water-soluble sugars (glucose, fructose, sucrose), low and high molecular weight fructans, and starch). (a) At anthesis (day after emergence, DAE) 89 during 1994, the uppermost fully expanded sunlit flag leaf and all physiologically active leaves beneath the flag leaf (i.e., flag-1, flag-2, flag-3, and flag-4) were sampled in sequential order from the same culm, for each of 10 culms, from the top of the canopy height (Z) denoted by the left-most y axis. (b) The whole-canopy TNC profile represents the sum of the flag through flag-4 leaves denoted by right-most y axis. Each mean datum was derived from four individual assays (repeated measures) performed on a 30-mg sample (i.e., pooled sample taken from the thoroughly mixed powder of 10 freeze-dried leaves) across four replications sampled at MD (i.e., means based on n = 16). Symbols in legend and results from ANOVA same as described in Fig. 1. Vertical bars are 1 SE of replication means (i.e., n = 4). Percentages in parentheses above each mean histogram datum indicate relative CO2 enhancement at Dry and Wet, respectively. Source of variance in ANOVA are as follows: replication (R); CO2 (C); irrigation (I); and, canopy height (Z). The corresponding error term for each effect is given in parenthesis as follows: C (R x C); I (R x I); C x I (R x C x I); Z (R x Z); Z x C (R x Z x C); Z x I (R x Z x I); Z x C x I (R x Z x C x I). The above illustration was derived from measurements of as many as 800 leaves.

View larger version (39K): [in this window] [in a new window]   Fig. 13. Seasonal trend in the (a, b) total root biomass (BR) in a 1-m soil profile and (c, d) shoot biomass (BS), and (e, f) the root to shoot ratio (BR/BS) for spring wheat during the (a, c, e) 1993 and (b, d, f) 1994 growing seasons. Sampling occurred on day after emergence (DAE) 16 and DAE 3 at the 3-leaf, DAE 36 and DAE 26 at tillering, DAE 63 and DAE 40 at stem-elongation, DAE 92 and DAE 102 at anthesis, DAE 113 and 118 at soft dough, and DAE 159 and DAE 154 at hard dough developmental stages during the 1993 and 1994 growing seasons, respectively. Each mean datum for BR was derived from four replications during vegetative development (i.e., 3-leaf, tillering, stem-elongation) and two replications during reproductive development (i.e., anthesis, soft dough, hard dough) (i.e., means and SE of replication means based on n = 4 and n = 2, respectively). During 1994, the BR was collected for four replications during vegetative and three replications during reproductive development (i.e., means and SE of replication means based on n = 4 and n = 3, respectively). The total BS (i.e., aboveground green and brown leaves, sheaths, culms and ears, and belowground crown tissue) was collected for four replications both seasons (i.e., means based on n = 4, SE of replication means based on n = 4). Symbols in legend, source of variance, and results from ANOVA same as described in Fig. 1.

Hypothesis 1: An elevated C_a–based decrease in g_s will reduce depletion of _S, thereby causing less negative _M

Since _S was used to derive _M, patterns in Y, D, C, and I and their interaction effects were similar (Table 1). For well-watered wheat (Wet), _M remained relatively constant at field capacity (i.e., _M –0.03 MPa) throughout the season (Fig. 1a, c). Notwithstanding, because of a slight reduction in whole-canopy ET (Table 2), presumably because of an elevated C_a–induced reduction in g_s (Fig. 2; Table 3), a slight reduction in soil–water depletion was observed for well-watered wheat grown in elevated C_a. The mitigating effect of elevated C_a in conserving _S via a reduction in g_s, however, can be counterbalanced by the relatively greater stimulation of elevated C_a in increasing root/shoot biomass (Fig. 13) and subsequent leaf area (Kimball et al., 1995; Samarakoon and Gifford, 1995; Pinter et al., 1996). Under ample water supply, therefore, elevated C_a had only a nominal effect on soil–water depletion. Nevertheless, _M was not sensitive enough to detect these minor differences in soil–water depletion, whereas the soil–water (Hunsaker et al., 1996) and energy (Kimball et al., 1995) balance methods were. Therefore, under ample water supply we accept Hypothesis 1 for _S, but reject it for _M.

View this table: [in this window] [in a new window]   Table 2. Means ± standard error (SE) across both years and all soil dehydration cycles for field-grown spring wheat [Triticum aestivum (L.) cv. Yecora Rojo] reared under two atmospheric carbon dioxide concentrations [Control (C) at 370 µmol mol–1, FACE (F) at ambient +180 µmol mol–1], and two irrigation regimes [Dry (D) at 50%, Wet (W) at 100% replacement of evapotranspiration] during the 1993 and 1994 growing seasons. Season-long average given for edaphic parameters include volumetric soil–water content (S), soil matric potential (M), evapotranspiration rate (ET), and cumulative evapotranspiration (ETc) obtained from Hunsaker et al. (1996). Growth parameters include season-long average living root (BR) and total shoot (BS) biomass, root to shoot ratio (BR/BS), and the season-long maximum living root (BR(max)) and total shoot (BS(max)) biomass. Also shown are the ratios of the carbon dioxide (F/C) and irrigation (D/W) treatments.

View this table: [in this window] [in a new window]   Table 3. Means ± standard error (SE) across both years and all soil dehydration cycles for field-grown spring wheat [Triticum aestivum (L.) cv. Yecora Rojo] reared under two carbon dioxide [Control (C) at 370 µmol mol–1, FACE (F) at ambient + 180 µmol mol–1], and two irrigation regimes [Dry (D) at 50%, Wet (W) at 100% replacement of evapotranspiration] during the 1993 and 1994 growing seasons. Gas exchange parameters for fully expanded upper canopy sunlit leaves measured at midmorning (MM: 2 h before solar noon), midday (MD: solar noon) and midafternoon (MA: 2 h after solar noon) include stomatal conductance to water vapor (gs), net assimilation rate (A), transpiration rate (TR), leaf temperature (Tl), leaf minus air temperature (T), both actual (WUE, A/TR) and intrinsic (IWUE, A/gs) water use efficiencies, internal CO2 concentration in mesophyll tissue (Ci) and the ratio of Ci to atmospheric CO2 concentration (Ca) (Ci/Ca). Water relation parameters include total plant water potential measured with a pressure chamber (bomb) (W(PB)), and total (W(PSY)) and osmotic () potentials measured with thermocouple psychrometers to derive turgor potential (i.e., P = W(PSY) – ). Carbohydrate parameters include leaf tissue concentrations of simple sugars (CHO: glucose, fructose, sucrose), complex carbohydrates (FRU: low and high molecular weight fructans; S: starch), and the total nonstructural carbohydrate (TNC). Also shown are the ratios of the carbon dioxide (F/C) and irrigation (D/W) treatments.

Water Vapor Exchange Rates
Hypothesis 2: Conservation of water via an elevated C_a–based reduction in g_s will enable stomata to remain open for a longer period into a drought
The overall treatment response ranking for g_s was CW > FW CD > FD (Fig. 2). Across most of the sunlit period an approximately 180 µmol mol^–1 increase in C_a reduced g_s for well-watered wheat by 33%, regardless of time of day (Fig. 2; Tables 1, 3). Furthermore, any differences in g_s resulted from a decrease in stomatal aperture, rather than stomatal density (Estiarte et al., 1994). Morison (1993, 1998) reported that a C_a as high as twice (i.e., 700 µmol CO₂ mol^–1) that in the present atmosphere generally resulted in a reduction in g_s of well-watered wheat at midday by about 40% (±5%). The elevated CO₂–based reduction in g_s across the sunlit period for well-watered wheat reported herein was consistent with a 20% lower TR measured in a leaf cuvette (Table 3), and with companion studies that showed a 7 to 23% lower TR obtained from stem flow gauges (Senock et al., 1996), an 8% lower latent heat flux derived from energy balance measurements (Kimball et al., 1995, 1999), a 5% lower rate of _S extraction derived through a soil–water balance technique (Hunsaker et al., 1996), lower apparent root hydraulic conductivity (Wall, unpublished data, 1994), a 1.5°C increase in individual leaf temperature measured in a cuvette (Table 3), a 0.6°C increase in daytime canopy surface temperatures (Kimball et al., 1995), and improved (less negative) _W(PB) of 0.04 MPa (3%) (Fig. 3–5; Tables 1, 3).

Under water stress, g_s was reduced by as much as 70% (Fig. 2; Tables 1, 3). Furthermore, under drought conditions, the effect of elevated C_a on g_s was dependent on T (Fig. 2; Tables 1, 3) because g_s was reduced by an insignificant 2% at MM and MD, but it was reduced by a significant 25% at MA (Fig. 2). This elevated C_a–based reduction in g_s across the sunlit period for water-stressed wheat lowered TR measured in a leaf cuvette by 29% (Table 3) and improved (less negative) _W(PB) by 0.16 MPa (9%) (Fig. 3–5; Tables 1, 3). Significant C x I interaction effects occurred because of a greater disparity in TR between FD and CD than between FW and CW (Tables 1, 3). Under drought conditions and ambient C_a levels, the more negative values of _W(PB) induced a reduction in stomatal aperture. But, under elevated C_a, less negative values of _W(PB) enabled less of a proportionate midafternoon depression in g_s by as much as 25% (Fig. 2). Therefore, we accept Hypothesis 2.

Plant Water Relations
Hypothesis 3: During the diurnal period, when transpiration rates are at their maximum, elevated C_a will improve (less negative) midday _W, and during the nocturnal period greater recovery in plant water status will be observed at predawn and sunset
More negative _M usually cause more negative _W and , and lower _P in field-grown wheat in both irrigated and dryland soil moisture regimes (Denmead and Millar, 1976; Millar and Denmead, 1976). Our results were consistent with this trend because as _S became depleted throughout the ontogeny of the crop, _M became more negative (Fig. 1a, c). And as _M became more negative, season-long values of _W(PB) also became progressively more negative (Fig. 3 and 4). Furthermore,_W(PB) was significantly more negative in Dry than Wet and for Control compared with FACE (Fig. 3; Tables 1, 3), thereby resulting in a treatment response ranking (i.e., less negative) for _W(PB) of FW > CW > FD > CD (Fig. 3). During 1993, the first significant C x I interaction effect for _W(PB) was observed at MA during inflorescence emergence on DAE 75 (Fig. 3a), but during 1994 it occurred one soil dehydration cycle earlier at stem-elongation on DAE 69 (Fig. 3g). These significant C x I interaction effects occurred because, on a relative basis, a greater disparity in _W(PB) was observed between FD and CD than between FW and CW.

At predawn and sunset, the dependency of _W(PB) on _M was more pronounced in Dry than Wet (Tables 1, 3). Although FACE fostered greater recovery in values of _W(PB) to almost predawn (Fig. 4a, b) levels by sunset (Fig. 4e, f), over the nocturnal period, enough time had elapsed to enable recovery of predawn value of _W(PB) to that of _M, regardless of C_a level. Nevertheless, FACE caused slightly less negative values of _W(PB) at both predawn and sunset.

The _S became the most depleted during the third and fourth soil dehydration cycles (see circled letters e–j; Fig. 1c). Although water stress was mild during the third soil dehydration cycle (DAE 96, 106) (Fig. 1c), _W(PSY) was less negative and was more negative in FACE compared with Control (Fig. 5b). But, little difference in _P was observed (Fig. 5a). In contrast, during the fourth soil dehydration cycle, a parabolic trend reversal in _P occurred as _S became depleted from fully rehydrated (_P: ranking of CD > FD > CW > FW) at mid milk ripe on DAE 111 to dehydrated (_P: ranking of FW > CW > FD > CD) at soft dough on DAE 124, as evidenced by a significant C x I interaction during the fourth soil dehydration cycle on DAE 120 (Fig. 5a). Following soil rehydration on DAE 111 (Fig. 1c), greater recovery in _P occurred between CD and FD than between CW and FW. As the number of days since the last irrigation in the Dry treatments increased from 1 on DAE 111 to 14 on DAE 124 (Fig. 1c), compared with CD, _W(PSY) for FD remained less negative and became more negative (Fig. 5b). This resulted in more positive _P (Fig. 5a). This parabolic trend reversal in _P between FACE and Control occurred despite the fact that at soft dough on DAE 124 _M became as negative as –0.91 MPa in FD but was only –0.27 MPa in Control (Fig. 1c).

Regardless of C level, _W(PSY) and became more negative with increasingly negative _M. But, as _W(PSY) became increasingly negative, became negative 11% slower for Control than for FACE (P = 0.20) (Fig. 5c). When _M was at its most negative at –0.91 MPa at soft dough on DAE 124 (Fig. 1c), was –3.2 MPa in Control compared with –3.7 MPa in FACE (Fig. 5c). Conversely, as _W(PSY) became increasingly negative, _P became less positive 11% faster for Control than for FACE (P = 0.20) (data not shown). When