The mid-day responses of wheat ear CO2 and water vapour exchange to full-season CO2 enrichment were investigated using a Free-Air CO2 Enrichment (FACE) apparatus. Spring wheat [Triticum aestivum (L). cv. Yecora Rojo] was grown in two experiments under ambient and elevated atmospheric CO2 (Ca) concentrations (approximately 370 μmol mol−1 and 550 μmol mol−1, respectively) combined first with two irrigation (Irr) schemes (Wet: 100% and Dry: 50% replacement of evapotranspiration) and then with two levels of nitrogen (N) fertilization (High: 350, Low: 70 kg ha−1 N). Blowers were used for Ca enrichment. Ambient Ca plots were exposed to blower induced winds as well the Ca× N but not in the Ca× Irr experiment. The net photosynthesis for the ears was increased by 58% and stomatal conductance (gs) was decreased by 26% due to elevated Ca under ample water and N supply when blowers were applied to both Ca treatments. The use of blowers in the Ca-enriched plots only during the Ca× Irr experiment (blower effect) and Low N supply restricted the enhancement of net photosynthesis of the ear due to higher Ca. In the latter case, the increase of net photosynthesis of the ear amounted to 26%. The decrease in gs caused by higher Ca was not affected by the blower effect and N treatment. The mid-day enhancement of net photosynthesis due to elevated Ca was higher for ears than for flag leaves and this effect was most pronounced under ample water and N supply. The contribution of ears to grain filling is therefore likely to increase in higher Ca environments in the future. In the comparison between Wet and Dry, the higher Ca did not alter the response of net photosynthesis of the ear and gs to Irr. However, Ca enrichment increased the drought tolerance of net photosynthesis of the glume and delayed the increase of the awn portion of net photosynthesis of the ear during drought. Therefore, the role of awns for maintaining high net photosynthesis of the ear under drought may decrease when Ca increases.
Ear photosynthesis plays an important role in the yield formation of wheat. Its contribution to grain yield is higher for species with awned ears and narrow leaves and increases with water stress ( Blum 1985). Analysing 11 durum wheat varieties, Araus et al. (1993) found a 59% yield reduction after ear photosynthesis was suppressed by covering the wheat head. However, gas exchange of the ears has been widely ignored in previous studies of the impact of long-term atmospheric CO2 (Ca) enrichment on wheat ( Teramura, Sullivan & Ziska 1990; Andre & Du Cloux 1993; Mitchell et al. 1993 ; Nicolas et al. 1993 ). Therefore, in the context of current attempts to simulate the behaviour of wheat in a future high-Ca environment, the question arises whether ear and flag leaves differ markedly in their response to elevated Ca. Several studies regarding CO2 and water vapour exchange of cereal ears found distinct phenomenological features compared with flag leaves, such as a higher water use efficiency ( Araus et al. 1993 ; Blum 1985), better ability for osmotic adjustment ( Morgan 1980), and lower carboxylation efficiency. Furthermore, a higher CO2 concentration is usually required to saturate photosynthesis in ears than in flag leaves ( Knoppik, Selinger & Ziegler-Jöns 1986). The net CO2 uptake per unit one-sided organ area is lower for ears than for leaves because of the larger amount of heterotrophic tissue (grain). However, reabsorption processes limit the diminishing effect of the CO2 released from grain respiration on net assimilation ( Knoppik et al. 1986 ; Bort, Brown & Araus 1995, 1996). Apart from these inconsistences, ear photosynthesis is not fundamentally different from leaf photosynthesis because it follows the C3 photosynthetic pathway ( Bort et al. 1995 ).
As previous studies have shown, growth and photosynthetic responses to elevated Ca are greater under drought than under ample water supply ( Morison 1993) and more pronounced for plants that are sensitive to drought than for drought-tolerant plants ( Hunt et al. 1991). Therefore, since gas exchange of ears ( Xu & Ishii 1990; Xu & Ishii 1991) and, in particular, awns ( Blum, Mayer & Golan 1988) is relatively drought-tolerant, one could hypothesize that rising Ca leads to a decreasing relative contribution of ear photosynthesis compared with that of flag leaves to yield formation (‘lower response’ predisposition). This effect could be strengthened further by an increasing awn proportion of the wheat head. The enhancement of wheat photosynthesis due to higher Ca may also be lower for ears than for flag leaves because of a lower carboxylation efficiency of ears and a relatively higher stimulation of leaf photosynthesis by Ca enrichment under lower light conditions ( Long 1991). Alternatively, comparing ears and flag leaves, the higher Ca required to saturate photosynthesis of ears and the higher ambient temperatures of ears ( Long 1991) may lead to a higher enhancement of the net photosynthesis (A) of the ear than the flag leaf under elevated Ca (‘higher response’ predisposition). Both predispositions may be altered further if nitrogen (N) shortage affects the enhancement of net photosynthesis of both ear and flag leaf in different ways due to higher Ca. Reduced N supply may eventually down-regulate ear photosynthesis more than flag leaf photosynthesis under higher Ca. Down-regulation may appear in both organs as a consequence of N-induced sink limitations ( Rogers et al. 1996). However, during grain filling, sink limitations for carbon accumulation in the grains might be more closely coupled to ear photosynthesis than to the more remote flag-leaf photosynthesis.
To evaluate the significance of the two alternative predispositions in a high Ca environment, we explored the CO2 gas exchange of wheat ears within two Free-Air CO2 Enrichment (FACE) experiments. Measurements were carried out under ambient and elevated atmospheric CO2 for wheat ears grown under two regimes of irrigation and N supply, respectively.
MATERIAL AND METHODS
Experimental site and treatment description
Two FACE Experiments were carried out with Ca as the first factor and irrigation (1992–93) and N fertilization (1995–96), respectively, as second factors. The experiments were conducted with hard red spring wheat [Triticum aestivum (L). cv. Yecora Rojo] on a Trix clay loam [fine-loamy, mixed (calcareous) hyperthermic Typic Torrifluvent] at the University of Arizona, Maricopa Agricultural Center located 50 km south of Phoenix, Arizona USA, (33·05 °N, 112·00 °W). Full details of the site, cultivation, irrigation (Irr), fertilization and the free air CO2 enrichment apparatus have already been described ( Wall & Kimball 1993; Pinter et al. 1996 ; Kimball et al. 1999 ). In brief, the experimental design was a strip-split-plot, with two main treatments, Ca (370 and 550 μmol mol−1 during daytime), replicated in four blocks (REP). Each of the eight circular main plots was split into two semicircular subplots to test the effect of the second factor: Irr in the Ca× Irr experiment and N fertilization in the Ca× N experiment. The Ca enrichment was achieved in the high Ca plots (FACE) by using a blower technique. This technique was only applied to the FACE plots in the Ca× Irr experiment. Further evaluation of the blower technique found a significant acceleration of plant development due to the use of blowers in FACE (see below). Therefore, blowers were applied to both Ca treatments in the Ca× N experiment. In accor-dance with previous publications, the ambient Ca treatment of the Ca× Irr experiment is labelled as ‘Control’, the ambient Ca treatment of the Ca× N experiment will be called ‘Blower’.
Ca× Irr experiment
One semicircle of each Ca main plot was irrigated at a target rate based on 100% replacement of potential evapotranspiration (Wet), and the other was irrigated at a target rate of 50% (Dry). The combination of Ca and Irr gave four treatment combinations; Control-Dry (CD), Control-Wet (CW), FACE-Dry (FD) and FACE-Wet (FW).
Ca× N experiment
The main circular Ca plots were divided into halves with each half receiving either 350 (High) or 70 (Low) kg N ha−1 of NH4NO3 fertilizer through the drip irrigation system as described for the Ca× Irr experiment with an additional 33 and 30 kg N ha−1, respectively, from the Irr water itself. Both halves were irrigated at a target rate based on 100% replacement of potential evapotranspiration. The combination of Ca and N resulted in the four treatment combinations Blower-Low (BL), Blower-High (BH), FACE-Low (FL) and FACE-High (FH).
Wheat seeds were sown on 15 December 1992 and 14–15 December 1995. The planting densities at emergence were 130 and 189 plants m−2 in 1992 and 1995, respectively. Fifty percent crop emergence occurred in both experiments on the first Day of Year (DOY); DOY was thus equivalent to day of crop appearance. The crops were harvested on 21 May 1993, and 28–29 May 1996. After sowing, the FACE equipment was constructed on site ( Hendrey 1993). Elevation of Ca commenced with crop emergence (1 January 1993 and 1996), and terminated at the time of grain maturity in the FACE plots (16 May 1993 and 15 May 1996). In the Ca× Irr experiment the absolute Ca concentration in the FACE plots was kept constant at 550 μmol mol−1, which diminished the Ca differences between ambient and elevated Ca plots during the night. An elevated Ca of 200 μmol mol−1 above ambient was maintained throughout 24 h in the Ca× N experiment. The average daytime CO2 concentrations in the FACE and Blower plots were 548 and 363 μmol mol−1, respectively, whereas the night-time values were 598 and 412 μmol mol−1. Further details are given by Kimball et al. (1999) .
Gas exchange measurements
Gas exchange measurements for CO2 and water vapour of wheat ears were carried out through a sequence of days after mid-anthesis (DAA) using a portable closed-exchange (transient) system with a 1 L plexiglass chamber (Model LI-6200; LI-COR, Inc., Lincoln, NE, USA). All measurements were executed in the field. The entire wheat ear was placed inside the cuvette, maintaining the natural orientation of the ear. Ear temperature was monitored using a thermocouple. Net photosynthesis (A) and transpiration (E) were directly measured. Stomatal conductance for water vapour (gs) and internal CO2 concentration (Ci) were calculated as suggested by LI-COR (1990) following von Cammerer & Farquhar (1981). Possible effects of CO2 released by the grains on Ci were not corrected for. All measured ears were harvested and dried at 72 °C for 2 days. After drying, dry weight, head length without awns, and ear surface area were determined and related to the gas exchange parameters. In order to minimize possible interference among subplots, only ears from plants, which were at least 1 m from any plot boarder, were measured.
Boundary layer conductance
The boundary layer conductance (gb, μmol s−1 head−1) was calculated using a linear relationship between ear length without awns (LH, cm) and gb. LH correlated well with the awn area (see below) and also with the number of awns. The number of awns should have a major influence on gb, assuming that individual awns build a network of parallel resistances to the fluxes of water vapour and CO2. The gb–LH relationship was estimated for a subsample of eight dried ears. Ears were sprayed with water, and afterwards the water evaporating from the ear surface was measured and converted into gb. The following linear regression was obtained: gb = (84·4 LH– 709·9)/2, R2 = 0·89, P < 0·01. The multitude of parallel operating awn-resistances to ear fluxes of water vapour and CO2 resulted in gb values (normalized to ear area) that were relatively high in comparison with gs. Because of the high gb level, which relates to a low boundary layer resistance (1/gb), further potential influences on ear gb beside organ size could be neglected.
Gas exchange of intact wheat ears
Gas exchange of wheat ears was measured between 1100 and 1500 h Mountain Standard Time in both experiments. It was repeated seven times in the Ca× Irr experiment (DAA 10, 11, 12, 16, 17, 19, 25, 32 with measurements on DAA 11 and 12 pooled) and four times in the Ca× N experiment (DAA 22, 23, 24 and 27) for all treatments in at least two blocks of the experiment. Additional measurements were carried out in one (DAA 21, 39) or two blocks (DAA 23, 30), respectively, of the Wet sites in the Ca× Irr experiment. On DAA 21, 1993, the mid-day gas exchange measurements were repeated over a range of Ca in CW and FW of the Ca× Irr experiment. On DAA 23, 1996, mid-day gas exchange of wheat ears was measured at 360 and 550 μmol mol−1 Ca in BL, BH, FL and FH of the Ca× N experiment. Before changing Ca during A(Ca) measurements, the cuvette was always reopened, which avoided temperature and humidity changes within a sequence of A(Ca) measurements for one ear.
Gas exchange of awns and glumes in the Ca× Irr experiment
On DAA 10, 11, 12, and 16 the gas exchange of the ear was measured before and after its awns were removed. The values of E and A of the awnless ear are thought to represent bulk rates of ear bracts (hereafter referred to as the glume rate), whereas the difference between awned and awnless heads in gas exchange should represent the awn portion of the total gas exchange. Glume and awn temperatures were calculated using energy balance as suggested by LI-COR (1990). Beside possible water and CO2 releases, Blum (1995) did not find a significant alteration of A and E of glumes after awn removal. However, awn removal might have contributed to an overestimation of glume E and awn A and an underestimation of awn E and glume A, which might have finally resulted in an overestimation of the difference in instantaneous water use efficiency (WUEi = A/E) between awns and glumes. In experiments using an indirect (isotope discrimination technique) and direct methods for measuring water use efficiency of growth and WUEi, Aurus et al. (1993) could not confirm the very high magnitude of differences in WUEi between awns and glumes reported by Blum (1985) for water stress-free wheat. Therefore, our conclusions about the effect of higher Ca on awns and glumes will be drawn from the changing patterns of gas exchange variables rather than from the magnitude of absolute effects.
The one-sided, half-lateral projected surface areas of the intact and awnless ears were measured first with the leaf area meter (Model LI-3100; LI-COR, Inc.). The leaf area meter could not completely resolve the awn area. Applying a correction factor for the intact ears, however, it could be used to estimate the awnless surface of both intact and awnless ears. The half-lateral, projected awn surface areas were measured next using a camera connected to an image analysis software package (Model WV CD20; Panasonic CCTV Camera and AgImage plus, version 1·08; Decagon, Pullman, WA, USA). The awn surface areas were directly measured in all cases when excised awn samples were available from DAA 10, 11, 12, and 16 in 1993. The following regression between half-lateral awn surface area (HAawn, cm) and LH, which did not change during grain filling, was used for the estimation of the awn surface area of intact ears: HAawn = 2·912 LH– 24·95, P < 0·01, R2 = 0·78.
The necessary application of blowers for adequate distribution of CO2 ( Hendrey 1993) led to an artificial increase in night-time temperatures of 1 K, which accelerated crop development in the corresponding plots of both FACE experiments ( Pinter et al. 1900 ). The temperature increase was caused by disruption of the inversion layer above the crop, which pulled warmer air down onto the canopy. The faster crop development, which was not due to higher Ca, led to earlier senescence and crop maturity in the FACE plots of the Ca× Irr experiment and thus reduced the yield differences between FACE and Control ( Table 1). This effect was more pronounced in Wet than in Dry. In Dry, the blower-induced differences in leaf senescence between Control and FACE were largely diminished by the later development of drought under higher Ca. In the Ca× N experiment, blowers were applied to both Ca treatments, injecting ambient and enriched air inside the Blower and FACE plots, respectively. As a consequence of this modified design, the differences in crop development between Ca treatments found in the Ca× Irr experiment were not observed again in the Ca× N experiment ( Table 2, Pinter et al. 1900 ). Nevertheless, despite the blower effect, new insights about the impact of higher Ca on ear gas exchange, particularly in combination with different regimes of N and water supply, still could be gained when the data of both FACE experiments were analysed together.
Table 1. Treatment means ± SE for selected parameters of growth and development for the Ca× Irr experiment a,b,c
BL, Blower–Low; BH, Blower–High; FL, FACE–Low; FH, FACE–High; DOY, day of year.
Days to mid anthesis
88 ± 1·6
89 ± 1·7
88 ± 1·4
87 ± 0·2
Days to maturity
124 ± 0·2
132 ± 0·6
124 ± 0·2
131 ± 1·1
LAI, DOY 109
2·2 ± 0·4
4·1 ± 0·4
1·9 ± 0·25
4·1 ± 0·2
Head area (cm2), DOY 109
17·6 ± 1
24·2 ± 1
20·7 ± 1
23·7 ± 1
Ear number, DOY 109
420·5 ± 15·2
555 ± 27·3
439·5 ± 36·9
560·8 ± 11·9
Head area in percentage of LAI, DOY 109
Statistical analysis [analysis of variance ( ANOVA), regression] used the general linear model procedure in SAS (PROC GLM, SAS Institute Inc., Cary, NC, USA). A NOVA and least-square mean comparisons were carried out for treatment effects on A and gs measured at mid-day for intact heads, glumes and awns. Treatment effects (Ca, Irr, N, Ca× Irr, Ca× N) were separately analysed for each experiment for single days and on pooled data (DAA 11 and 12 of the Ca× Irr experiment referred to DAA 12). According to the experimental design that has been described, Ca was split by a nonrandomized Irr and N effect, respectively. Therefore the errors used for evaluating the main effects (Ca, Irr and N) were (Ca× REP), (Irr × REP) and (N × REP).
Intermittent cloud cover reduced the received photosynthetic photon flux on DAA 16, 17, 25 and 32 ( Fig. 1a) during the Ca× Irr experiment whereas all days had predominantly clear skies during the Ca× N experiment ( Fig. 1b). Further information about the variation of weather variables during the measuring periods is given in Fig. 1 for both experiments. Table 3 summarizes direction and significance of treatment effects observed under these weather conditions in both experiments for A and gs of intact ears, glumes and awns. Further details are given below.
Table 3. Treatment effects on A and gs of the intact ear, glumes and awns in the Ca× Irr and Ca× N experiment, where ***, **, *, NS symbolize significance levels P≤ 0·01, 0·05, 0·1 and not significant, respectively, from ANOVA and regression analysis for both experiments; + and – instead of * indicate the direction of change, + and – in front of NS indicate the direction of a nonsignificant change > 10%
+ + +
+ + +
+ + +
+ + +
Ca enrichment increased A and decreased gs of wheat ears in both experiments ( Figs 2 & 3). Under ample N supply, the enhancement of A for the ear due to higher Ca was lower in the Ca× Irr than in the Ca× N experiment ( Fig. 4a & c). However, drought of ears diminished this difference between both experiments ( Figs 2 & 4c). The impact of higher Ca on gs of ears was similar in the Ca× Irr and the Ca× N experiments ( Fig. 4b & d). It was not affected by N or Irr treatment.
Limited soil moisture reduced A and gs in Control and FACE ( Figs 2 & 5). N deficiency decreased A but not gs ( Figs 3, 4a & b). Regression and variance analysis revealed no general interaction effect Ca× Irr for A and gs of ears. However, interactions were significant for A on DAA 10 and 32 and for gs on DAA 10, 25 and 32, when measurements were taken after rewatering ( Figs 2 & 5). The interaction Ca× N was not significant for gs, but it was consistently significant for A ( Fig. 3). However, the enhancement of A for the ear due to higher Ca under Low N supply was closer to that under High N supply when measurements were taken after a few days of dry-down (DAA 22, 24) and not immediately after Irr (DAA 23, 27).
Awns and glumes differed very often in the responses of A and gs to drought and higher Ca ( Fig. 6). Throughout the dry-down period from DAA 10–16, A decreased for the glumes but not for the awns. Consequently, the awn proportion of net photosynthesis of the ear increased after DAA 10 from 52% in all treatments to 64 and 84% (P < 0·2) on DAA 12 and then to 82 and 112% (P < 0·2) on DAA 16 in Wet and Dry, respectively. The awn proportion of the one-sided ear area remained unaltered by treatment and time during that period (about 45%). Carbon uptake of glumes was consistently stimulated by higher Ca ( Fig. 6a). This effect increased with the duration of the dry-down period from DAA 10 to DAA 16. A significant Ca× Irr interaction was found for net photosynthesis of the glumes on DAA 16. Higher Ca also increased net photosynthesis of awns, but this effect was lower than that for the glumes. The Ca-induced enhancement of net photosynthesis of the awn also did not enlarge with an increasing level of drought across Irr treatments and sample dates ( Fig. 6b). Therefore, the awn proportion of net photosynthesis of the ear increased less under higher than under ambient Ca during the dry-down period from DAA 10 to DAA 16. On DAA 16, the awn share of ear net photosynthesis was 113% under ambient Ca and only 81% under Ca enrichment (P < 0·05). The treatment relations for net photosynthesis of awns and glumes often showed a complementary pattern. Smaller treatment effects for the awns were accompanied by larger treatment differences for the glumes and vice versa ( Fig. 6a & b).
Higher Ca reduced gs of awns and glumes significantly ( Fig. 6c & d). Greater water supply increased the value of gs for both organs. However, rewatering on DAA 16 lead to an increase of gs only for the glumes but not for the awns. Significant and nearly significant Ca× Irr interactions were found for the gs of the awns but not for that of the glumes.
Short-term fluctuations in Ca altered the treatment levels of net photosynthesis of the ear. The treatment levels of ear gs were mostly insensitive to short-term Ca changes ( Figs 7 & 8). In the Wet plots of the Ca× Irr experiment, measurements were taken before recovery of A from previous dry-down. Under these conditions, the A–Ci relationships was similar for ears in CW and FW ( Fig. 7a). Short-term Ca changes did not indicate any photosynthetic acclimation of net photosynthesis of the ear to higher Ca in the High plots of the Ca× N experiment ( Fig. 8a). Under Low N supply, however, the carbon uptake at similar Ca was significantly lower in FACE than in Blower ( Fig. 8a).
Under ample N and water supply, net photosynthesis for the ear at mid-day was enhanced by about 58% due to higher Ca in the Ca× N experiment. A lower stimulation of net photosynthesis of the ear was found for the com-parable Wet plots of the Ca× Irr experiment because of the systematic error that was introduced by the omission of blowers in the Control plots. The exclusive use of blowers in the high Ca plots accelerated senescence and therefore decreased the photosynthetic performance sooner in FW than in CW. Otherwise the enhancement of net photosynthesis of the ear due to higher Ca in Wet of the Ca× Irr experiment most likely would have been similar to that observed in High of the Ca× N experiment. The latter is particularly supported by Fig. 9, where A for FACE ears is plotted versus virtual blower rates simulated for the Ca× Irr experiment. The virtual FACE–Blower relationships for both Irr treatments of the Ca× Irr experiment appear very similar to those observed for the High plots of the Ca× N experiment. The virtual Blower rates, ABlower calc., were calculated from the Control rates, AControl, following ABlower calc. = Acontrol–ΔA. The variable ΔA accounts for the potential nonstomatal effect (see below) of blowers on instantaneous A of ears grown under ambient Ca. This potential blower effect was derived for every DAA and both Irr treatments from the difference between the observed A of FACE ears and the reference A, which was determined from the High treatment of the Ca× N experiment ( Fig. 4a), also as depicted in Fig. 4c.
The + 58% enhancement of mid-day net photosynthesis for the ear ( Figs 4 & 9) is much larger than the parallel increase of 26% observed in mid-day net photosynthesis of flag leaves ( Fig. 10) due to Ca enrichment under ample N and water supply. The intrinsic water use efficiency of ears was higher than that of flag leaves as well; a 28% increase in net photosynthesis of flag leaves due to better water supply corresponded to a 66% increase in gs ( Fig. 10b & d), whereas a 17% increase of ear gs was sufficient for an increase of net photosynthesis of the ear by 20% due to higher Irr ( Fig. 5). Referring back to our initial hypotheses, a dominance of the ‘higher response’ predisposition of net photosynthesis of the ear to elevated Ca is indicated by the results from this study. It suggests that the potential of ears to contribute to grain filling increases more than that for flag leaves under higher Ca. The higher intrinsic water use efficiency of ears than that of flag leaves did not coincide with a lower sensitivity of ears to higher Ca in comparison with flag leaves. Differences in microclimate, morphology and the characteristics of photosynthetic processes between flag leaves and ears as described in the introduction are possible reasons for this behaviour. The enhancement of A observed for mid-day ear and flag leaf photosynthesis was higher than the final yield increase due to higher Ca under ample water and N supply. One possible explanation is a down-regulation of ear and flag leaf A due to sink limitations under higher Ca in the afternoon. Garcia et al. (1998) reported corresponding results from flag-leaf measurements in the Ca× Irr experiment.
Low N treatment and earlier senescence suppressed the Ca-induced enhancement of net photosynthesis of the ear. Stomatal limitations can be excluded as possible causes in both cases. Although FACE ears senesced earlier than Control ears in the Ca× Irr experiment, the effect of higher Ca on gs of the ears did not differ from that in the Ca× N experiment. The results from the latter experiment also show that the induced N deficiency did not affect the FACE–Blower (i.e. FACE–Control) relationship for gs of ears. Processes at either the source or sink sites could be responsible for the observed down-regulation of photosynthesis. At the source sites, decreased pools of soluble proteins in the green ear tissue due to low N treatment and earlier senescence, respectively, could have restricted the enhancement of net photosynthesis of the ear under higher Ca. At the sink sites, limitations for grain filling could have initiated a down-regulation of net photosynthesis by decreasing the Rubisco share in the soluble protein pool under higher Ca ( Rogers et al. 1996 ). Comparing FW with CW and FL with BL, respectively, sink limitations could have resulted from the advanced stage of grain filling due to the blower effect and from the lower grain number developed under Low N treatment, respectively.
A down-regulation of net photosynthesis due to early senescence and Low N treatment was reported also for flag leaves ( Adam et al. 1997 ; Wall et al. 1997 ; Garcia et al. 1998 ). However, it occurred later and was less pronounced than that observed for the ears. In particular, the corresponding results from the Ca× N experiment suggest that the ear portion of canopy net photosynthesis might increase less under limited than under ample N supply in a high Ca environment. However, it has been reported previously for wheat flag leaves that differences in net photosynthesis due to the N status may diminish with increasing water stress ( Xue & Chen 1990). In fact, the down-regulation of net photosynthesis of the ear almost disappeared in both Ca experiments when drought became more severe. Therefore, the N status of the green ear tissue may be important for the effect of higher Ca on net photosynthesis of the ear only under ample water supply. The underlying mechanism is unclear and deserves more research.
There is evidence from both experiments that, unlike net photosynthesis, the gs of ears was quite insensitive to short-term Ca and Ci changes ( Figs 7 & 8). Sionit, Hellmers & Strain (1980), Sionit et al. (1981) and Wall et al. (1995) found experimental evidence for a reduced stomatal responsiveness to treatment differences in Ca when the accumulation of osmotic solutes increased. Ears accumulate more carbohydrates than do leaves during drought for osmotic adjustment ( Morgan 1980), and therefore, their stomates might be less sensitive to short-term Ca changes. Relatively higher levels of carbohydrates in awns than in glumes ( Blum et al. 1988 ) might have decreased also the stomatal sensitivity of awns to rewatering compared with glumes. This would explain the absence of any response of awn gs to rewatering on DAA 16 ( Fig. 6d), although the gs of glumes strongly increased ( Fig. 6c). Furthermore, a higher level of accumulated carbohydrates in FACE than Control ears could have established a lowered sensitivity of ear stomates to altered water supply under enriched Ca leading to the Ca× Irr interaction for ear gs observed on days after rewatering ( Fig. 2b).
The Ca treatment did generally not affect the impact of drought on A and gs of wheat ears in the Ca× Irr experiment ( Fig. 5). The general increase of net photosynthesis for the ear corresponds with the general increase of gs for the ear due to higher Irr under ambient and enriched Ca. Therefore, stomatal effects must have dominated the Irr response of net photosynthesis of the ear in both Ca treatments. The enhancement of mid-day net photosynthesis due to higher Ca for flag leaves was also similar in Dry and Wet ( Wall et al. 1995 ). However, both findings contrast with other more spatially and temporally integrated results from the same experiment which suggest second order benefits from a later development of water stress under higher Ca due to reduced soil water extraction ( Hunsaker et al. 1996 ), stem water flow ( Senock et al. 1996 ) and canopy transpiration ( Kimball et al. 1995 ). The yield and growth reduction caused by lower Irr in the Ca× Irr experiment was substantially smaller in FACE than in Control ( Table 1). Kimball et al. (1995) also found less reduction in canopy assimilation due to lower Irr in FACE than in Control between stem elongation and maturity. The response differences were always greater than the corresponding differences in LAI and ear area together ( Table 1), which suggests that the drought-induced reduction in canopy carbon uptake per unit leaf area was also less under higher than under ambient Ca. The differences between the upper layer and the whole canopy response to drought under ambient and higher Ca may indicate that water status, photosynthetic capacity, and finally life span of the lower rather than the upper canopy layers benefited from the extra water available in FACE compared with Control under Dry conditions.
Although the impact of drought on net photosynthesis of intact wheat ears was generally not decreased by Ca enrichment, gas exchange measurements for glumes and awns indicate that the importance of awns to the achievement of high yields under drought will be reduced under elevated Ca. The awn proportion of ear photosynthesis increased with the level of drought ( Fig. 6a & b), which is consistent with the better developed drought tolerance of awns ( Shmat’ko et al. 1982 ; Araus et al. 1994 ). Under higher Ca, however, the awn portion of net photosynthesis for the ear increased less because of the improved tolerance to drought of the net photosynthesis of the glume. The magnitude of this effect could have been overestimated here, however, if we postulate a faster senescence of awns than glumes due to a greater blower-induced night warming of the awns than the glumes. In fact, a reversal in the mean FACE–Control relationship of net photosynthesis was observed for the awns but not for the glumes on DAA 10 that could be due to a lower N status of FACE than Control awns caused by a blower-induced acceleration in senescence. Yet, the measurements on DAA 12 and 16 were taken during a period of severe drought and therefore should have been only slightly affected by the blower-induced differences in glume and awn senescence as discussed above.
The often-seen complementary pattern of treatment differences for net photosynthesis of awns and glumes could indicate a coupling mechanism between awn and glume net photosynthesis, which may stabilize the net photosynthesis of the ear under changing environmental conditions. However, more detailed studies are necessary to test this hypothesis.
Results from both FACE experiments suggest that the relative role of the ear for grain filling is likely to increase in a future high-Ca environment due to a greater stimulation of net photosynthesis of the ear than the flag leaf by elevated Ca. In addition, the significance of awns for maintaining higher photosynthetic rates of ears under drought conditions will decrease with increasing Ca. N deficiency may limit the increase of net photosynthesis of the ear under higher Ca. This limitation is more pronounced under ample than under deficient water supply.
This research was supported by the Agricultural Research Service, US Department of Agriculture, including the US Water Conservation Laboratory, Phoenix, AZ. Operational support was also contributed by the Bundesministerium für Forschung und Technologie and Hochschulerneue-rungsprogramm, Germany; by Grant #DE-FG03–95ER-62072 from the Department of Energy Terrestrial Carbon Processes Research Program to the University of Arizona, Tucson and Maricopa, Arizona; and by Interagency Agreement No. IBN-9652614 between the National Science Foundation and the USDA-ARS US Water Conservation Laboratory as part of the NSF/DOE/NASA/USDA Joint Program on Terrestrial Ecology and Global Change (TECO). Special thanks are given for technical assistance to Mr Matt Reaves and Mrs. Laura Olivieri.