• Changes in the growth and yield of field-grown potato (Solanum tuberosum cv. Bintje) induced by season-long elevated CO2 and/or ozone concentrations are reported.
• Open-top chambers and unchambered field plots were used to examine crop responses to three CO2 (ambient, 550 and 680 µmol mol−1) and two ozone (ambient and 65 nmol mol−1, 8 h d−1 seasonal mean) treatments applied throughout the 105 d growing season.
• Elevated CO2 increased both above- and below-ground biomass at intermediate and final harvests. Tuber yield at final harvest was increased by c. 40% due to an increase in mean tuber weight rather than tuber number; tuber yield did not differ significantly between the 550 and 680 µmol mol−1 CO2 treatments. Elevated ozone had no significant effect on growth or yield except for the largest size category of tubers, despite extensive visible foliar injury. Significant CO2 × ozone interactions were detected only for senescent leaf number and green leaf ratio.
• Elevated CO2 increases biomass and tuber yield in S. tuberosum cv. Bintje even at high ozone concentrations; these findings are discussed in relation to predicted future atmospheric changes.
Atmospheric carbon dioxide (CO2) concentration has increased dramatically from 280 µmol mol−1 in preindustrial times to the current level of c. 365 µmol mol−1, and could exceed 700 µmol mol−1 by the end of the present century if emissions continue to rise at current rates (IPCC, 1996). Ozone (O3) is considered to be the most important air pollutant in many parts of Europe, North and Central America and the Far East (Heath, 1994; Krupa et al., 1995), not only because of its phytotoxicity (Ashmore & Bell, 1991), but also because the concentration in the troposphere has increased considerably during the past 60 yr (Anfossi et al., 1991) and is likely to continue to increase at an annual rate of 0.5–2.5% in the northern hemisphere (Ashmore & Bell, 1991; Hertstein et al., 1995; Stockwell et al., 1997).
Although the individual effects of elevated CO2 and O3 have been examined under field conditions for many crop species, including potato, there is little information concerning their combined effects on the growth and yield of this economically important crop. Although daily mean CO2 concentrations remain relatively constant during the growing season, tropospheric ozone levels exhibit substantial diurnal and seasonal variation dependent on the prevailing climatic conditions and local geographical factors. Atmospheric CO2 concentration may influence crop responses to O3 by affecting the uptake of the pollutant and/or the supply of photoassimilate to support repair processes and growth (Polle & Pell, 1999). Interactive effects of CO2 and O3 have previously been described for radish (Barnes & Pfirrmann, 1992), wheat (McKee et al., 1997; Mulholland et al., 1998; Donnelly et al., 1999) and soybean (Fiscus et al., 1997). These studies, and a recent review by Polle & Pell (1999), suggest that the severity of ozone-induced foliar injury is generally reduced by elevated CO2, a conclusion that may also apply to growth and yield. The protective influence of elevated CO2 results primarily from reductions in stomatal aperture which decrease the flux of O3 entering the leaves (McKee et al., 1995). However, elevated atmospheric CO2 does not always protect against ozone-induced injury, and responses may depend on: the timing and duration of ozone exposure (Mulholland et al., 1998; Donnelly et al., 1999; Black et al., 2000); and the extent of stomatal closure induced by elevated CO2, which varies considerably between taxa (Barnes & Wellburn, 1998).
There is little information concerning the interactive effects of elevated CO2 and O3 on field-grown potato crops. However, a previous study of cv. Bintje showed that, although elevated CO2 increased above-ground and tuber dry weight during the early stages of the season, this effect was not sustained to maturity, when tuber numbers, but not tuber yield, were increased (Lawson et al., 2001). Season-long exposure to elevated O3 (seasonal 8 h d−1 mean of 50 nmol mol−1) reduced above-ground dry weight but did not affect tuber yield. The objectives of the present study were: (1) to establish the impact of season-long exposure to elevated atmospheric CO2 and/or O3 on the growth and yield of field-grown potato; and to determine whether elevated atmospheric CO2 ameliorates any detrimental effect of ozone when significantly higher seasonal mean O3 concentrations are encountered (65 nmol mol−1).
Materials and methods
Site details, open-top chambers and experimental design
The experimental site was situated at the Sutton Bonington Campus of the University of Nottingham (52°N, 1°15′W). The soil, a sandy loam of the Astley Hall series, was chisel-ploughed and deep-harrowed before planting, and the pre-emergence herbicide (Simazine) was applied at a rate of 1.0 dm3 ha−1. The open-top chambers (OTCs) were of the Heagle design (Heagle et al., 1973), 3.1 m in diameter and 2.4 m in height, with a 45° sloping frustum. Chambers were spaced 10 m apart (north–south and east–west) to avoid mutual shading. The aluminium frame of the OTC was covered with 200 µm PVC in three sections (frustum, middle and bottom). The bottom cover was double-skinned and its inner wall was perforated with 306 × 25 mm diameter holes, through which air from a fan box (Model PSA 402/2, Jones and Attwood, Stourbridge, UK) was evenly distributed. The ventilation rate was sufficient to provide 3–4 air changes min−1 in each chamber. A factorial design was used comprising 18 OTCs, containing three CO2 and two O3 treatments and an unchambered treatment, randomised within three blocks.
Certified seed tubers (Solanum tuberosum L. cv. Bintje) were obtained from the Netherlands. This cultivar is commercially important in Europe for the production of processed food products, and was believed to be susceptible to ozone-induced damage on the basis of controlled environment studies (B. Kollner, pers. comm.). The tubers were stored at 4–5°C before use, in accordance with standardized Changing Climate and Potential Impacts on Potato Yield and Quality (CHIP) protocol. They were then placed upright in wooden boxes and left to sprout in growth rooms at 12°C under continuous light for 4 wk. Cores containing one subsidiary eye were cut from each tuber, to ensure that single-stemmed plants were obtained (Gill et al., 1989). These were taken by inserting the sharp end of a potato peeler into the tuber and gently scoring around the eye to remove a conical piece of tissue containing a single eye; each core weighed approx. 8 g. The corer was disinfected with alcohol after each incision to avoid possible transfer of infection. The cores were placed on trays lined with damp paper, covered with transparent polythene, and placed in a growth room at 18°C for 48 h to allow them to suberise. The remaining tubers were used to plant guard areas around the OTCs.
The OTCs were placed on a prepared seedbed with a flat soil surface (i.e. without ridges). The cores were planted at a depth of 10 cm in a diamond pattern on 12–13 May 1999 to provide a density of 13.3 plants m−2. Spacings between plants were 30 cm within rows and 25 cm between rows.
Gas supply and environmental monitoring
Elevated CO2 was applied for 24 h d−1 between emergence and final harvest, while ozone was applied to the elevated O3 treatments for 8 h d−1 (0900–1700 GMT) for 5 d wk−1 except during periods when rain fell. The three chambered CO2 treatments comprised ambient air, 550 and 680 µmol mol−1 CO2 (chAA, c550 and c680, respectively); the elevated O3 treatment involved a target seasonal 8 h d−1 mean of 60 nmol mol−1 under both ambient (oz) and elevated CO2 conditions (oz550 and oz680, respectively). In each case, CO2 and/or O3 was added to unfiltered ambient air. CO2 was supplied from a 5 tonne storage tank with a vaporization facility capable of producing up to 50 kg h−1 at a working pressure of 2 kg cm−2 (Hydrogas, Middlesex, UK) and distributed using a quarter turn ball valve (20 mm diameter), controlled by a 24-V DC actuator (Model R-O POS 4/20, J + J Electric Actuators, Worcester, UK) using a computer. Ozone was generated from O2 by electrical discharge (Model LN103, Ozonia, Switzerland). Gas concentrations were routinely measured in eight plots (one from each treatment plus a roving line to test plots which were not routinely monitored) at 30 min intervals. Air temperature, soil temperature and solar radiation were also logged at 30 min intervals (21X Micrologger, Campbell Scientific, Shepshed, UK). Soil moisture was monitored using septum tensiometers (Skye Instruments Ltd, Powys, Wales, UK) and maintained above 70% of field capacity. Further details of environmental monitoring are provided by Lawson et al. (2001).
After determining the nutrient status of the soil, fertilizer was applied at the following rates according to recommended agronomic practice: N, 250 kg ha−1; P, 100 kg ha−1; K, 150 kg ha−1. Slug pellets and rat bait were placed in each chamber and over the surrounding area. The crop was sprayed at weekly intervals between emergence and 53 d after emergence (DAE) and then at 10 d intervals until final harvest, using a fungicide–insecticide mixture containing Fubol (metalaxyl and mancozeb 100 g l−1) and Metaphor (demeton-S-methyl 1 ml l−1). The plots were weeded by hand and watered using trickle irrigation. The plants were supported using bamboo canes to prevent lodging in accordance with standard CHIP protocol.
Nondestructive growth analysis
Ten randomly selected plants in each plot were tagged and used for weekly nondestructive growth measurements; these included plant height, total leaf number, number of senescent leaves and visual estimates of ozone damage; the scoring procedure used to quantify visible injury is described by Donnelly et al. (2001). The ratio of green leaves to the total number of leaves per plant was used to estimate green leaf ratio (GLR).
Destructive growth analysis
Destructive harvests were carried out at: (1) tuber initiation (15 June, 22 DAE); (2) approximately half-way through the growing season (12 July, 49 DAE); and (3) when 50% of the leaves had senesced (final harvest, 105 DAE). At the final harvest, above-ground biomass was harvested on 6 September (105 DAE) and the below-ground biomass 1 wk later on 13 September 1999 (112 DAE), according to standard agronomic practice. After crop emergence, selected areas of each plot remained undisturbed for specific harvests. Five randomly chosen plants (excluding peripheral plants) were harvested from these predetermined areas within each plot at 22 and 49 DAE; 15 plants were harvested from each plot at crop maturity. The plants were subdivided in to above- and below-ground biomass before determining the number and dry weight of green and senescent leaves, stem dry weight, tuber number and tuber dry weight. Leaf area index (LAI) was calculated from measurements of green leaf area, while specific leaf area (SLA) was calculated from green leaf area and the corresponding leaf dry weight. The above-ground material was then placed in paper bags and dried at 80°C for 48 h. Tubers were excavated using a garden fork, placed in paper potato sacks and stored in darkness at 8°C. Total tuber fresh weight was determined for each plot, before separating the tubers into < 35 mm, 35–50 mm, 50–60 mm and > 60 mm size classes according to the standardized CHIP protocol, which followed commercial grading categories. A subsample of the tubers from each category was oven-dried at 80°C for 72 h. Individual tuber weights were recorded for each category.
Data were analysed as a seven-treatment randomized block by ANOVA using Genstat 5 (Lawes Agricultural Trust, IACR, Rothamsted, UK). As the seven treatments comprised a factorial three CO2 × two O3 treatment structure plus an unchambered ambient air treatment, the treatment sum of squares was partitioned into a chambered vs unchambered contrast and then into the three CO2 × two O3 factorial within the OTC treatments. Effects of CO2 were partitioned between (a) ambient vs. elevated (combined 550 and 680 µmol mol−1) treatments, and (b) 550 vs. 680 µmol mol−1 CO2. Split plot analysis was used to determine whether the number and weight of tubers in each size class at final harvest differed significantly between treatments. Repeated measures analyses were used for measurements made for the same plants at different times during the season. Where there were no significant CO2 × O3 interactions, the combined means for the two ambient (chAA and oz), 550 nmol mol−1 (c550 and oz550) and 680 nmol mol−1 CO2 treatments (c680 and oz680) were used to improve the sensitivity of the statistical analysis when testing for significant CO2 effects.
Climatic conditions and CO2 and O3 concentrations
Climatic conditions in the OTCs differed from those in the unchambered plots mainly through a reduction in incident solar radiation and an increase in air temperature (Table 1). The OTC frames and covers decreased mean daily solar radiation by 26% relative to the unchambered plots (Table 1). Mean air temperature was 1.3°C higher in the OTCs, although the difference in soil temperature was small (0.1°C). Daily mean saturation vapour pressure deficit was 10% higher in the OTCs.
Table 1. Solar radiation, air and soil temperatures and saturation vapour pressure deficit (SVPD) in the ambient air (AA) and open-top chamber plots (OTC)
Data for OTCs are means for all six treatments ± 1 SE.
Solar radiation (MJ m−2)
Average daily total
15.5 ± 0.36
11.5 ± 0.30
Accumulated seasonal total
1689 ± 37.0
1254 ± 33.0
Air temperature (°C)
14.9 ± 0.51
16.2 ± 0.14
Daily mean maximum
16.7 ± 0.67
17.8 ± 0.22
Daily mean minimum
9.5 ± 0.49
9.9 ± 0.13
Soil temperature (°C)
6.1 ± 0.16
16.2 ± 0.15
0.31 ± 0.04
0.34 ± 0.03
Seasonal mean CO2 and O3 concentrations did not differ greatly from target levels (Fig. 1a,b). Daily mean ambient CO2 mole fraction was 398 µmol mol−1, while values for the elevated CO2 treatments were within 2% of their target values (543 and 694 µmol mol−1). Seasonal mean 8 h d−1 O3 mole fraction in the elevated O3 treatment was 64.7 nmol mol−1 (Fig. 1b); this equates to a seasonal mean of 82 nmol mol−1 for the 5 d wk−1 when O3 was applied. Seasonal mean values (7 d wk−1) for the chambered and unchambered ambient air plots were 25 and 27 nmol mol−1. The mean AOT40 (accumulated ozone exposure above a threshold of 40 nmol mol−1) value for the ambient O3 OTC treatments was 1822, compared with 2392 nmol mol−1 in the unchambered field plots; the mean value for the elevated O3 treatment was 27 113 nmol mol−1 h.
Effects of elevated O3
There was no significant effect of elevated O3 on plant height (Fig. 2a), leaf number (Fig. 2b), leaf area index (Fig. 3a) and specific leaf area (Fig. 3c). However, the dry weight of senescent leaves was significantly greater under elevated O3 at 49 DAE (P < 0.05) but not at 22 DAE (Tables 2, 3). The increase in the number of senescent leaves with time differed with the CO2 and O3 treatments (P < 0.01; Fig. 2c), as the number of senescent leaves in O3-treated plants increased more rapidly under ambient than under elevated CO2; the reverse pattern was apparent for green leaf ratio (P < 0.01; data not shown). The first symptoms of ozone injury were recorded at 53 DAE as necrotic lesions on the adaxial surface of the leaves; these increased in size as the season progressed and in some cases > 50% of the leaflet surface area showed visible O3 injury at final harvest (Donnelly et al., 2001). No significant O3 effect was detected for any of the variables measured destructively at tuber initiation (22 DAE; Table 2), although the dry weight of senescent leaves was increased by 36% at the intermediate harvest (49 DAE; P < 0.05; Table 3). At the final harvest (105 DAE), above-ground dry weight, tuber numbers and dry weights within individual size categories were all unaffected by ozone (Table 4). However, single tailed t-tests revealed that tuber fresh weight in the largest size category was reduced by elevated O3 in the ambient and 680 µmol mol−1 CO2 treatments (P < 0.05), but not in the 550 µmol mol−1 treatment.
Table 2. Effects of elevated CO2 and O3 on biomass components expressed of Solanum tuberosum cv. Bintje per ground area at tuber initiation (22 d after emergence)
Total plant dry weight includes above- and below-ground dry weights.
2 Below-ground dry weight includes stolons, tubers and roots.
AA, unchambered ambient field plots; OTC plots were supplied with ambient, 550 or 680 µmol mol−1 CO2 under either ambient (chAA, c550 and c680) or elevated O3 (oz, oz550 and oz680, respectively). ANOVA summaries are presented. ch, comparison between the unchambered and all chambered plots. Significance levels are represented by P-values; ns denotes no significant difference at the 5% level.
Table 3. Effects of elevated CO2 and O3 on biomass components and tuber number of Solanum tuberosum cv. Bintje expressed per unit ground area at the intermediate harvest (49 d after emergence)
Total plant dry weight includes above- and below-ground dry weights.
2 Below-ground dry weight includes tubers and roots.
AA, unchambered ambient field plots; OTC plots were supplied with ambient, 550 or 680 µmol mol−1 CO2 under either ambient (chAA, c550 and c680) or elevated O3 (oz, oz550 and oz680, respectively). ANOVA summaries are presented. ch refers to the comparison between the unchambered and all chambered plots. Significance levels are represented as P-values; ns denotes no significant difference at the 5% level.
Table 4. Effects of elevated CO2 and O3 on biomass components and tuber numbers of Solanum tuberosum cv. Bintje in various size categories at final harvest (105 d after emergence for shoot material, 112 d after emergence for tubers). AA indicates unchambered ambient field plots; the open-top chamber (OTC) plots were supplied with ambient, 550 or 680 µmol mol−1 CO2 under either ambient (chAA, c550 and c680) or elevated O3 (oz, oz550 and oz680, respectively)
Total plant dry weight includes above- and below-ground dry weights. ANOVA summaries are presented. ch refers to the comparison between the unchambered and all chambered plots. Significance levels are presented as P-values; ns denotes no significant difference at the 5% level. SED, standard error of the difference.
Plant height and total leaf number were significantly lower (P < 0.001) under 680 µmol mol−1 CO2 than in the 550 µmol mol−1 and chambered ambient air treatments during the latter stages of the season (Fig. 2a,b); this effect resulted from an earlier and more pronounced slowing of growth in the 680 µmol mol−1 treatment. No CO2 effect was detected for either the number of senescent leaves per unit ground area (Fig. 2c) or green leaf ratio (data not shown). The only parameter affected at tuber initiation (22 DAE) was SLA, which was reduced by elevated CO2 (P < 0.05; Fig. 3c). At 49 DAE, LAI was increased by 21 and 19%, respectively, in the 550 and 680 µmol mol−1 CO2 treatments relative to chambered ambient air control plants; green leaf dry weight per unit ground area was increased by 25 and 31% (P < 0.05; Fig. 3a,b). There was no significant difference in response between the 550 and 680 µmol mol−1 treatments.
Elevated CO2 had no significant effect on the above- and below-ground dry weight parameters examined at tuber initiation (22 DAE; Table 2). However, by 49 DAE, total plant, above-ground, stem and green leaf dry weights were all c. 30% greater in the 550 and 680 µmol mol−1 treatments than in the chambered ambient air control (P < 0.05; Table 3; Figs 3a,b, 4a). A similar pattern was apparent at the final harvest, when above-and below-ground biomass were again significantly greater under elevated CO2 (Table 4). Total plant dry weight was increased by 32 and 37% in the 550 and 680 µmol mol−1 treatments (P < 0.01); above-ground, stem and leaf dry weights were increased by 20–25% (P < 0.05; Table 4).
Elevated CO2 had no effect on tuber number per unit ground area at 49 DAE (Table 3). However, total tuber dry weight was significantly greater under elevated CO2 at both the intermediate (P < 0.05) and final harvests (P < 0.001; Table 4). At 49 DAE, tuber dry weight was increased by 46 and 73% in the 550 and 680 µmol mol−1 CO2 treatments; the corresponding increases at final harvest were 36 and 40%. At the final harvest, the tubers were divided in to four size classes (< 35, 35–50, 50–60 and > 60 mm). Although elevated CO2 induced no significant change in tuber number or weight expressed on a fresh or dry weight basis in the two smallest size classes, tuber dry weight in the two largest size classes and tuber number in the largest size class were increased (P < 0.05 and 0.002, respectively; Table 4). Elevated CO2 also increased tuber numbers in the 100–200 and > 200 g weight classes and reduced tuber numbers in the < 100 g category (P < 0.05; Fig. 4).
CO2 × O3 interaction
There were significant CO2 × O3 interactions (P < 0.05) for both the number of senescent leaves (Fig. 2c) and green leaf ratio; the number of senescent leaves was increased by elevated O3 in the ambient and 550 µmol mol−1 CO2 treatments but not in the 680 µmol mol−1 treatment. No significant interaction was detected for any of the other variables examined at the three destructive harvests (Tables 2, 3, 4).
Effects of open-top chambers
Plant height and stem dry weight were c. 33% lower in the unchambered field plots than in the OTC treatments at tuber initiation (22 DAE; P < 0.01; Fig. 2a; Table 2). A significant chamber effect was also apparent at 49 DAE when tuber number was significantly lower in the unchambered plots (P < 0.05; Table 3).
The absence of significant adverse effects of elevated O3 on tuber yield, with the sole exception of tuber fresh weight in the largest size category (Table 4), is surprising in view of the observation that visible foliar injury became apparent in the elevated O3 treatment at 53 DAE and increased in severity as the season progressed (Donnelly et al., 2001); the dry weight of senescent leaves was increased at the intermediate harvest (Table 3). A similar lack of effect was reported by Foster et al. (1983a), Finnan et al. (unpublished) and Lawson et al. (2001), all of whom found that yield components were unaffected despite the presence of visible foliar injury. However, in an observation analogous to the present study, Pell et al. (1988) reported that the total number and dry weight of tubers in the largest size class (> 60 mm) were reduced by elevated O3 (seasonal 10 h d−1 mean of 51 nmol mol−1) in cv. Norchip, whereas the weight of tubers in the smallest size class (< 35 mm) increased. These changes reduced the total weight of tubers produced under elevated O3 in the absence of effects on overall tuber number. Ozone-induced reductions in total tuber number have also been observed in experiments using container-grown plants (Foster et al., 1983b; Pell & Pearson, 1984), although Foster et al. (1983a) suggested that the containers may have influenced the impact of O3 by restricting tuber initiation. In contrast to our findings, Clarke et al. (1990) reported that tuber yields were reduced by > 30% when 75% of the crop exhibited foliar injury in a study carried out under ambient O3 conditions in the eastern United States using eight varieties which differed in O3 sensitivity. They found that up to 75% of the crop exhibited foliar injury in the most sensitive varieties when the cumulative ozone exposure reached or exceeded 110 µmol mol−1 h, and concluded that tuber yields were reduced only when moderate to severe foliar symptoms developed.
Tuber yield for the variety used here, cv. Bintje, has proved to be relatively insensitive to O3 in experiments carried out in the UK (Lawson et al., 2001 and present study) and Ireland (Finnan et al., unpublished). In tests of the O3 sensitivity of 12 potato cultivars in the United States, Heggestad (1973) concluded that the more sensitive varieties originated from areas experiencing low ambient O3 concentrations. In OTC experiments in California, Foster et al. (1983b) showed that a cumulative dose of 44.2 µmol mol−1 h O3 induced severe foliar damage and reduced tuber yield by 45%; however, other studies of defoliation by Colorado beetle indicate that potato is capable of compensating for extensive foliar damage (Hare & Moore, 1988) as no yield losses occurred even after 50% defoliation. Clarke et al. (1990) suggested that similar adaptive compensatory mechanisms were induced irrespective of the factor(s) responsible for defoliation. Moreover, Mosley et al. (1978) suggested that early maturing varieties may be more sensitive to O3-induced injury than late-maturing varieties. These factors may explain why cv. Bintje proved less sensitive to O3 than initially anticipated: this variety was bred in the Netherlands where ambient O3 concentrations are relatively high compared to the British Isles, and the seed tubers used were obtained from a Dutch supplier; cv. Bintje is also a late-maturing variety harvested in September or October rather than an early variety harvested in July.
AOT40 maps produced for wheat (May-July, daylight hours only) using data collected between 1990 and 1994 show a gradient from 2000 nmol mol−1 h in northern Europe to values exceeding 10 000 nmol mol−1 h in central Europe (PORG, 1997), although values approaching 30 000 nmol mol−1 h have been recorded in some parts of southern Europe (Fowler et al., 1999). The mean AOT40 value for the elevated O3 treatment in the present study of 27 113 nmol mol−1 h therefore exceeded those typically experienced in central Europe, was substantially greater than those generally recorded across France, Belgium and the Netherlands, and was 8–10-fold higher than in the much of the northern UK and Scandinavia; the AOT40 applied therefore substantially exceeds current ozone exposures over much of Europe. The results suggest that tuber yields in cv. Bintje are unlikely to be greatly affected by O3 levels currently experienced across Europe. Lawson et al. (2001) have previously reported that tuber yield for the same cultivar was unaffected following exposure to an AOT40 of 12 460 nmol mol−1 h.
The observation that plant height and leaf number per unit ground area were significantly lower in the 680 µmol mol−1 CO2 treatment during the latter stages of the season (Fig. 2a,b) contrasts with previous reports that height, leaf number and leaf area were unaffected by elevated CO2 (Miglietta et al., 1998; Lawson et al., 2001). However, plant height and leaf number were not significantly affected by elevated CO2 at 49 DAE (Fig. 2a,b), suggesting that effects on crop development were restricted to the latter stages of the season. An acceleration of phenological development by elevated CO2 has previously been reported for potato by Miglietta et al. (1998), who showed that flowering occurred earlier in the FACE rings used in their experiment. Although Wheeler & Tibbitts (1997) reported an increase of up to 50% in stem length under elevated CO2, their experiments were carried out in growth rooms under relatively low light conditions (photosynthetic photon flux density (PPFD), 22 mol m−2 d−1).
The impact of elevated CO2 on crop performance was much greater than that of O3, although the extent of the effects induced varied during the season. None of the variables examined showed significant CO2 effects at tuber initiation (22 DAE), suggesting that the duration of exposure to elevated CO2 was insufficient to affect crop growth by this stage. However, significant responses were observed at the intermediate (49 DAE) and final harvests (105 DAE), when above- and below-ground biomass were both increased by elevated CO2. An important factor contributing to this effect was the greater LAI produced under elevated CO2, which increased the photosynthetic area available for assimilate production. Interestingly, no differences in the dry weight of the various plant components were detected between the 550 and 680 µmol mol−1 treatments, suggesting that a saturation concentration may exist, above which no further benefits of CO2 enrichment are obtained. A similar pattern was observed in an earlier experiment with the same variety (Lawson et al., 2001), whereas studies of wheat (Mulholland et al., 1998; Donnelly et al., 1999) suggest that grain yield increases linearly as atmospheric CO2 concentration rises from ambient to 550 and finally to 680 µmol mol−1.
The influence of elevated CO2 on tuber size was greatest for the two largest size classes, for which tuber number and weight were both increased; no significant effect was detected for the smaller size classes (Table 4). Finnan et al. (unpublished) also found that elevated CO2 increased tuber number in the largest size class, thereby increasing total yield within this category. These results suggest that enhanced tuber growth under elevated CO2 increases mean tuber size at maturity, increasing the proportion of tubers within the larger size classes (Fig. 4).
Tuber yield, expressed either on a fresh or dry weight basis (Tables 2, 3), was significantly increased by elevated CO2 at both the intermediate and final harvests due to an increase in mean tuber weight, rather than tuber number per unit ground area; similar results were reported by Finnan et al. (unpublished). Miglietta et al. (1998) also found that tuber yield was increased when potato was grown under elevated CO2 in a FACE system, although the observed yield enhancement in their experiment resulted from an increase in tuber number, which may reflect the relatively low planting density used. By contrast, De Temmerman et al. (2000) and Lawson et al. (2000) found no increase in tuber yield at final harvest under elevated CO2, perhaps because the substantially greater planting density employed (20 plants m−2 vs 13 plants m−2 in the present study) increased the extent of competitive interactions between adjacent plants. Wheeler & Tibbitts (1997) also found no increase in tuber yield under elevated CO2, even though additional tubers were initiated during the early stages of crop development. Experiments with three potato varieties in controlled environment chambers have shown that tuber yield was increased by elevated CO2 under conditions of low PPFD, but was decreased under high PPFD (Wheeler et al., 1991). There is also evidence that yield may be reduced by supra-optimal CO2 levels. For instance, Mackowiak & Wheeler (1996) reported that tuber yield reached a maximum at 1000 µmol mol−1 CO2, but was reduced at 10 000 µmol mol−1; they concluded that assimilate partitioning to the tubers was adversely affected in the latter treatment. In the present study, yield was significantly increased under 550 µmol mol−1 CO2 relative to the ambient air treatment, but no further increase was observed at 680 µmol mol−1. These observations suggest that interactions between cultivar, atmospheric CO2 concentration and potentially limiting environmental variables such as radiation, water and nutrient availability are important in determining the direction and extent of the yield responses induced.
Although an AOT40 exceeding 27 000 nmol mol−1 h was applied in the present study, this was insufficient to reduce tuber yield despite the occurrence of extensive visible foliar injury during the latter stages of development. By contrast, elevated atmospheric CO2 increased tuber yield by 40% as a result of increases in mean tuber weight rather than tuber number. There was little difference in response between the 550 and 680 µmol mol−1 treatments, suggesting that a threshold concentration may exist above which no further benefits for crop growth are obtained. The results suggest that tuber yields in potato cv. Bintje are likely to be significantly increased by predicted future increases in atmospheric CO2 and unaffected by concurrent increases in tropospheric O3 levels.
This work was funded by EU Contract No. ENV4-CT-0498. Thanks also to Ann-Marie Tulloch for technical assistance.