Stomatal development and CO2: ecological consequences


Author for correspondence: F. I. Woodward Tel: +44 (0)114 2224374 Fax: +44 (0)114 2220002 Email:


  •  Stomatal density responses by 48 accessions of Arabidopsis, to CO2 enrichment, broadly parallel interspecific observations.
  •  Accessions differing in the degree of stomatal response to both CO2 and drought differed in flower production. Under well watered conditions flowering benefits from a small reduction in stomatal density with CO2 enrichment, but benefits from a large reduction under drought.
  •  Stomatal density increases with altitude in Vaccinium myrtillus but is also strongly influenced by exposure. Exposed plants had higher stomatal densities than plants at the same altitude but in a community of individuals. This difference might be explained by systemic signalling within the plant as mature leaves detect both irradiance and [CO2], subsequently controlling the response of stomatal development in developing leaves.
  •  Plants with the highest stomatal densities also had the highest stomatal conductances and photosynthetic rates. This suggests that signalling from mature to developing leaves predetermines the potential of the developing leaf to maximize its photosynthetic potential, including associated features such as nitrogen allocation, during early stages of development in the enclosed bud.


The observation that changes in atmospheric CO2 concentration have led to changes in stomatal density (Woodward, 1987) provided evidence that plants can detect and respond to anthropogenic influences on atmospheric composition. The initial, clear-cut observations suggested that a reduction in stomatal density with CO2 enrichment is a general response by plants. Accumulated observations from a wider range of species (Woodward & Kelly, 1995; Royer, 2001) indicated that although the average stomatal response was a reduction with CO2, this masked a wide range of species-specific responses from very large reductions to large increases. This wide variability of response makes the identification of mechanisms difficult because of likely pleiotropic responses due to the very different interspecific genetic backgrounds. Such differences also hamper attempts to identify the ecological consequences of the stomatal responses to changing CO2 concentration.

Access to the full genotype of Arabidopsis thaliana (The Arabidopsis Initiative, 2000), and the availability of a wide range of accessions with different genetic mutations, indicate an ideal model species for investigating the consequences of the stomatal density response to CO2. The work presented here first examines the stomatal responses of different accessions of Arabidopsis to CO2, then explores impacts on physiological and ecological behaviour and finally investigates the application of this understanding to the field.

Materials and Methods


Accession responses to CO2 doubling 48 accessions of Arabidopsis (from the Nottingham Arabidopsis Stock Centre, UK, were grown from seedlings for 6 wk in controlled environments chambers (Fitotron SGC2352/FM/HFL Sanyo Gallenkamp, UK). The mean growth temperature was 20°C during the 16 h photoperiod and 16°C at night, with a constant rh of 55%. The mean irradiance at plant height was 200 µmol m−2 s−1 and two CO2 mole fractions of 350 and 700 µmol mol−1 were provided in different growth cabinets. The positions of the plants within the chambers were randomised every two days and all plants were swapped weekly between chambers. Ten fully expanded leaves from each CO2 mole fraction, and for each accession were analysed for stomatal density on the adaxial and abaxial leaf surfaces, and for the mid-region of the leaf on both sides of the rachis.

Response to drought The Arabidopsis accessions Col-0 and Ws were grown under the same environmental conditions and CO2 mole fractions as indicated above. 16 plants of each accession were grown under well-watered conditions and 16 grown under droughted conditions (pots were watered to saturation when the soil water potential reached −0.2 MPa, determined gravimetrically), in both CO2 concentrations. Plants were grown from seedling to flowering and stomatal density (on rosette leaves) and flower production were recorded under all treatments for both accessions. The whole experiment was repeated on two occasions.

Key to processes Initial observations indicated that high irradiance during growth led to marked anthocyanin accumulation in the leaves of Arabidopsis and a diminished stomatal response to CO2 enrichment in the accession Col-0. These observations were quantified for 4 different accessions of Arabidopsis: Ct, Cri, Kas and Ms, with different responses of stomatal density to CO2 enrichment. Plants were grown in two treatments, in the conditions outlined above and under a higher irradiance of 510 µmol m−2 s−1.

Key to mechanismsLake et al. (2001) described a technique for investigating the nature of CO2 sensing in Arabidopsis. In this technique the ambient CO2 environment of expanding leaves is that of the growth chamber, while the CO2 environment around recently fully expanded leaves is controlled independently with a transparent and air conditioned cuvette. The accession Col-0 was used for the majority of the experiments. The experiments were also repeated using the jasmonate deficient mutant, fad-4 (Hugly et al., 1991).

Field observations Plants from six different altitudinal populations of Vaccinium myrtillus L. were investigated for the developmental responses of stomata to the natural changes in environment, including atmospheric CO2 concentration, associated with altitude at a range of field sites in Scotland (Butler, 1985; Woodward, 1986). Stomatal densities were determined on 10 leaves from each existing population. In addition two populations, both from 200 m of altitude, were transplanted to an exposed plateau at an altitude of 905 m. Stomatal density was recorded for these individuals in the first and second growing seasons at the higher altitude.

Profiles of photosynthetically active radiation were determined through the vegetation at the highest altitude with a quantum sensor (Woodward & Yaqub, 1979).


Accession responses

The responses of stomatal density to CO2 doubling in 110 species (Woodward & Kelly, 1995; plus new unpublished measurements) indicate (Fig. 1) a mean reduction in density of 29% with CO2 doubling. The histogram indicates the responses of the 48 accessions of Arabidopsis to CO2 doubling. The mean response is slightly lower, at 22%, than the different species but indicates that a wide range of stomatal density responses is possible within one species and also suggests Arabidopsis as an ideal model species for investigating the nature and ecological relevance of the CO2 response of stomatal development.

Figure 1.

Responses of stomatal density to CO2 doubling: between species (solid line) (Woodward & Kelly, 1995; unpublished observations) and for different accessions of Arabidopsis (histogram).

Drought experiment

The drought experiment with the accessions Col-0 and Ws indicated the significantly (P < 0.05) greater sensitivity of stomatal density to CO2 in Ws, under both well-watered and droughted conditions (Fig. 2, the CO2 × water supply–accession interaction was significant at P < 0.05). The sensitivity of Col-0 to CO2 increased significantly (P < 0.05) under drought, but the sensitivity of Ws did not change significantly.

Figure 2.

Response of stomatal density (mean of both leaf surfaces) to CO2 enrichment under well-watered and droughted conditions for Col-0 (open bars) and Ws (dotted bars) accessions of Arabidopsis.

At a CO2 concentration of 350 ppm drought significantly reduces flower production (Fig. 3) in Ws but not for Col-0. However, at 700 ppm the situation is almost reversed, Col-0 shows a significant reduction in flower production with drought while the response of Ws is much smaller. Compared with plants at a CO2 concentration of 350 ppm, CO2 enrichment virtually restores the well-watered response of flower production in Ws but not in Col-0.

Figure 3.

Responses of flower production in Col-0 and Ws to CO2 enrichment and water supply. Cross hatched columns, well watered; open columns, droughted. Columns shown with 95% confidence limits.


The responses of the adaxial stomatal density to CO2 doubling (Fig. 4), under the standard growth conditions, indicate a wide range of responses by the four accessions of Arabidopsis, Ct, Cri, Kas and Ms, from increases to large decreases. Under the higher irradiance, anthocyanin production was observed for all accessions and in all cases the stomatal response to CO2 enrichment was significantly diminished and effectively negated. It was notable that this effect occurred irrespective of the direction of change in stomatal density under the control conditions.

Figure 4.

Responses of adaxial stomatal density in four accessions of Arabidopsis, Ct, Cri, Kas and Ms, to standard growth conditions (dotted columns) and to high irradiance (open columns).


Lake et al. (2001) demonstrated that independently controlling the CO2 concentration, around developing and mature leaves, shows that mature leaves detect and control the CO2 response of stomatal initiation in developing leaves. This response is shown for the stomatal index responses of Col-0 (Fig. 5), where there is no significant effect of the developing leaf CO2 concentration (P > 0.05) on stomatal index (the fraction of epidermal cells which are stomata). However the same analysis of stomatal density indicates a different response from index (Fig. 6). In this case a high CO2 concentration around the expanding leaf over-rides significantly (P < 0.05) any signal for increased density when the mature leaves are at the lower CO2 concentration. As stomatal index is not influenced, then this suggests an impact on leaf area expansion.

Figure 5.

Responses of stomatal index to the CO2 concentration surrounding developing and mature leaves.

Figure 6.

Responses of stomatal density to the CO2 concentration surrounding developing and mature leaves.

It was observed that stomatal index in the fad-4 mutant slightly increased with CO2 enrichment, opposite to the response shown by Col-0 the parental background of fad-4 (Fig. 7) and significantly so at P < 0.05.

Figure 7.

Stomatal index responses to CO2 doubling for Col-0 (open columns) and fad-4 (dotted columns).

Field observations

Five of the populations of V. myrtillus were located in moderately dense communities of plants. One population at 1095 m occurred in a very exposed site of only scattered individuals of plants. It was noted at altitudes of 900 m, and greater, that the growth form of V. myrtillus changed markedly from about 15 cm tall in areas of late snow lie to 2–3 cm tall in exposed areas. Subsequent growth of transplants from these populations to the lowlands (20 m) demonstrated that the dwarf form was genetically controlled, with no significant difference in height growth over 2 yr of observation. Therefore this population, from 1095 m, has been treated separately in this analysis.

Adaxial stomatal density increases with altitude (Fig. 8) in a close to linear fashion. Abaxial density did not change significantly with altitude. The prostrate population developed with a significantly higher stomatal density than the nearby population located in a community of plants. The individuals transplanted from 200 m to an exposed site at an altitude of 905 m developed with marked increases in adaxial stomatal density (Fig. 9, data shown for first year of growth but similar responses were shown in the second year of growth).

Figure 8.

In situ observation of stomatal density with altitude for Vaccinium myrtillus. Data points shown with 95% confidence limits. The isolated point is from the high altitude exposed site. Regression line, excluding the isolated point, r2 = 0.95.

Figure 9.

Data as for Fig. 8 but including observations of stomatal density for individuals originating at 200 m and transplanted to 905 m (open square).


The first publications describing the reductions of stomatal density with increasing CO2 concentration (Woodward, 1986, 1987) demonstrated that plants in experimental systems and also stored as herbarium material could respond to changes in atmospheric composition caused by human activities, but the ecological consequences were unknown. Subsequent analysis of the δ13C composition of the same herbarium leaves (Woodward, 1993) indicated that leaf water use efficiency had been increasing through the same period, a response in concert with the reduced stomatal density. Since these observations there has been limited effort to investigate the contemporary ecological consequences of this stomatal response, with most effort involved with increasing the database of species’ responses.

One problem in comparing the responses of different species to a particular treatment is the different genetic backgrounds of the species, and the likelihood that differences in response may not be connected with the target response of stomatal development. Studies have been extended to investigate within species responses, in order to reduce some of the problems associated with interspecific studies. A wide range of accessions in the species Arabidopsis thaliana is available for determining its utility as a model plant. The stomatal responses of 48 accessions to CO2 enrichment broadly parallel the interspecific observations (Fig. 1), indicating the acceptability of Arabidopsis for this study.

The generality of the stomatal response to CO2, from short-term experiments to evidence from fossils, indicates a selective advantage to this response. This could occur through changes in water use efficiency (Woodward, 1993), but no matter what the actual mechanism it should be possible to demonstrate an effect on fitness. This was investigated by experiment, using two accessions of Arabidopsis, Col-0 and Ws. The accession Ws shows a much greater reduction in stomatal density with CO2 enrichment than Col-0 (Fig. 2). However the reduction for Col-0 is significantly increased (P < 0.05) under mild drought. The impacts of the treatments on fitness were assessed by measuring flowering (Fig. 3). Under well-watered conditions Col-0, with the smallest change in stomatal density, shows the greater enhancement in flowering with CO2. When droughted, flower production is enhanced under CO2 enrichment for Ws, but with no change for Col-0. These responses suggest that the reduction in stomatal density with CO2 has complex effects on fitness. Under well-watered conditions there appears a benefit from a small response, contrasting with a benefit from a large response under mild drought.

The ambient CO2 environment around a leaf and a plant may vary through the year and throughout the life of the plant (Bazzaz & Williams, 1991). This raises the question of what particular CO2 concentration influences leaf development. Is it the CO2 concentration around the developing leaf, or is it a CO2 concentration lower down the plant? Lake et al. (2001) demonstrated that the developing leaf does not have this capacity and that mature leaves control the stomatal developmental response in the developing leaf to both CO2 concentration and light. Therefore irrespective of the CO2 concentration or the irradiance around the developing leaf, it is the CO2 concentration and irradiance around the mature leaf that determines stomatal development, in terms of stomatal index (Fig. 5). However the response of stomatal density is rather different (Fig. 6). In this case stomatal density is significantly reduced by CO2 enrichment over either the developing leaf or the mature leaf. This reflects a direct stimulation by CO2 enrichment of epidermal cell and leaf expansion in the developing leaf.

The influence of mature leaves on stomatal initiation in the developing leaf suggests a more reliable method of monitoring the local environment – specifically both CO2 concentration and light. The developing leaf may not experience full irradiance or ambient CO2, being surrounded, at least in the early stages of significant stomatal initiation (Croxdale, 2000), by other developing leaves which will change the local environment. A similar response is seen by the responses of different accessions of Arabidopsis to high irradiance (Fig. 4) when, irrespective of the response of stomatal density to CO2 enrichment, in all cases the stomatal response to CO2 is inhibited. In all cases the mature and subsequently developed leaves developed with anthocyanin accumulation. The anthocyanin production may have a protective role against high irradiance (Merzlyak & Chivkunova, 2000) and a likely involvement with the jasmonate and lipoxygenase (LOX) pathways (Bate & Rothstein, 1998), which are involved in orchestrating plant responses to abiotic and biotic stresses (Ellis & Turner, 2001). The possibility of the connection between high light, anthocyanin production, and the jasmonate and LOX pathways was investigated with the fad-4 mutant of Arabidopsis. This mutant (Hugly et al., 1991) is unable to accumulate jasmonate because of a deficiency in the LOX pathway and therefore is unlikely to develop anthocyanin (Bate & Rothstein, 1998). Experiments with the fad-4 mutant (Fig. 7) indicated that the CO2 response of stomatal index was also inhibited. This indicates that it is unlikely to be anthocyanin, or jasmonate accumulation that inhibits the response. Research with other mutants of Arabidopsis aims to identify the site of this control.

The field studies with V. myrtillus (Figs 8 and 9) indicated that stomatal density increased with altitude and that this response could be induced reliably by transplanting plants to different altitudes. Controlled environment experiments indicated that the changes in temperature, wind speed and mean irradiance characteristic of this transect (Woodward, 1983; Woodward & Kelly, 1995; F. I. Woodward, unpublished) exerted little impact on stomatal density. However, the changes in CO2 concentration observed along the altitudinal gradient induced significant changes in stomatal density (Woodward, 1986). Both the prostrate upland and snowbed upland populations showed very similar responses of stomatal density to changes in CO2 concentration (Woodward, 1986).

The field response for plants growing in a community of other plants appears to be less than plants growing individually, and for both existing plants (Fig. 8) and transplants (Fig. 9). The topography of the two high altitude sites (Fig. 10) indicates that the lower stomatal density is found for plants in a snow bed community. The dwarf plants from the exposed position developed with higher stomatal densities (Fig. 10). This response can be explained on the basis of the signalling mechanism identified by Lake et al. (2001) in which the fully expanded leaves are in a lower irradiance in the snow bed community than in the exposed site. This would lead to a reduction in the stomatal density of the expanding leaves at the top of the canopy, assuming that the signalling response determined for Arabidopsis and Sinapis is also true for V. myrtillus. The ecological significance of this response is most likely to emerge through physiological responses. Woodward (1986) showed, over an altitudinal range from 200 to 1100 m, that the plants from the highest altitude, with the highest stomatal densities, also had the highest stomatal conductances and photosynthetic rates and leaf nitrogen concentrations. This suggests that signalling from mature to developing leaves may serve to maximise the photosynthetic potential of the developing leaf while it is in early stages of development in an enclosed bud. In addition it is expected that there will be metabolic crosstalk, where many signalling pathways interact to produce the final integrated response (Knight & Knight, 2001).

Figure 10.

Ground profile representing the small-scale topography of a late snow bed, with snow patch vegetation (depth of the depression c. 20 cm and plants drawn to scale) and an exposed area, at an altitude of 1100 m in Scotland. Profiles of photosynthetically active radiation (PAR), through the vegetation, are shown for the two areas.


We are grateful to the University of Sheffield for a postgraduate studentship to J.A.L.