CASE STUDY 15.3 More plant for less water?

Printer-friendly version

Paul Kriedemann and Jay Anderson



Figure 1 Ms Vikki Fischer of CSIRO Plant Industry measuring stomatal conductance on different wheat genotypes. (Photograph courtesy 'biologic', February 1998)

 Vascular plants (Figure 1) typically transpire between 100 and 1000 g of water per gram dry mass formed (less under humid conditions, more under dry conditions). This large cost in water for such a meagre amount of carbon fixation is a direct consequence of a huge difference in the water vapour partial pressure inside leaves compared with their ambient atmosphere (a difference of 3500 µPa on a hot dry day; see also leaf versus air water potential in Section 15.2.4).

During transpiration, water molecules move from moist substomatal cavities to a much drier neighbouring atmosphere. By contrast, CO2 molecules diffusing into leaves have further to go (Figure 1.5) and are responding to a much more modest difference in partial pressure, typically 35–36 µPa outside (late 1990s level) compared with 24.5–25.2 µPa inside. The inward flux of CO2 molecules commonly reaches 20–30 µmol m–2 s–1 in a well-nourished C3 leaf, whereas the cor-responding outward flux of water molecules will be about 2000–3000 µmol m–2 s–1!

CO2 molecules entering leaves are thus buffeted by a much stronger efflux of water molecules. Moreover, during this inward diffusion from outside air to fixation sites inside chloroplasts, a subtle fractionation occurs between two of the naturally occurring forms of CO2. Heavier 13CO2 lags behind the slightly lighter 12CO2, so that the 13C/12C ratio in fixation products is slightly less than that of air.

CO2 occurs naturally as two stable isotopic forms, 12CO2 and 13CO2, where the superscripts 12 and 13 refer to the atomic weight of carbon making up a particular molecule. These two molecules have identical chemical properties but slightly different molecular weights. CO2 molecules con-taining 18O versus 16O also fractionate, but are not considered here.

In nature, 12CO2 is much more abundant (98.89% of outside air), whereas 13CO2 is rare by comparison, and represents only 1.11%. A minor fractionation (4.4 per thousand, i.e. 4.4‰) occurs during diffusion through stomata and intercellular spaces; a more substantial fractionation (29‰) occurs during biochemical fixation within chloroplasts.

Photosynthesising leaves in effect discriminate against 13CO2, so that 12C is fixed more readily than its slightly heavier counterpart 13C. The relative abundance of these two carbon isotopes in dried plant material can be measured by a ratio mass spectrometer with remarkable accuracy. That measurement then provides a ‘carbon isotope signature’ for a given material, represented here by Δ. Pioneering studies in the 1980s (see Farquhar et al. 1982), established that:

Δ = a ((papi)/pa) + b (pi/pa) (1)

and simplifying,

Δ = a + (ba) pi/pa (2)

where a represents fractionation in air (c. 4.4‰), and b represents net fractionation caused by carboxylation (c. 29‰) (see Case Study 1.1 for underlying theory on inward diffusion of CO2 and subsequent fixation by chloroplasts).

According to this simplified version (Equation 2), the extent of discrimination against 13CO2 during leaf gas exchange varies according to total CO2 partial pressure (12CO2 + 13CO2) inside leaves (pi) compared with that out-side leaves (pa). Higher CO2 partial pressure inside leaves permits greater discrimination, hence a numerically larger Δ, while lower CO2 partial pressure inside leaves restricts dis-crimination against the heavier isotope, hence a numerically smaller Δ.

When the carbon isotope signature of dried plant material is determined by mass spectrometry, a larger value for Δ implies that a higher pi/pa ratio prevailed at the time that carbon was being fixed, whereas smaller values imply that a lower pi/pa ratio must have prevailed. This ratio of pi/pa provides a link between Δ and water use efficiency during photosynthesis.

For a given stomatal conductance, inward diffusion of CO2 will be enhanced by a low pi/pa ratio, whereas transpira-tion will remain unaffected. Consequently, water use efficiency (expressed as moles of CO2 fixed per mole water vapour transpired) will be greater when pi/pa is lower.

Having validated this principle on barley and peanut (Farquhar et al. 1989), Condon et al. (1990) went on to assemble data from 16 wheat genotypes to show that transpiration efficiency of both well-watered and droughted plants was significantly correlated with Δ (Figure 2).



Figure 2 A highly significant relationship exists between transpiration efficiency (above-ground dry matter accumulated per unit mass of water transpired) and carbon isotope signature (DELTA) of plant material so formed. These 16 genotypes, including 14 hexaploid wheats and 2 tetraploid wheats, were grown in a green-house under either (a) we1l-watered conditions or (b) gradually increasing terminal drought stress. (Based on Condon et al. 1990)



Figure 3 A notional representation of stomatal and photosynthetic influences on pi/pa and hence carbon isotope signature (Δ) of photoassimilate formed. Either low photosynthesis (weak A, where A = assimilation) or high stomatal conductance (large gs, where gs = stomatal conductance) would contribute to a high pi/pa ratio during gas exchange. A high pi/pa ratio would result in a numerically larger value for Δ of fixed carbon. By contrast either high photosynthesis (strong A) or low stomatal conductance (small gs) would predispose to low pi/pa and hence a numerically smaller value for Δ. (Original unpublished drawing courtesy PE. Kriedemann)



Figure 4 Contrasting relationships between carbon isotope discrimination (Δ) and grain yield. Different wheat lines (derived from single F2 plants of one cross) were grown in test plots over two consecutive years. Yields in both years were favourable, and above the long-term average for that region. The drier year (1992) returned a negative relationship, whereas the wetter year (1993) returned a positive relationship (Based on original observations by A.G. Condon and R.A. Richards and cited in Richards 1996; reproduced with permission of Kluwer Academic Publishers).

During leaf gas exchange, pi/pa stabilises at a set point that varies according to environmental inputs, but is also subject to internal controls and especially genetic factors. Indeed, strong genetic variation in Δ has been established unequivocally (Condon et al. 1993) and with a broad sense heritability of around 60–90% in wheat, 80% in peanut and 71% in bean (broad sense heritability is derived from statistical analysis of population data where variation due to genetic factors is expressed as a percentage of total variation which includes environmental factors plus gene × environment interactions).

Large values for broad sense heritability imply in this case that a relatively small number of constitutive genes are involved, and that significant genetic gain in water use efficiency can be derived by selecting for low Δ lines. How-ever, any gene that affects either assimilation or stomatal conductance will influence pi/pa and hence Δ. Low Δ lines might just as well be an outcome of consistently low stomatal conductance as of fast assimilation. The relationship between Δ and pi/pa during leaf gas exchange is robust, and can be taken as a reliable indicator of pi/pa during leaf gas exchange, but in surveying populations for potential yield based on Δ, a distinction has to be made between stomatal and photo-synthetic factors as predominant sources of variation in Δ (Figure 3).

Contrasts in yield versus Δ relationships for dryland cereals (Figure 4) can now be addressed via this concept. In one year (1992), a negative relationship between Δ and yield indicated that a potentially higher water use efficiency was indeed associated with higher yield, and consistent with expectation. In another year (1993), a positive relationship emerged, implying that lines of lower water use efficiency achieved higher yields, or at least those lines where pi/pa was generally higher during the growing season returned higher yield.

This apparent anomaly can be resolved in terms of stomatal versus assimilatory influences. In those seasons where soil moisture was abundant, and replenished by rain at strategic points in vegetative growth and reproductive development (1993 in Figure 4), profligate water use oc-casioned by constantly high stomatal conductance would favour yield. This would be reflected in a positive correlation between Δ and yield. By contrast, in seasons where soil moisture was limiting gas exchange at critical stages (1992 in Figure 4), efficient use of water would carry more important implications for yield, and those lines with an intrinsically high water use efficiency would be superior, hence a negative correlation.



Figure 5 Carbon isotope signature (Δ) of young leaves (rapidly expanding and importing photoassimilate) was taken as indicative of current discrimination against 13C. Values collected from 11 species of eucalypt growing in a common garden at Wagga Wagga (part of the CSIRO Wagga Effluent Plantation Project) were negatively related to evaporative conditions during peak growing season in their original habitats. Species are numbered right to left (from arid to mesic environments): 1, Eucalyptus camaldulensis (Lake Albacutya provenance); 2, E. dives; 3, E, melliodora; 4, E. maculata; 5, E. pilularis; 6, E. botryoides; 7, E. saligna; 8, E. dunnii; 9, E. paliformis; 10, E. maidenii; 11, E. grandis (Coffs Harbour provenance); and 12, E. grandis (Woondum provenance) (Based on Anderson et al. 1996)

One further practical outcome of our conceptual model for leaf gas exchange (Figure 3) is that surveys of populations for intrinsic differences in water use efficiency need to be conducted in an environment where differences in pi/pa, and hence Δ, can be attributed to expression of underlying (genetic) differences in photosynthetic activity rather than seasonal responses in stomatal conductance. Well-nourished plants in a high humidity should behave this way because stomatal conductance would be uniformly high so that variation in Δ will be largely photosynthetic in origin. This projected outcome enabled Anderson et al. (1996) to link carbon isotope signature of eucalypt species to native habitat (Figure 5). A species trial at Wagga Wagga (New South Wales) with high inputs of water and nutrients was surveyed for Δ. Those species which had evolved under conditions of high evaporative demand during their growing season (represented in Figure 5 by the January potential evaporation of their current native habitats) returned a numerically lower value for Δ compared to those from more mesic habitats. By implication, high water use efficiency must have carried a selective advantage for evolution on drier sites.


Anderson, J.E., Williams, J., Kriedemann, P.E., Austin, M.P. and Farquhar, G.D. (1996). ‘Correlations between carbon isotope discrimination and climate of native habitats for diverse eucalypt taxa growing in a common garden’, Australian Journal of Plant Physiology, 23, 311–320.

Condon, A.G., Farquhar, G.D. and Richards, R.A. (1990). ‘Genotypic variation in carbon isotope discrimination and transpiration efficiency in wheat. Leaf gas exchange and whole plant studies’, Australian Journal of Plant Physiology, 17, 9–22.

Condon, A.G., Richards, R.A. and Farquhar, G.D. (1993). ‘Relationships between carbon isotope discrimination, water use efficiency and transpirational efficiency for dryland wheat’, Australian Journal of Agricultural Research, 44, 1693–1711.

Farquhar, G.D., Ehleringer, J.R. and Hubick, K.T. (1989). ‘Carbon isotope discrimination and photosynthesis’, Annual Review of Plant Physiology and Plant Molecular Biology, 40, 503–537.

Farquhar, G.D., O’Leary, M.H. and Berry, J.A. (1982). ‘On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves’, Australian Journal of Plant Physiology, 9, 121–137.

Richards, R.A. (1996). ‘Defining selection criteria to improve yield under drought’, Plant Growth Regulation, 20, 157–166.