Previous PageTable Of ContentsNext Page

Measurement of the depth of water extraction by individual plants in a plant nursery, with 18O.

Jean-Louis Durand1, Thierry Bariac2, Mia Victoria Gonzalez Dugo1, Marc Ghesquire3, Patricia Richard2 and Philippe Biron2.

1 Unit d’Ecophysiologie des Plantes Fourragres Email
Unit BIOMCO Email
Unit de Gntique et Amlioration des Plantes Fourragres Email


Screening in plant breeding nurseries for depth of water extraction or rooting depth is hardly possible using standard techniques of soil moisture measurements. A new methodology based on the discrimination of 18O in soil water is proposed. It consists in measuring the natural gradient in 18O which generates in soil following several days of evaporation, and the δ18O of the water extracted from the plant. The study compares the actual plant δ18O and the value computed combining neutron probe and soil δ18O measurements. The study shows that, provided the plant sample is appropriate, there is consistency between the plant δ18O and the soil water δ18O actually extracted. Tested here on forage grasses, the method can be easily used for any crop.

Media summary

Natural 18O discrimination by evaporation could provide a new method enabling plant breeders to screen for depth of water extraction in large populations.


Natural isotope abundance, drought resistance, rooting depth, grasses, plant breeding, festulolium.


A European breeding program focuses on the hybridisation between the drought resistant grass forage species Festuca arundinaceae ssp glaucescens and the high quality, highly productive species Lolium multiflorum (Ghesquire et al. 2002). The main trait associated with drought resistance in tall fescue was found to be rooting depth (Durand et al. 1997). However, screening for depth of water extraction (DWE) under nursery conditions on a large number of plants is difficult. Plants growing in such conditions take the transpired water from different horizons according to rooting density. In addition to root depth and density, axial and radial hydraulic conductance and root diameter are only some of the parameters which play an important role in the complex soil-root interactions. A direct estimate of water extraction is therefore a prerequisite for screening elite genotypes. Direct measurement of soil water uptake requires a sequence of two measurements of water content during a dry period. This would require access tubes or TDR probes on each individual.

The methodology proposed here relies on the sampling of plant organs for later analysis of oxygen isotopic composition (δ18O). Under dry conditions due to natural discrimination of evaporation against heavy isotopes in water and to the diffusion in soil, a gradient in soil δ18O is generated. Recent studies showed that the xylem water isotopic composition and the soil water composition were related. The use of the same methodology under nursery conditions could, in principle, give a ranking of the depths at which different genotypes take water. To test this, it was decided to compare the actual plant δ18O to a theoretical δ18O derived from both soil water isotopic composition and soil water extraction. The objectives of the work reported here were hence (i) to determine which plant's compartment reflected the total water extracted and (ii) to compare the depth of water extracted to indirect measurements.


Nursery conditions experiment.

If the proportion of water extracted at different depth z (ez) is known, and if the isotopic composition (18O) of the soil water at each depth δ (z) is known, the isotopic composition of the water collected by the plant plant) should be


δplant cannot be the mean δ of the plant bulk water as large variations due to the evaporation occurring in leaves induce high enrichment in 18O. Strictly, the measurements of all variables considered in the equation (1) should be done at the same time. This is hardly conceivable in natural conditions but one can expect the stability of distribution of water extraction in soil to be large enough to use soil moisture variations over a period of one or two weeks, depending on water flux. Genotypes were chosen among parent species Lolium multiflorum (Lm) (cv Fastyl) and of Festuca glauscescens (Fg) and their intergeneric hybrids recombined with Lm. Seeds were germinated in petri dishes and plants were transplanted later in the nursery at the beginning of autumn 2001, 50 cm apart in a greenhouse. The soil of the nursery was a homogenous silt loam. Plants were fully irrigated until 24 April 2002 and received no more water until 24 July.

Access tubes for neutron probe measurements were set as close as possible to the centre of the turf of ten individual plants. Neutron probe measurements were made on 24 April, 3 May and 10 June. Readings were made every 15-cm from the surface down to 100 cm.

On 10 June, 3 soil cores were sampled between plants down to 100 cm, every 2-cm between 0 and 15 cm depth, and every 10 cm below, in the vicinity of the individual plants measured. Tillers from the plants equipped with an access tube were sampled on 11 June between 11h00 and 12h00 TU on 11 June. 5-cm long tiller segments were cut from the ground. The organs in these samples consisted of the base of sheath of mature leaves and the base of growing leaves. All dead material and the sheath of the oldest leaf were discarded, in order to remove any non-active organ which could be transpiring. Water extraction and 18O analysis were made following the procedure described in Bariac et al. (1994).

Controlled environment experiment

In order to determine the compartment of the plant, the water of which represents the water collected from the whole rhizosphere, five Lolium and five Festuca plants were grown in the laboratory in hydroponics under one 400 watt HQI lamp with a photoperiod of 14 hours at room temperature. Tillers were collected from different plants before and 10 hours following turning on the lamp. Tillers were separated between lamina, outer sheath and the rest of the tiller, which consisted of a 5-cm long stubble containing sheaths of older leaves and bases of growing leaves. The water was extracted from the plant as in Bariac et al. (1994). The δ18O of the extracted water of the different plant compartments and of the nutrient solution were analysed.

In a second experiment under similar conditions, the roots of each plant were split between two containers, where they were allowed to grow for two weeks before sampling. One nutrient solution was at 0.1 MPa osmotic pressure, the other maintained at 0.6 MPa using PEG, roots of each of 8 plants per species growing in both containers (split roots system). The water in the high π container was enriched in 18O (δ = + 15 o/oo) to assess the possible implication of the species in the partitioning of water. The two species were grown in the same containers. The experiment was repeated 4 times with the same results.


Controlled environment

Table 1: Isotopic composition (δ 18O) of the water extracted from different compartments from tillers sampled on Lolium and Festuca plants grown under controlled environment. Each number is in o/oo and is the mean of two measurements (maximum coefficient of variation was 6 % under high irradiance and for lamina). Each sample consisted of a group of 4 tillers.

Transpiration regime


Lolium multiflorum

Festuca arundinacea

High irradiance




outer sheath






nutrient solution



Low irradiance




outer sheath






nutrient solution



The isotopic composition of the 5 cm stubble without the outer sheath was the same as for the nutrient solution, both under high and low transpiration regime (Table 1). Outer sheaths were approximately 0.5 - 1 /oo higher than the nutrient solution. The lamina exhibited much higher values than the nutrient solution, especially under high irradiance. This was due to the enrichment of the leaf water as evaporation increased. The difference between species was approximately 1 and 2 /oo under low and high irradiance, respectively, suggesting a higher transpiration rate in Festuca than in Lolium.

In the split root experiment, the proportion of water transpired coming from the container where the osmotic potential of the solution was raised to 0.6 MPa with PEG did not significantly depend on the species (Table 2). Nor did it depend on the quantity of roots in each container. Among different plants, the root proportion varied between 21 and 75 %, whereas the contribution only varied between 15 and 23 %. Hence, the water potential of the nutrient solution entirely controlled its contribution to transpiration stream, in accordance with the classical electric analogy of the soil plant atmosphere continuum. This also indicated that the δ18O of the water in plant stubbles reflected the origin of water, even under heterogeneous conditions for roots, irrespective of species.

Table2: Proportion of roots in- and water extracted from the container with nutrient solution enriched in PEG at 0.6 MPa ospotic pressure for the two species studied during the split root experiment. Mean values +/- standard error.


% of water transpired
coming from the high π container

% roots in the high π container

Festuca arundinacea

17 +/- 2.3 %

54 +/- 6.6 %

Lolium multiflorum

20 +/- 1.0 %

62 + 3.9 %

Nursery conditions.

Among genotypes, the average DWE measured with the neutron probe varied between 17 and 35 cm between 24 April and 3 May, and between 15 and 60 cm between 3 May and 10 June. On both dates, a gradient for water isotopic composition was found in the soil (Fig1). δ18O varied between 3 and – 6 o/oo, the latter value being the value of regional rainfall water. The gradient began at approximately 30 and 50 cm on 3 May and 10 June, respectively. The difference between the two dates came from the time elapsed between the last watering and measurement. In the top 3 cm on 10 June, the δ18O increased with depth until a cusp like point between 2 and 3 cm. In this region, the δ18O was directly influenced by the vapour phase, which is relatively poorer in 18O than the liquid water. This indicated the depth of the evaporation surface in the soil at that time. It is also clear that the gradient was smaller on 10 June, due to the effect, with time, of diffusion of water in the soil towards the drying soil surface. Presumably, the depth of liquid evaporating water being 3 cm below the ground, evaporation was greatly reduced.

Figure 1: a) soil water isotopic composition (δ18O) from soil cores sampled under bare soil in the plant nursery. On each date, the difference between two measurements at the same depth was less than 0.05 o/oo. Open symbols 3May; closed symbols: 10 June. b) Diagram of the disposition of soil sampling and moisture measurement.

The δ18O of the water extracted from the stubbles varied between –3 and –4.2 o/oo and between –4.2 and –5.2 o/oo on 4 May and 11 June, respectively. Combining the profiles of water extraction (neutron probe) and the profiles of soil δ18O, the theoretical δ18O of stubble was computed using equation (1). The comparison with observed values revealed a high consistency between both assessments (Fig 2) when plotted over both dates, but the correspondence was less consistent when each date was considered separately. Indeed, a plant exhibiting a δ18O of -4 o/oo had a DWE of 27 and 50 cm on 3 May and 11 June, respectively.

To our knowledge, this was the first direct use of the DWE (neutron probe) to predict the isotopic signature of plant water (Dawson et al. 2002). Although the water extraction was measured over at least 10 days, and soil δ was measured between plants, the stubble δ18O matched the predicted results. This suggests that instantaneous values for δ reflect the average depth of water extraction over a longer period of time. It also suggests that although the soil δ might differ between the area just below the plant and between plants in the top horizons, it does not alter the conclusions derived from measurements made in the latter region.

Figure 2: Comparison between the computed plant δ18O using soil moisture and soil water δ18O profiles to the measured plant water δ18O. Each point represents one plant associated to one neutron probe access tube. open sympols: 3 May; closed symbols 10 June.


Can the ranking of individual plants in a nursery according to their DWE be made using the isotopic composition of the water in their non-transpiring aerial organs? This requires certain condition to be met. (i) The organ from which the water is extracted for analysis must not transpire and should contain the bulk of water flowing through the plant. (ii) The soil δ18O gradient must be large enough to allow for significant range of δ18O values. (iii) The soil gradient must extend over a distance which is deep enough to test difference between species. Within the Lolium multiforum and their hybrids, it was possible to distinguish such differences, but more work is required to develop this technique as a tool for screening plants adapted to dry conditions.


This work was partly supported by the European Union within the project SAGES. V Gonzalez Dugo has a grant from the Region Poitou-Charentes.


Bariac T, Gonzalez-Dunia J, Katerji N, Bthenod O, Bertolini J.M. and Mariotti, A, (1994).- Variabilit spatio-temporelle de la composition isotopique de l'eau dans le continuum sol - plante - atmosphre : approche en conditions naturelles. Chemical Geology 115, 317-333.

Dawson TE, Mambelli S, Plamboeck AH, Templer PH, and Tu, KP. (2002). Stable isotopes in plant ecology. Annual Review of Ecological Systems 33, 507-559.

Durand JL, Gastal F, Etchebest S, Bonnet AC and Ghesquire, M (1997). Interspecific variability of plant water status and leaf morphogenesis in temperate forage grasses under summer water deficit. European Journal of Agronomy, 7, 99-107.

Ghesquire M, Barre P, Bguier V, Durand JL, Garnier S, and Laurent, V (2002). Interspecific hybridization in the fescue-ryegrass complex for better sustainability of European grasslands. In Durand JL, Emile JC Huyghes C and Lemaire G editors, Grassland Science in Europe 7, 298-299.

Previous PageTop Of PageNext Page