Comparisons of WUE in twelve genotypes of winter wheat and the relationship between δ13C and WUE

Twelve winter wheat (Triticum aestivum) genotypes were examined for differences in grain yield, water use efficiency (WUE), and stable carbon isotope composition (δ13C) in flag leaves. The plants were subjected to rain-fed treatment and supplemental irrigation at the jointing and anthesis stages, during the 2015–2016 and 2016–2017 winter wheat growing seasons. The relationships between δ13C with grain yield and WUE were analyzed under two different water environments. The results indicated that there were significant differences in δ13C, grain yield, and WUE among wheat genotypes both under rain-fed and supplemental irrigation conditions. The δ13C values increased with grain-filling proceeding, the δ13C being lower under supplemental irrigation treatment than that under rain-fed treatment. The relationships between the average of δ13C with grain yield and WUE were significantly positive during three measurement periods (R2 = 0.5785 − 0.8258), whether under rain-fed or irrigation environments. This suggests that δ13C might be associated with the grain yield and WUE in winter wheat under rain-fed and supplemental irrigation conditions in the climate region of the northwest Huang-Huai-Hai Plain of China.


INTRODUCTION
More than half of the winter wheat in China is supplied by the Huang-Huai-Hai Plain of China, which covers a 32,000 km 2 area (Zhang et al., 2010). As the main winter wheat production area in China, the Huang-Huai-Hai Plain does not have consistent annual rainfall, with less than 200 mm of rainfall, but more than 400 mm of evapotranspiration during the winter wheat growing season (Li et al., 2012;Fulvia et al., 2012). Moreover, climate change is forecasted to further compromise the water quality and availability for supply (Sowers, Vengosh & Weinthal, 2011;Garcia & Pargament, 2015). The Gravity Recovery and Climate Experiment (GRACE) detected a significant groundwater storage (GWS) depletion rate of 7.2 ± km 3 /yr in the Huang-Huai-Hai Plain of China during 2002(Feng et al., 2018. New drought-resistant and water-saving cultivars are needed to ensure production capacity under climate change scenarios. Water use efficiency (WUE) represents the relationship between carbon assimilation and water used. Thus it is vital to increase WUE for sustaining wheat production with the increasing water scarcity and drought trends (Bacon, 2004).
There are several scales of WUE. The intrinsic WUE is the ratio of the photosynthesis rate divided by the stomatal conductance. The leaf WUE or leaf transpiration efficiency (TE) is the ratio of the photosynthesis rate divided by transpiration rate. The WUE at the individual plant level is the ratio of biomass or economic weight divided by water transpired. The field WUE is the economic product or biomass divided by water used, including transpiration and evaporation (Mei et al., 2013).
Carbon is an important life element, and there are two stable forms ( 13 C and 12 C) in nature. Due to the difference in physical and chemical properties of 13 C and 12 C, isotope fractionation occurs in plants during the process of assimilating CO 2 (Farquhar, Ehleringer & Hubick, 1989), and the stable isotope composition before and after the reaction are different. Carbon isotope composition (δ) and carbon isotope discrimination ( ) are parameters that are utilized to characterize carbon isotope discrimination in plants (Xu et al., 2007). Because of the large differences in δ 13 C among natural plants (Paneth & O'Leary, 1985), δ 13 C can integrate the long-term photosynthetic, various physiological, and morphological characteristics of plants. δ 13 C is widely used in ecology, agronomy, and global change research with the advantages of simplicity, rapidity, and accuracy. For it analyzes the carbon accumulated in the leaves over a long period of time, δ 13 C is then used to evaluate the characteristics of WUE in leaves and plants during growth (Damesin & Learge, 2003;Scartazza, Mata & Matteucci, 2004).
Over the past forty years, many researchers have found that plant-stable carbon isotope composition (δ 13 C) is rather useful. In recent years, there have been many studies that used δ 13 C in plant leaves to indicate WUE, and some important conclusions have been made. A large number of studies have shown that there is a significant correlation between δ 13 C value with crop yield and WUE, and the δ 13 C in plant leaves can be the indicator of plant long-term WUE (Farquhar & Richards, 1984;Hubick & Farquhar, 1987;Sarange et al., 1999;McDowell, Brooks & Fitzgerald, 2003). Studies have found that δ 13 C is heritable (Condon & Richards, 1992). Simultaneously, grain yield is positively associated with in grain at maturity and in leaves at anthesis (Xu et al., 2007). The stage of sampling, the analyzed organ or tissue, and the water regime can affect the correlation between and grain yield (Xu et al., 2007).
The highest negative correlations between δ 13 C in mature grains and grain yield across durum wheat genotypes were shown under rain-fed and irrigation conditions. While the flag leaf, δ 13 C was negatively correlated with grain yield only under rain-fed conditions (Bort et al., 2014). In addition, leaf WUE was affected by the ratio of CO 2 concentration in the leaves and CO 2 concentration in the atmosphere (Ci/Ca), while the δ 13 C values were closely related to Ci/Ca (Ehleringer, 1993). This suggested that there may be some relationships between δ 13 C value with the grain yield and WUE. Studies on grain WUE could be conducted to supply guidance for agricultural practices on the North China Plain (Mei et al., 2013). However, there are some arguments on the stability of relationships of 13 C in various tissues with yield and WUE, especially concerning crop WUE on the field scale (Anyia et al., 2007;Seibt et al., 2008;Yousfi, Serret & Araus, 2013). TE on individual plant scale and effective amounts of water used for production have no correlation with 13 C using isogenic materials with large variation in TE (Devi et al., 2011).
Furthermore, the selection criteria for grain yield and WUE are important aspects of water-saving agricultural research. A large number of studies have been carried out from the aspects of morphology, physiology, and molecular biology, aiming at finding out reliable identification indicators for selecting drought-resistant and water-saving crop genotypes. To date, however, there are not many rapid, simple screening methods that will assist mass selection in breeding programs.
Although field WUE provides the most meaningful guidance for agricultural practices, it is laborious and time consuming, thus not suitable for evaluation of a large number of varieties. Essentially, the data on the correlation between leaf δ 13 C concentrated on different grain-filling stages, grain yield, and WUE on a field scale are lacking. In our experiment, 12 different genotypes of winter wheat with different morphological and physiological characteristics were selected for study. The δ 13 C values in the flag leaves of different genotypes and their relationships with yield and WUE were studied under different water conditions with rain-fed and supplemental irrigation at the jointing and anthesis stages to provide a basis for screening drought-resistant and water-saving varieties.

Experimental site
The experiment was carried out at the Shunyi Experimental Station (20 m above sea level, 40 • N and 116 • E), located in the northern part of the Huang (Yellow River)-Huai (Huai River)-Hai (Hai River) Plain of China. The area is in a warm temperate semi-humid continental monsoon climate zone with most of annual rainfall falling during the summer season. In this region, the mean rainfall is 625 mm, of which approximately 75% falls between July and August, and the annual average temperature is 11.5 • C, with average annual sunshine of 2,750 h. The experiment was carried out in the 2015-2016 and 2016-2017 winter wheat growing seasons. Each experimental plot was 10.0 m ×2.4 m, and the plot consisted of 12 lines separated by 20 cm. The levels of organic matter, total nitrogen, rapidly available phosphorus, rapidly available potassium, and pH in a 0-20 cm soil layer were 14.4 g kg −1 , 1.1 g kg −1 , 24.5 mg kg −1 , 10 mg kg −1 , and 7.7, respectively; 112.5 kg ha −1 of nitrogen, 75 kg ha −1 of phosphorus pentoxide, and 75 kg ha −1 of potassium oxide were present, and the previously planted crop was field peas (Pisumsativum spp. arvense). The winter wheat was planted at a density of 6.75 × 10 6 plants ha −1 on October 5, 2015 and October 3, 2016. Winter wheat was harvested on June 12, 2016 and June 11, 2017. Eight near isogenic lines (NILs) along with their donor parent Jing 411 and recurrent parent Jinmai 47 and Shijiazhuang 8 and Yumai 18 were adopted as materials in this experiment. Twelve winter wheat genotypes have different drought-resistant and water-saving characteristics.
In the 2015-2016 and 2016-2017 winter wheat growing seasons, the precipitation totals were 173.4 and 119.6 mm, respectively ( Fig. 1), which were less than the average seasonal precipitation (242.6 mm ) by 69.2 and 123 mm, respectively. In the 2015-2016 winter wheat growing season, precipitation occurred mainly from April to June, accounting for 62.9% of the total annual precipitation. By contrast, during the 2016-2017 winter wheat growing season, precipitation occurred mainly from October to November, and the precipitation during this period accounted for 64.4% of the total annual precipitation.

Experimental design
Twelve winter wheat genotypes, which differ in population water use efficiency, were selected as materials, based on the previous studies regarding evaluation of water consumption traits. Rain-fed and supplemental irrigation conditions were set as two water treatments. For rain-fed treatments, no irrigation was conducted after sowing in the winter wheat growing season. For supplemental irrigation treatments, 60 mm water was applied to each plot at the jointing and flowering stages by sprinkling water with a flow meter to record the irrigated water amount. Before sowing, each plot of the two treatments was watered uniformly to field capacity to ensure that the seeds would sprout and grow into intact seedlings. Each experimental growing plot was 10m in length and 2m in width. The plots were arranged in randomized blocks with three replications. Soil moisture content was measured at the sowing stage and final harvest stage, grain yield was determined after harvest, and leaf samples were taken in early, middle, and late grain filling stages for measurement of δ 13 C.

Soil moisture content
The soil moisture content obtained at every 20 cm down to 160 cm of soil was measured by the oven-drying method (Li et al., 2015). Measurements were performed at the sowing stage and final harvest stage. The soil sample is placed immediately in an aluminum box after it is taken. The fresh weight is placed in a drying oven and dried at 105 • C for 24 h to constant weight. The dried soil is weighed and the soil moisture content is calculated. Soil moisture content=(fresh soil weight-dry soil weight)/dry soil weight ×100%.

Grain yield
When the winter wheat had attained maturity, 1-m lengths of 5 rows were manually harvested at random in every experimental plot and air dried.

Evapotranspiration
Evapotranspiration (ET) was calculated using the soil water balance equation (Li et al., 2012): where ET denotes evapotranspiration (mm), I denotes the irrigation amount (mm), P denotes precipitation (mm) that was measured by the weather station at the site, R denotes the surface runoff (mm) and runoff was not observed due to the low precipitation, D denotes the downward drainage below the crop root zone (mm), and it was negligible because the soil moisture measurements indicated that drainage at the experimental site was negligible, and therefore, deep percolation was ignored, and SW denotes the change in water storage in the soil profile exploited (0-160 cm in depth) from planting to harvest by crop roots (mm).

Water use efficiency
Water use efficiency (WUE) in relation to the grain yield is defined as follows (Sadras & Lawson, 2013): where Y (kg ha −1 ) denotes the winter wheat grain yield, and ET (mm) denotes the total evapotranspiration during the entire winter wheat growing season derived from Eq. (1).

Sample collection
Wheat leaf samples were taken on May 13, 22, and June 1, 2016, and May 11, 20, and 29, 2017, corresponding to early, middle, and late stage of wheat grain filling. 10-15 fully expanded flag leaves from 5 spots in each plot were taken and pooled. A portion of this pooled sample was then analyzed.

Measurement of stable carbon isotope composition (δ 13 C)
The samples were placed in paper envelopes and brought back to the laboratory, and then, they were dried to a constant weight at 70 • C, crushed, and sieved through 200 mesh. δ 13 C was determined using a combination of a Vario PYRO cube element analyzer and ISOPRIME-100 stable isotope mass spectrometer (Elemental Microanalysis, Okehampton, UK). The standard substance used for sample determination was a Low Organic Soil Standard, standard value : δ 13 Cv-PDB: −26.66, C%: 1.61. In the analysis process, every 12 samples were interspersed with a laboratory standard for calibration. The long-term

Notes.
Proportion can be achieved by dividing the sum of squares of each factor by the total sum of squares.
standard deviation of the instrument was 0.2 , and C is a calibration result based on the international standard V-PDB.

Statistical analysis
The collected data for all of the measurements were statistically analyzed with three-factor ANOVA to test the difference among wheat genotypes and between two water treatments, using SAS software (version 9.2; SAS Institute, Cary, NC, USA). The multiple comparisons were performed at α = 0.05 level of significance with LSD test. The Cluster Procedure (Single Linkage Cluster Analysis) has been used for factors. Correlation analysis was conducted to relate the δ 13 C with grain yield and WUE.

Comparison of WUE in 12 genotypes of winter wheat
There were significant differences in WUE in different wheat genotypes under the same water conditions and water condition accounting for 14.51% of the total variance together with two-factor and three-factor interaction between genotype and water condition had significant effect on WUE (Table 1). The method of ''Minimum Distance Between Clusters'' was used to generate Figs. 2 and 3. Under rain-fed condition, the materials were divided into four groups with different WUE (Fig. 2). There were five genotypes in group I, four genotypes in group II, two genotypes in group III, and just one genotype in group IV. The average value of WUE in group I was significantly higher than that in groups II, III, and IV. The genotypes in group I exhibited a higher grain yield and lower water consumption. The genotypes in group II exhibited moderate grain yield and moderate water consumption. The genotypes in groups III and IV exhibited lower grain yield and higher water consumption. Under supplemental irrigation conditions, the materials were divided into three groups with different WUE (Fig. 3). Both water consumption and grain yield in winter wheat genotypes were significantly different regardless of rain-fed or irrigation treatment (Tables 2-3). Water condition had the largest effect on grain yield and water consumption compared with other factors and interactions. The average grain yield difference between two water conditions was 1,592 kg ha −1 in 2015-2016, and 1,563 kg ha −1 in 2016-2017 growing seasons, respectively. However, genotypes differed largely in response to soil moisture alteration (Tables 4-5).
The largest increase in grain yield occurred in genotype '908032' with 1,925 kg ha −1 , and the smallest in genotype '908188' with 1,267 kg ha −1 in 2015-2016. The largest increase in grain yield was in genotype '908120' with 2,494 kg ha −1 , and the smallest in genotype '908274' with 835 kg ha −1 in 2016-2017 growing seasons. This implied different responses to supplemental irrigation applied at jointing and anthesis stages, which brought about ultimate different yield responses, among wheat genotypes. The variation of the WUE among 12 genotypes was basically consistent with the grain yield. Under irrigation treatment, the average WUE for wheat genotypes was higher than that under rain-fed conditions by 1.5 kg ha −1 mm −1 in 2015-2016 growing season, and by 1.2 kg ha −1 mm −1

Notes.
Proportion can be achieved by dividing the sum of squares of each factor by the total sum of squares.
in 2016-2017 growing season, respectively. Similarly, genotypes differed largely in WUE increase under irrigation treatment, compared with rain-fed condition. The largest increase was in genotype '908032' at 3.07 kg ha −1 mm −1 , and the smallest increase was in genotype '908188' at 0.13 kg ha −1 mm −1 in 2015-2016 growing season. The largest increase was in genotype '908188' at 2.43 kg ha −1 mm −1 and the smallest increase was in genotype 'Shijiazhuang 8' at 0.3 kg ha −1 mm −1 in 2016-2017 growing season. Also, with the same amount of water consumption, genotypes differed significantly in grain yield. These results comprehensively reflected the differences in response to supplemental irrigation during critical water demand stages among genotypes. Moreover, under rain-fed and supplemental irrigation conditions, the grain yields were significantly and linearly positively correlated with the WUE (R 2 = 0.8231 − 0.9375).

Stable carbon isotope composition (δ 13 C) in flag leaves
The 12 winter wheat genotypes exhibited significant differences in δ 13 C values of flag leaves during early, middle, and late grain filling stages, both under rain-fed and supplemental irrigation conditions (Tables 6-7). For the three measurement time points from early to late grain filling stage, δ 13 C of some genotypes did not change substantially, while some genotypes increased significantly (P < 0.001) under both water conditions (Tables 6-7). The order of 12 winter wheat genotypes in term of δ 13 C values remained nearly the same

Notes.
Values without the same letters in the same line are significantly different at 5% level; Difference here indicated the gap of grain yield and WUE between irrigation and rain-fed treatments.

Notes.
Values without the same letters in the same line are significantly different at 5% level; Difference here indicated the gap of grain yield and WUE between irrigation and rain-fed treatments.  at three measurements under the two water treatments, indicating the stability of δ 13 C in characterizing plants. The δ 13 C values of flag leaves were significantly higher under rain-fed conditions than that under supplemental irrigation conditions (Tables 6-7). ANOVA of δ 13 C showed that the factors investigation year (Y), genotype (G), stage (S), water condition (W), and block (B) all had significant effects on δ 13 C (Table 8). Variance in δ 13 C from Y was the largest, covering 43.94% of the total variance. G ranked the second to Y, occupying 25.64%. Proportion of variance from S, W, and B were rather smaller than that of Y and G. This indicated that Y was the most influential variable for δ 13 C, and G had important effects on δ 13 C. Moreover, multiple-factor interactions among Y, G, W, S, and B had significant effects on δ 13 C, except two-factor interactions between S × G, S × W, and S × Y, though the proportions of variance were rather smaller, compared with that from Y and G.

Relationship between the δ 13 C value, grain yield, and WUE
The reliability of using δ 13 C to characterize WUE and grain yield was examined by analyzing the correlation between δ 13 C values, grain yield, and WUE. There was a linear positive significant correlation between the δ 13 C values, the grain yield, and WUE during different grain-filling periods under rain-fed and supplemental irrigation conditions (Figs. 4-7). The δ 13 C values increased during early grain-filling stage to the late grain-filling stage (Tables 6-7). However, the changes in the δ 13 C values did not cause changes in the linear relationships between δ 13 C values, the grain yield, and WUE (Figs. 4-7). Whether under rain-fed conditions or supplemental irrigation conditions, the genotypes with high δ 13 C values exhibited higher grain yield and WUE. On the contrary, the grain yield and WUE in genotypes with lower δ 13 C values were relatively lower. This indicated that the δ 13 C value measurement in winter wheat can be performed at any time during the grain-filling period without affecting its correlation with WUE and grain yield.

DISCUSSION
In practical terms, WUE at the field population scale deserves the most attention rather than the WUE at leaf level and the individual or intrinsic WUE (Mei et al., 2013). As well known, however, evaluation of WUE at the field population scale is laborious, time consuming, thus impossible to evaluate a large number of materials. The present study explored how the easily measured trait δ 13 C correlated with field population WUE in 12 selected winter wheat genotypes under rain-fed and supplemental irrigation treatments. This might provide helpful information for finding reliable surrogate traits for WUE evaluation.

Effects of different factors on WUE and δ 13 C
The results indicated that there were significant differences in grain yield and WUE among winter wheat genotypes. Grain yield and WUE were higher under supplemental irrigation conditions in comparison with those under rain-fed conditions. These were consistent with previous researches (Li et al., 2014). The lower WUE of different genotypes under rain-fed conditions was mainly due to the fact that less water supplied during the experimental years and a severe degree of drought stress subsequently affected plant growth. Midgley & Moll (1993) conducted research on the cyclic drought stress in perennial shrubs in South Africa and pointed out that WUE increased with moderate water content of the soil, and when the soil water content was too low, stomatal conductance became conservative and eventually stabilized, and WUE began to decline again. Previous studies have also shown that there are significant differences in δ 13 C values in C 3 , C 4 , and CAM plants (Mook, Bommerson & Staverma, 1974;O'Leary, 1984;Farquhar, Ehleringer & Hubick, 1989). Due to the synergistic effect of the absorption and diffusion phases in the enzymatic reaction during CO 2 formation from carboxyl groups after isotope fractionation, the same photosynthetic pathway plants also exhibit the phenomenon of high degree of differentiation. The 12 winter wheat genotypes belonged to C 3 plants, and the δ 13 C values in flag leaves in winter wheat genotypes exhibited significant differences for 3 measurements. The above theories sufficiently explain the results of the present study. Our study also showed that δ 13 C values slowly increased during the three measurements, whether under rain-fed conditions or supplemental irrigation at the jointing and anthesis stages. After plants develop into grain filling stage, strong photosynthesis for rapid accumulation of assimilates began to recede gradually. Instead, distribution and conversion of assimilates to grain turn to predominate this stage. The weakened photosynthesis leads to increased Ci/Ca, which in turn result in reduction of CO 2 entering into mesophyll cells. At this time, the 13 C could not be fractionated in a timely fashion by the carboxylase of the plants, and then organic matter was directly synthesized with high δ 13 C values ( Van de Water, Leavitt & Betancourt, 1994;Feng & Epstein, 1995).
As shown by previous studies, water stress induce stomatal closure, then cause decrease in stomatal conductance and Ci, which in turn leads to an increase in the δ 13 C values of fixed carbon in plant photosynthesis. Conversely, a decrease in δ 13 C values comes out under more water and high nutrient conditions. The δ 13 C values for the different genotypes of wheat were higher under rain-fed conditions than that under corresponding irrigation conditions. The differences in the δ 13 C values among the genotypes in this experiment were greater than that between different measurement periods and different water treatments (Tables 6-8), which indirectly indicates that the δ 13 C values in flag leaves of winter wheat have relative stability. Dramatically, the factor investigation year had the largest effect on δ 13 C, differing from its influence on yield, water consumption, and WUE. δ 13 C was affected by intrinsic photosynthetic capacity, driven by the amount of carboxylating enzyme (Bort et al., 2014;Condon et al., 2002). Precipitation in the wheat grain filling stage of 2016-2017 growing season was larger than that in 2015-2016 growing season (Fig. 1). Plant photosynthesis sensitively responds to water status, as well known. Thus we would rather ascribe the substantial influence of year on δ 13 C to the precipitation difference in grain filling stage between the two investigation years.

Correlations between δ 13 C and WUE
The negative relationship between carbon isotope discrimination and WUE at leaf level for wheat genotypes was first proposed in 1984 (Farquhar & Richards, 1984). In the current study, the δ 13 C values in leaf material showed a significant linear positive correlation with WUE at field population level under the experimental conditions (Figs. 4-7). The correlation between δ 13 C values during the three determination periods at the grain-filling stages with grain yield and WUE reached an extremely significant level, and different water conditions had no effect on the correlation between the δ 13 C values with grain yield and WUE, which is consistent with a large number of previous studies (Wright, Rao & Farquhar, 1994;Condon, Richards & Rebetzke, 2004). The data implied that comparisons might be carried out at any time during the grain-filling stage under various moisture conditions to indicate WUE in winter wheat with δ 13 C values as an indicator.
The theoretical basis of the linear positive correlation between the δ 13 C values and leaf WUE can be obtained from the calculation formula of δ 13 C values and WUE derived from Farquhar, O'Leary & Berry (1982) and Condon, Richards & Rebetzke (2004). Their researches indicate that plant δ 13 C values and WUE have the same influencing factor of Ci/Ca. Therefore, WUE can be deduced by measuring the plant δ 13 C values. However, there are also reports that the δ 13 C values are negatively correlated or not correlated with yield and WUE (Delucia & Schlesinger, 1991;Fengjun et al., 2006;Chamekh et al., 2016). This might be ascribed to the differences in water conditions and growth environments. It can be seen that the degree of water stress may have an effect on the correlation between δ 13 C values with grain yield and WUE, although the specific reasons require further study. Therefore, it is necessary to define water conditions clearly for using δ 13 C value as surrogate trait in screening water-saving crop varieties.
In addition to the effect of water conditions discussed above, the correlation between δ 13 C and WUE seems to be related with crop species and temporal-spatial scale of WUE. Transpiration efficiency (TE), different from field WUE in scale, is biomass accumulation divided by water consumed through transpiration of individual plant in a period of time. In legume plants, TE was reported to negatively correlate or have no correlation with δ 13 C. A research conducted in the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) found TE was negatively associated with 13 C in groundnut leaves at the final harvest after drought stress imposition, but fairly weak (R 2 = 0.13) (Krishnamurthy et al., 2007). In lentil (Lens culinaris Medikus), chickpea (Cicer arietinum L.), and lupin (Lupinus angustifolius L.) 3 grain legumes, leaf δ 13 C and TE in the vegetative phase were not significantly correlated under well-watered conditions (Turner et al., 2007). Also, no correlation between TE and leaf 13 C in groundnut isogenic materials with large variation in TE was found regardless of the water regime (Devi et al., 2011). In wheat, however, positive correlations between 13 C and field WUE or yield are apt to be reported more by the researches. Leaf 13 C at anthesis positively correlated with grain yield for spring wheat in the post-anthesis water stress conditions of northwest Mexico. Both leaf 13 C and grain 13 C were strongly and positively correlated to grain yield under rain-fed conditions of northern China (Xu et al., 2007). These are in agreement with our results in winter wheat. Therefore, to elucidate and then efficiently use the relations between δ 13 C and WUE in screening water saving varieties, crop species and WUE scale need to be fully considered.
In general, though the accumulating studies proved high stability of δ 13 C characterizing plants, the relationship between δ 13 C value and WUE were determined by diverse environmental factors, crops species, along with WUE scale. Therefore, in order to improve the reliability and scientific accuracy of the δ 13 C value as an indicator to assist with breeding water-saving genotypes, further studies need to be performed to elucidate some definite relations under specific environmental conditions.

CONCLUSIONS
Leaf δ 13 C values at grain-filing stage significantly correlated with WUE of field population level in both rain-fed and irrigation status under the climate conditions of the northwest Huang-Huai-Hai Plain, as proved by our experiments conducted in two consecutive growing seasons. This may imply that leaf δ 13 C of grain filling stage might be a potential useful surrogate trait for field population WUE, which is of practical guidance but of difficulty in evaluation. Referring to the previous relative studies, the relations between leaf δ 13 C and WUE are considered to depend on multiple factors, such as soil water status, climate conditions, crop species, etc. Therefore, to improve the reliability of δ 13 C value as an indicator to assist with breeding and selecting water-saving cultivars, further studies need to be performed. Enlarged number of wheat materials, experimental sites under different climate, and different soil water regimes all remain to be taken account. To conclude, for obtaining definite correlations between δ 13 C and population WUE, it is absolutely essential to specify environmental conditions by conducting further systematic researches.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
This work was supported by research grants from the National Key Research and Development Program (2017YFD0201702) and the Innovation Program of Chinese Academy of Agricultural Sciences (CAAS-XTCX2016019). The funders had no role in