Published in Crop Sci 39:1792-1797 (1999)
© 1999 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Crop Science 39:1792-1797 (1999)
© 1999 Crop Science Society of America
CROP ECOLOGY, PRODUCTION & MANAGEMENT
Assessment of Drought Resistance of Crop Genotypes by Means of the Water Potential Index
A.J. Karamanosa and
A.Y. Papatheoharia
a Lab. of Crop Production, Agricultural Univ. of Athens 75, Iera Odos, 11855 Athens, Greece
akaram{at}auadec.aua.gr
 |
ABSTRACT
|
|---|
The evaluation of crop genotypes for their adaptability is often performed by means of regression techniques of yields against some kind of an environmental index as independent variable. The weakness of these techniques lies in the lack of a direct assessment of a given environment by specific environmental factors. In this work, a new index, the water potential index (WPI), is suggested as a measure of the total water stress experienced by any crop in a given environment for a specific time interval. The index is derived from the integral of the course of leaf water potential over time. Its usefulness is demonstrated in evaluations of field-grown bread wheat [Triticum aestivum (L.) em. Thell] and faba bean (Vicia faba L.) genotypes of contrasting characteristics by yield vs. WPI linear regression analysis. In these regressions, the intercept represents the "potential" yield at no stress conditions and the slope represents the "adaptability" of each genotype. It was found that the dwarf wheat cultivars Yecora and Siette Cerros exhibited higher potential yields and lower adaptability in comparison with the tall cultivar Generoso. In faba beans, the smaller the seed size of a cultivar, the lower was the potential yield and the better the adaptability. On the basis of this regression analysis, different response scenarios of hypothetical crop genotypes are described. Furthermore, the "relative adaptability" concept is presented, in which the potential yield effect is removed from the adaptability of any genotype. Examples of the use of WPI for obtaining information on the involvement of water shortage in specific growth stages of the wheat crop or for assessing the sensitivity of faba bean yield components to water shortage are presented.
Abbreviations: WPI, water potential index • WPD, water potential duration • CWSI, crop water stress index • 
, plant water potential • l, leaf water potential • a, regression intercept • b, regression slope • bN, relative adaptability to drought
|
INTRODUCTION
|
|---|
THE IDENTIFICATION
of one or more physiological parameters as indices of drought resistance has been the subject of many research works (for a review, see Karamanos, 1984). Despite the sound physiological base and the usefulness of most of these parameters as water stress indicators, only in a few cases were they associated with an indication that the examined genotypes were also satisfactorily productive under drought (Turner, 1997). Some attempts to devise simple tissue tests for drought resistance (e.g., Sullivan and Ross, 1979) were not successful because of the multiplicity of factors and factor interactions contributing to drought resistance of crops in the field. From the agronomic point of view, the ways in which crop plants respond to drought cannot be evaluated without assessing their impact on the productivity of the stands. For example, a reduction of water loss by means either of leaf shedding or prolonged stomatal closure may be undesirable since these two responses also reduce dry matter production.
The difficulty in suggesting a certain physiological parameter as a reliable indicator for satisfactory productivity under dry conditions led the plant breeders towards using only yield performance over a range of environments as the main indicator for drought resistance. Thus, a number of regression techniques of yields against some kind of an environmental index as independent variable have been developed to evaluate genotype adaptability. Finlay and Wilkinson (1963) and Eberhart and Russell (1966) used expressions of the mean yield of all the examined genotypes at a given site as a measure of the local environment. Fischer and Maurer (1978) proposed the "drought susceptibility index" (yield of a genotype under drought as a function of the yield without drought), whereas Lin and Binns (1988) used the "superiority index" (the mean square of the distance of the yield of a genotype from the maximum yield of all genotypes at a given location) as estimates of genotype adaptability over a range of environments.
The weakness of the regression techniques mentioned above lies in the lack of a direct quantification of a given environment by specific environmental factors (Eberhart and Russell, 1966). This point is important when specific environmental stresses (e.g., drought, temperature) are to be considered in genotype evaluation trials. In this case, the regression techniques and all indices based on yield give only a bulk estimate of the combined effects of all environmental factors with no possibility of evaluating them separately. As an estimate of the water stress experienced by crop plants, Idso et al. (1981a) suggested a "crop water stress index" (CWSI) derived from the increase in average canopy temperature in relation to that of a well-watered reference plot and evaluated by infrared thermometry. The CWSI was linearly related to available soil water for a number of crops (Jackson et al., 1981; Nielsen and Anderson, 1989) and, thus, it could be useful for irrigation scheduling (for a review, see Gardner et al., 1992). However, its relation with plant water potential () appears to depart from linearity at high degrees of water stress (Idso et al., 1981b).
Because is an adequate expression of plant water balance at any time, it could be a useful and objective indicator of the intensity of water stress in genotype evaluation trials. By taking regular measurements, either of its highest (pre-dawn) or its lowest (at midday) diurnal value throughout the growing season, we form an integrated view of the water stress history experienced by a given plant or a crop. The integral of the course of over time (i.e., the shaded area in Fig. 1)
describes the "duration" of the water potential (WPD, in MPa days) for a given period:
 | (1) |
where t is the water potential at Day t within the observation period, i.e., from Days 1 to v.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 1 The time course of leaf water potential () for the wheat cultivar Yecora grown under rainfed conditions during the 1984-1985 season. The shaded area represents the water potential duration (WPD), and the vertical bars the standard errors of the means
|
|
Taking into account that the periods under consideration may differ in length for various reasons (e.g., differences in the duration either of the biological cycle or of the period of measurements), the values of the WPD can be made comparable among different cases by dividing them by the length of the period of study. The value so obtained can be called "water potential index" (WPI, in MPa) of a plant or crop:
 | (2) |
where n is the length of a period in days.
WPI can be used as an objective physiological index when plant genotypes are assessed for their drought resistance. In this work, regression analyses of final yields against the WPI will be undertaken for different genotypes of two crop plants, bread wheat and faba beans, in an effort to evaluate its usefulness as an environmental index.
|
Materials and methods
|
|---|
The results come from two series of experiments conducted in the field of the Agricultural University of Athens, in which the responses to water shortage of bread wheat and faba bean cultivars of contrasting types were studied. In the first series, three wheat cultivars, two semidwarf of Mexican origin (Yecora and Siette Cerros) and one tall and awnless of Italian origin (Generoso) were grown in the field for two seasons (1984-1985 and 1985-1986) under two irrigation treatments (irrigated and rainfed) in three replicates. The plots were arranged in a split-plot design with the cultivars as main plots and the irrigation treatments as subplots. In the second series, five faba bean cultivars belonging to the subspecies major (the Italian VT-1), equina (Brocal from Spain and the determinate FLIP-86 from Syria), and minor (the Greek KY-188 and the French R29-T) were cultivated for three seasons (1988-1989, 1989-1990, and 1990-1991). In the first and third seasons, plants were grown under dryland conditions in randomized blocks with four and three replicates, respectively. In the second season, a split-plot design was adopted with the cultivars as main plots and two irrigation treatments (irrigated and rainfed) as subplots randomized in three replicates. In all irrigation treatments, the amount of water was adjusted to bring soil to field capacity. Soil water was recorded by means of gypsum blocks placed at four different depths (10, 30, 50, and 70 cm). Irrigation, applied by means of drip irrigation systems, started 20 to 30 d before anthesis and continued up to maturity at approximately weekly intervals.
In all experiments, measurements of the water potential of the youngest fully expanded leaf (l) were taken around 1200 h (minimum daily value) twice a week throughout the growing seasons by means of the pressure bomb technique (Schollander et al., 1964) at a pressurization rate of 0.05 MPa s-1 Three to four plants per plot were sampled. To eliminate water loss, leaf blades were put in small polyethylene bags during cutting. The bags were then sealed and placed in dark and humid containers for the time required for determination. Seed yields were determined at harvest. Yield components were determined only in the series of faba bean experiments. WPD and WPI were calculated from the time courses of l according to Eq. [1] and [2].
|
Results and discussion
|
|---|
The values of the WPI of the wheat and faba bean cultivars for all series of experiments are shown in Tables 1 and 2
. In wheat (Table 1), WPI differed between irrigation treatments, illustrated more clearly in the second, drier, season (172 against 215 mm of rain in the first season). In faba beans (Table 2), irrigation caused only a slight increase in WPI in the second season because of the wet spring of 1990 (221 mm of rain). The spring was also wet in 1991 (256 mm), in contrast to 1989 when it was very dry (only 57 mm). These differences were reflected in the values of WPI in these seasons.
View this table:
[in this window]
[in a new window]
|
Table 1 The values of WPI among the wheat cultivars examined and irrigation treatments (W: irrigated, D: rainfed) in the two seasons (1984-1985 and 1985-1986). The LSDs (P = 0.05) for the cultivars and irrigation treatments are also shown
|
|
View this table:
[in this window]
[in a new window]
|
Table 2 The values of WPI among the faba bean cultivars examined and irrigation treatments (W: irrigated, D: rainfed) in the three seasons (1988-1989, 1989-1990, 1990-1991). The LSDs (P = 0.05) for the cultivars and irrigation treatments are also shown
|
|
The linear regressions of seed yields for the wheat and faba bean cultivars against WPI are shown in Fig. 2 and 3
. For each species, regressions were performed using data from all treatments and years. No consistent deviations along the fitted lines among the different seasons were detected (Fig. 4)
. However, differences in the regression parameters among the cultivars were observed. In wheat, significant differences among both the Y-intercepts and regression slopes were detected (Table 3) , whereas in faba beans (Table 4)
only the Y-intercepts differed significantly. In general, the higher values of the Y-intercepts (a) are indicative of a higher potential yield for a given genotype. In addition, the value of the slope (b) is an expression of the "adaptability" of the genotype at different levels of water shortage: a lower slope indicates the ability to maintain higher yields with increasing degrees of water stress, i.e., a higher adaptability, and vice versa. Accordingly, the higher values of the intercepts in the two semidwarf wheat cultivars definitely show higher potential yields (13.2 and 12.7 Mg/ha for Yecora and Siette Cerros, respectively) in comparison with the taller cultivar Generoso (7.2 Mg/ha) when water is not limiting. However, the lower slope of Generoso is indicative of a higher adaptability of this cultivar (Fig. 2, Table 3). In the faba bean cultivars (Fig. 3, Table 4), definite differences among the groups of cultivars according to seed size were observed. Potential yields were lowest (about 4.8 Mg/ha) in the two small-seeded (R29-T and KY-188), and highest in the medium- (Brocal and FLIP-86) and large-seeded (VT-1) cultivars (between 6 and 6.6 Mg/ha), whereas the slope was lowest (i.e., the adaptability higher), although not significantly, in the small-seeded cultivars.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 2 The fitted linear regressions between grain yields and water potential index (WPI) for the bread wheat cultivars Yecora, Siette Cerros, and Generoso. The points represent results from each plot. Data for both seasons (1984-1985 and 1985-1986) are included
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 3 The fitted linear regressions between seed yields and water potential index (WPI) for the faba bean cultivars VT-1, FLIP-86, Brocal, KY-188, and R29-T. The points represent results from each plot. Data for the three seasons (1988-1989, 1989-1990, and 1990-1991) are included
|
|
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 4 The fitted linear regressions between yields and WPI for the wheat cultivar Yecora (squares) and the faba bean cultivar KY-188 (circles). The data for the different seasons are shown [first season, open symbols; second season, filled symbols; third season (in faba beans only), half-filled symbols]. No marked deviations from the fitted lines among seasons are observed
|
|
View this table:
[in this window]
[in a new window]
|
Table 3 The parameters of the fitted linear regressions (a: Y-intercept, b: regression coefficient, r2: coefficient of determination) of grain yield against WPI for three bread wheat cultivars. The F-ratios of the comparisons among a and b are also shown
|
|
View this table:
[in this window]
[in a new window]
|
Table 4 The parameters of the fitted linear regressions (a: Y-intercept, b: regression coefficient, r2: coefficient of determination) of grain yields against WPI for five faba bean cultivars. The F-ratios of the comparisons among a and b are also shown
|
|
Keim and Kronstad (1979) pointed out that the value of the slope alone is not a sufficient indicator of adaptability of genotypes evaluated by the regression technique. A consideration of the yield performance either under stress or non-stress conditions would also be important. A combined study of both intercept and slope in the yield v. WPI regressions leads to different response-scenarios of hypothetical genotypes. In Fig. 5
, Line 1 presents the performance of an "ideal" cultivar which combines the highest potential yield (highest a) with the highest adaptability (lowest b). This cultivar exhibits the highest yields over a wide range of WPI. On the other hand, Lines 2 and 3 show situations similar to those found in this study, i.e., either low potential yields with low slopes (wheat cv. Generoso, faba bean cv. R29-T and KY-188) or high potential yields with high slopes (wheat cv. Yecora and Siette Cerros, faba bean cv. Brocal, FLIP-86, and VT-1). Line 4 describes the least desirable case, namely a low potential yield and a high slope. The association of high potential yields with low adaptability already mentioned here has also been detected in other works (Laing and Fischer, 1977). It is not surprising from the physiological point of view that the absolute reduction in yield for a given reduction in any resource is strongly influenced by the level of potential yield (Fischer and Maurer, 1978). Alternatively, a high potential yield with a low slope would be biologically less feasible, although the opposite combination is possible (Fischer and Maurer, 1978).
View larger version (11K):
[in this window]
[in a new window]
|
Fig. 5 Representation of the regression lines between yields and WPI for four hypothetical cultivars exhibiting different behavior under drought. 1: high potential yield, high adaptability. 2: high potential yield, low adaptability. 3: low potential yield, high adaptability. 4: low potential yield, low adaptability
|
|
A way to remove the effects of potential yield already mentioned is to divide the slope by the intercept in our regression model:
where bN could be named "relative adaptability" to drought. A similar normalization approach has also been used by others in other regression models (e.g., Chinoy, 1947; Blum, 1973; Fischer and Maurer, 1978). From the theoretical analysis made above, it follows that bN will take its lowest values for an "ideal" cultivar (i.e., exhibiting high a and low b) and its highest ones for the most undesirable case (low a and high b). Table 5
shows the values of bN for our wheat and faba bean genotypes. Wheat cultivars exhibited lower values (from 0.31–0.41) than those of faba bean (0.52–0.59), a fact that can be associated with the greater ability of the wheat plant to produce under arid conditions. Nevertheless, taking into account the findings of Rowe (1995) on the credibility of the use of ratios of two traits as selection criteria, more analysis on the meaning of the variation of bN in relation to the changes in a and b is required.
More information related to the physiological effects of water shortage on yield or yield components could be obtained by using either WPD or WPI for specific periods in the life history of crop plants. Table 6
shows the coefficients of determination of the regressions of wheat grain yields against WPI both before and after anthesis. In all cases, WPI after anthesis was more closely correlated with grain yields than WPI before anthesis. This implies more decisive effects of water shortage on yields through mechanisms operating after anthesis (e.g., grain filling) in our experiments, which could be further clarified if yield component data were available. Table 7
shows the coefficients of determination of the linear regressions of faba bean yield components against WPI. The numbers of fertile stems/plant as well as the numbers of seeds/pod appear to be independent of plant water status for all the examined cultivars. On the other hand, the number of pods/stem and the average seed weight exhibited significant correlations with WPI for most cultivars. These results are indicative of a different sensitivity of the faba bean yield components, and , hence, of the mechanisms related to them, to water shortage in our experiments.
View this table:
[in this window]
[in a new window]
|
Table 6 The coefficients of determination (r2) of the linear regressions of grain yields against WPI before and after anthesis for the examined wheat cultivars
|
|
View this table:
[in this window]
[in a new window]
|
Table 7 The coefficients of determination (r2) of the linear regressions of the yield components of the examined faba bean cultivars against WPI
|
|
|
Conclusions
|
|---|
The use of the WPD or WPI as indices of drought intensity in genotype evaluation studies provides both a physiologically sound and direct approach to the rational quantification of water stress for a number of reasons. First, at any time, leaf water potential is a combined result of interactions of soil water availability, evaporative demand, and plant responses, and, hence, a reliable indicator of plant water balance. Second, leaf water potential can be measured quickly and relatively easily with simple and cheap instrumentation accessible to any laboratory, such as the pressure chamber. Third, the destructive sampling involved is restricted to one leaf per plant, and does not create problems of gaps and non-uniformity of the stands even in small field plots. Fourth, it is possible to assess genotypes by inducing different degrees of water shortage in one location, instead of conducting field trials over a range of locations. Finally, the calculation of both WPD and WPI is very simple. It is important, however, to make frequent recordings of (at least twice a week) to obtain a sufficiently accurate figure of the water stress history over the growth period under consideration. We believe that the regression analyses of yield on WPD/WPI will provide adequate information on the drought adaptability of genotypes without the ambiguity of using indirect measures of either general or specific aspects of any environment.
Received for publication August 18, 1998.
|
REFERENCES
|
|---|
- Blum A. Components analysis of yield responses to drought of sorghum hybrids. Exp. Agric. 1973;9:159-167.
- Chinoy J.J. Correlation between yield of wheat and temperature during ripening of grain. Nature (London) 1947;159:442-444.
- Eberhart S.A., Russell W.A. Stability parameters for comparing varieties. Crop Sci. 1966;6:36-40.[Abstract/Free Full Text]
- Finlay K.W., Wilkinson G.N. The analysis of adaptation in a plant-breeding programme. Aust. J. Agric. Res. 1963;14:742-754.[ISI]
- Fischer R.A., Maurer R. Drought resistance in spring wheat cultivars. Aust. J. Agric. Res. 1978;29:897-912.[ISI]
- Gardner B.R., Nielsen D.C., Shock C.C. Infrared thermometry and the crop water stress index. II. Sampling procedures and interpretation. J. Prod. Agric. 1992;5:466-475.
- Idso S.B., Reginato R., Reicosky D., Hatfield J. Determining soil-induced plant water potential depressions in alfalfa by means of infrared thermometry. Agron. J. 1981;73:826-830 a.[Abstract/Free Full Text]
- Idso S.B., Jackson R.D., Pinter P.J., Reginato R.J., Hatfield J.L. Normalizing the stress-degree-day parameter for environmental variability. Agric. Meteorol. 1981;24:45-55 b.
- Jackson R., Idso S.B., Reginato R., Pinter P.J. Canopy temperature as a crop water stress indicator. Water Res. Res. 1981;17:1133-1138.
- Karamanos A.J. Ways of detecting adaptive responses of cultivated plants to drought. In: Margaris N.S., et al. , ed. Being alive on land. The Hague: Tasks for vegetation science. Dr. W. Junk, 1984:91-101.
- Keim D.L., Kronstad W.E. Drought resistance and dryland adaptation in winter wheat. Crop Sci. 1979;19:574-576.[Abstract/Free Full Text]
- Laing D.R., Fischer R.A. Adaptation of semidwarf wheat cultivars to rainfed conditions. Euphytica 1977;26:129-139.[ISI]
- Lin C.S., Binns M.R. A superiority measure of cultivar performance for cultivar x location data. Can. J. Plant Sci. 1988;68:193-198.
- Nielsen D.C., Anderson R.L. Infrared thermometry to measure single leaf temperatures for quantification of water stress in sunflower. Agron. J. 1989;81:840-842.[Abstract/Free Full Text]
- Rowe D.E. Characteristics of elite population selected on ratio citerion: I. Traits with equal genetic variances. Crop Sci. 1995;35:425-430.
- Schollander P.F., Hammel H.T., Hemmingsen E.A., Bradstreet E.D. Hydrostatic pressure and osmotic potential in leaves of mangroves and some other plants. Proc. Nat. Acad. Sci. (USA) 1964;52:119-125.[Free Full Text]
- Sullivan C.E., Ross W.M. Selecting for drought and heat resistance in grain sorghum. In: Mussel H., Staples R.C., eds. Stress physiology in crop plants. New York: John Wiley and Sons, 1979:263-281.
This article has been cited by other articles:
|
|
|
|
 
F. Rizza, F. W. Badeck, L. Cattivelli, O. Lidestri, N. Di Fonzo, and A. M. Stanca
Use of a Water Stress Index to Identify Barley Genotypes Adapted to Rainfed and Irrigated Conditions
Crop Sci.,
November 1, 2004;
44(6):
2127 - 2137.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
