| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Forestry, Oklahoma State Univ., Stillwater, OK 74078 USA
bcm{at}soilwater.agr.okstate.edu
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
) is a property that may be used to improve water-use efficiency (WUE). This study tested the association between and WUE with plant materials and growth conditions likely to disrupt the link between and WUE. The cultivated tomato, Lycopersicon esculentum Mill. cv. UC82B, a drought-resistant related species, L. pennellii (Cor.) D'Arcy accession LA716, and the F1 and F2 generations of the L. esculentum x L. pennellii cross were grown in containers in wet and dry field environments. The wet environment was repeated in a second year and plants were split into groups terminated early and late in the season. The F1 generation had greater mean WUE and dry weight (DW) than L. esculentum, but the DW advantage was not maintained in the F2 generation. Low of L. pennellii suggested that leaf WUE was high, but its whole plant WUE varied relative to the other plant materials in the different environments. There was a negative correlation between and WUE in the F2 generation, and WUE was generally positively correlated with DW. However, low was associated with large DW in only one environment and with small DW in three environments. Averaged across environments, the top 10% of the plants ranked by WUE had 47% greater WUE than the bottom 10%. In comparison, the bottom 10% ranked by had an average of 16% greater WUE than the top group, but in three of the four environments the bottom group accumulated 33 to 47% less DW than the top group. This study on tomato suggests that WUE can be increased by selecting low , but selecting low alone may identify a subpopulation of small plants. Dry weight could probably be increased by traditional breeding techniques.
Abbreviations: DW, dry weight • ea, ei, partial pressure of water vapor in the air and the intercellular air spaces • pa, pi, partial pressure of CO2 in the air and the intercellular air spaces • S/R, shoot/root ratio (w/w) • WU, water use • WUE, water-use efficiency • , stable carbon isotope discrimination • 
c, respiratory CO2 loss as a fraction of daily CO2 fixation • w, uncontrolled water loss as a fraction of the daytime stomatal water loss
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
production depends on the amount of water used for growth as well as on water-use efficiency, WUE. The simple relationship DW = WUE x WU implies that dry weight, DW, increases with increasing WUE for constant water use, WU. Thus, WUE is particularly important in those circumstances where growth ceases as a result of depletion of a finite and limiting water source. Numerous studies have noted considerable variation in WUE among plants (Martin and Thorstenson, 1988; Virgona et al., 1990; Ehdaie et al., 1991; Nageswara Rao et al., 1993; Ebdon et al., 1998), although Sinclair et al. (1984) proposed that exploitable variation is small. Little has been done to increase WUE largely because a tool to distinguish between varying degrees of WUE among field-grown plants is lacking, and maintenance of breeding populations in individual containers demands enormous manpower and cost. Large plants in a variable population usually have large WU but not great WUE. Development of a productive monoculture is not possible with this type of large plant, because in a monoculture neighboring plants with lower water demand would not be present.
Water-use efficiency at the level of instantaneous leaf gas exchange (photosynthesis rate/transpiration rate) and stable carbon isotope discrimination (, a measure of the proportion of 13C relative to 12C in plant organic matter), are independently linked to the ratio of the internal to external partial pressures of CO2, pi/pa, as Eq. [1] and [2] (Farquhar et al., 1982, 1989; Hubick and Farquhar, 1989) imply
 | (1) |
 | (2) |
Hubick and Farquhar (1989) combined Eq. [1] and [2] and introduced two inefficiency terms, c and w, to obtain Eq. [3] that describes the link between and long-term whole-plant WUE as follows:
 | (3) |
c is the amount of respired CO2 as a fraction of daily CO2 fixation, and w is uncontrolled water loss, such as night time stomatal transpiration and water loss by nonphotosynthetic plant parts, expressed as a fraction of daytime stomatal water loss in exchange for CO2.
Equations [1] through [3] form the framework for the proposition that provides a means of ranking WUE in C3 plants. To date, numerous studies have experimentally verified the link, and it has been advocated that breeders use to improve crop WUE (Farquhar and Richards, 1984; Hubick et al., 1986, 1988; Hubick and Farquhar, 1989; Condon et al., 1990; Johnson et al., 1990; Ehleringer et al., 1990, 1991; Ismail and Hall, 1992; Hall et al., 1994). However, crop improvement programs are costly and span many years, so the technique must be robust before it is incorporated into improvement programs.
While Eq. [3] formulates the theory for proposing the use of to evaluate WUE, the equation contains terms also suggesting that the association between and WUE may at times fail. In fact, values of different plants will exactly mirror WUE values only if is the single variable on the right side of the equation. The isotopic fractionation terms a, b, and d, should pose little problem because they relate to physical properties of CO2 and Rubisco and should vary little, if at all (Hubick and Farquhar, 1989; O'Leary, 1993). In contrast, ea changes with season, and because ei depends on leaf temperature, it varies both daily and seasonally. If leaf properties such as vapor conductance and radiation reflectance vary among plants, the resultant variation in ei could be particularly troublesome. Variation in the two vapor terms may disrupt the relationship between and WUE either if leaf temperatures differ among plants on a given day, or if biomass is predominantly produced in different parts of the season with different ea and ei. Equations [1] and [2] reveal that under such scenarios, two leaves may have the same pi/pa and thus incorporate carbon with identical isotopic fingerprints; yet because (ei - ea) differs, values of leaf WUE will differ. Moreover, the inefficiency terms c and w may vary among plants and disrupt the association between and WUE suggested by Eq. [3]. Hubick and Farquhar (1989) estimated that c may range from 0.3 to 0.5. The magnitude of variation in w is uncertain. However, the observation that cuticular transpiration accounts for about one third of total transpiration in plants from shady and moist habitats but only 1 to 2% of total transpiration in succulent species suggests that w might vary considerably among plants (Larcher, 1995).
Our previous work with plants grown both outside and in controlled-environment growth chambers showed that L. pennellii had greater WUE and smaller than L. esculentum and that WUE of the F2 generation was negatively correlated with as predicted by Eq. [3] (Martin and Thorstenson, 1988; Martin et al., 1993; Kebede et al., 1994). These differences were adequately explained by smaller leaf conductance caused by fewer and smaller stomata in L. pennellii than L. esculentum (Kebede et al., 1994). However, L. pennellii also had longer trichomes, lower chlorophyll content and Rubisco activity per unit leaf area, and proportionally larger mesophyll cell surface area exposed to intercellular air space than L. esculentum. The very small root of L. pennellii (1–4% of total DW) (Martin and Thorstenson, 1988; Martin et al., 1993) compared with L. esculentum (9–15% of total DW) might also cause differences in c among individuals in the F2 generation. Furthermore, L. pennellii is small seeded, germinates slowly, and shows indeterminate growth as opposed to the determinate growth type of the L. esculentum cultivar used here. The combined variation in all of these traits in the F2 generation used in this study was expected to reduce the magnitude of the association between and WUE.
Preliminary efforts to grow L. pennellii in the field in Oklahoma produced severely stunted plants with a generally unhealthy appearance, suggesting lack of adaptation to the environment. We recognized that the morphological and physiological variation in F2 plants mentioned above, combined with the potential variation in adaptation ranging from well adapted individuals resembling plants of L. esculentum to poorly adapted ones resembling L. pennellii, might offer an opportunity to disrupt the link between WUE and . Tomato was used solely as a model experimental material without considering its fruit crop. The first objective was to test the robustness of the relationship between and WUE under conditions thought to interfere with the linear association between and WUE expected from Eq. [3]. The second objective was to evaluate the potential for gain in WUE by indirect selection of .
|
Materials and methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
1995 Experiment. Each morning, one group of plants was watered to 95% of field capacity of the growth substrate (wet environment) and the other group to 45% of field capacity (dry environment). Water levels were maintained through a daily weighing and watering schedule previously described by Martin and Thorstenson (1988). After watering, each container was daily assigned a new random position in the field. On 29 June and four times between 20 July and 26 July, sufficient rain fell to exceed the field capacity of the growth substrate. After the rains, excess water was drained and collected by gently tilting the containers. The collected water was recorded and accounted for in the calculation of WU. During 1 August through 5 August, the plants were placed in a greenhouse to shelter them from forecasted intense rains. On 25 July, the plants were treated with fungicide (Maneb—active ingredient, manganese ethylenebisdithiocarbamate; Dexol Industries, Torrance, CA) and on 30 June and 10 July with insecticide (Dipel Worm Killer—active ingredient, Bacillus thuringiensis var. kurstaki spores and delta-endotoxin; Abbott Laboratories, North Chicago, IL). The experiment was terminated on 7 August.
1996 Experiment. All plants were maintained at 95% of field capacity, but one group of plants was terminated on 8 July (early environment) and the remaining group on 5 August (late environment). In this year, the plants were kept under a large rainout shelter consisting of a clear plastic canopy supported by aluminum beams to avoid problems with heavy rains. The shelter was open on all four sides to facilitate air movement. The plants were placed in new random positions each day and treated with insecticide (Dipel Worm Killer) on 29 June.
Carbon Isotope Discrimination
On the days experiments were terminated, 10 leaflets were randomly collected from each plant and dried at 70°C. The dry leaves were ground to a fine powder with a KSM2 Automatic Coffee grinder (Braun Inc., Woburn, MA) and sent to the Stable Isotope Ratio Facility for Environmental Research, University of Utah, Salt Lake City, for carbon isotope analysis. Isotope values are expressed in notation of discrimination, , with a 
13C value of CO2 in the air of -8.00
.
Dry Weight and Shoot/Root Ratio
After collecting leaf samples, the shoot was cut at the soil surface and the root was rinsed free of growth substrate with a gentle jet of water. Roots and shoots were placed in separate paper bags and dried to constant weight in an oven at 70°C. Plant DW was obtained by summing weights of root and shoot. The shoot DW included the weight of the sample collected for analysis. The shoot/root ratio, S/R, was calculated on a dry weight basis (g g-1).
Water Use
Plant WU was calculated as reported by Martin and Thorstenson (1988) by summing daily water loss from the containers. Corrections were made for soil evaporation obtained from containers without plants and for water drained off after a few large precipitation events (in 1995 only, see above).
Water-Use Efficiency
Plant WUE was calculated as DW divided by WU (g kg-1).
Statistical Analysis
Linear regression was used to fit lines to the data in Fig. 4 to 7 . Means of high and low groups in the simulated selection experiment (Table 1)
were separated by t-test. Heritability, h, was calculated from the relationship
, where Vh is variance due to genotype, and Ve is variance due to the environment. Vh was obtained by subtracting Ve from the total variance (variance of the F2 population). Computation of Ve as (VL. esc x VL. pen x VF1)1/3 is a modification of the procedure described by Briggs and Knowles (1967)(p. 104–105) who calculated Ve as the average variance of the parents. VL. esc, VL. pen, and VF1 are the variances of L. esculentum, L. pennellii, and the F1 generation, respectively.
|
|
|
|
|
|
Results |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
|
|
, the F1 generation was not heterotic (Fig. 3A to D)
. Most F2 individuals had lower values than L. esculentum and the F1, but few approached the low values of L. pennellii. In the late environment, of the F1 differed from the other environments by falling closer to L. pennellii than L. esculentum. Reduced availability of water and late growth tended to lower values, but less so in the parents than in the F1 and F2 generations.
|
A negative relationship was observed between WUE and across environments in both years
(Fig. 5). Water-use efficiency was negatively associated with in each of the four environments (r-values ranged between -0.38 and -0.23 with P < 0.05 to P < 0.001).
In 1995, DW was positively correlated with WUE in the dry environment
but not in the wet environment or across environments. Therefore, the regression line in Fig. 6A is for the dry environment only. Dry weight was positively correlated with WUE both in the early and late environments in 1996
. However, the slope was considerably greater in the late than the early environment (Fig. 6B).
Regression of DW on across both wet and dry environments showed a positive linear relationship in
. The relationship was very strong for each environment also (
, P < 0.01 in the wet and the dry environments, respectively). In 1996, DW was positively correlated with in the early environment
but the correlation became negative when the growth period was extended by 4 wk and DW more than doubled (
, Fig. 7B).
Simulated selection was performed from the 100 F2 plants in each of the four environments (Table 1). The two data columns on the left show the outcome of simulated direct selection of the 10 plants with the greatest WUE compared to the 10 plants with the smallest WUE. The two data columns on the right show the outcome of indirect selection for WUE, i.e., selection of the group of 10 plants with the smallest (predicted to have the largest WUE) and the group of 10 plants with the largest (predicted to have the smallest WUE). By means of direct selection, the top groups had 61, 49, 37, and 38% greater WUE than the bottom groups in the wet, dry, early, and late environments, respectively (Table 1). Indirectly selected by , the top groups (lowest ) had 30, 15, and 14% greater WUE than the bottom groups (highest ) in the first three environments. Indirect selection tended to increase WUE in the late environment also, but the difference in WUE between the low- and the high- groups was not significant (P < 0.11). The estimated heritability, h, of was 0.95, 0.83, 0.79, and 0.61 in the wet, dry, early, and late environments, respectively, whereas h of WUE was 0.77, 0.80, 0.62, and 0.84 in these environments. Indirect selection reduced DW by 33 to 47% and WU by 44 to 60% in the low- (high WUE) groups in all but the late environment. In the late environment, the low- group had 55% greater DW and 45% greater WU than the high- group. The groups selected directly for high WUE had lower in the late environment (in the other environments, tended to be lower also but the differences were not statistically significant), greater DW in the dry and the late environments, smaller WU in the wet environment, and greater WU in the late environment. The indirectly selected group with high WUE had the greater S/R ratio in wet, dry and early environments, while the directly selected group with high WUE had the greater S/R ratio in the early environment only.
|
Discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
At the instantaneous leaf level, should vary in concert with WUE because both and leaf WUE depend on pi/pa in accordance with Eq. [1] and [2]. However, Eq. [3] suggests several reasons relevant to this study why the link between and long-term whole-plant WUE may break down. First, the Oklahoma environment is unsuitable for growth of the L. pennellii. Numerous F2 individuals approached the small DW of L. pennellii (Fig. 1), suggesting that they were also unadapted to the Oklahoma environment. Thus, maintenance respiration and c might vary among the offspring in response to variable need to repair stress-damaged cell components, the need of repair being the greatest in offspring dominated by L. pennellii genes. Second, the parents have different relative root size (Martin and Thorstenson, 1988; Martin et al., 1993) so that root DW as a proportion of total DW ranged from a low of 2% to a high of 22% for the F2 plants in this study. To maintain existing roots and grow new root tissue, respiration likely differed between the parents and among individual F2 plants. Consequently, c might vary in the F2 generation. Third, L. pennellii grows indeterminately, is smaller-seeded, and requires a longer time from planting to emergence than does L. esculentum. As a result, relatively more carbon was probably acquired late in the season at greater (ei-ea) in F2 plants dominated by L. pennellii genes. Fourth, small vapor conductance resulting from small and few stomata, and possibly a thick boundary layer because of long trichomes (Kebede et al., 1994) trapping stagnant air on top of leaf surfaces, might elevate leaf temperature of L. pennellii-like genotypes as compared with L. esculentum-like genotypes. This would result in variation in leaf temperature and, therefore, ei. Finally, Martin et al. (1993) failed to detect water loss from severely water-stressed leaves of L. pennellii, whereas L. esculentum leaves seemed unable to completely stop water loss. As a result, w of the offspring may vary. Combined, the differences in timing of growth, leaf physiological and morphological properties, and lack of adaptation to the Oklahoma environment could interfere with the link between and WUE by means of substantial variation among F2 individuals in any or all of ei, ea, c, and w at times when most carbon was deposited. In this study, the plants were spaced at a considerable distance from each other. The range in WUE would likely be diminished in a closed canopy having reduced stomatal control of transpiration because of a large canopy boundary layer effect (Condon and Hall, 1997).
The close association between DW and WU (Fig. 4) dispels the concern that the greater scattter in Fig. 5 resulted from difficulties in accurately determining WUE. Were that the case, data in Fig. 4 would show more scatter as well. The scatter in Fig. 5 is found in other studies also and, we believe, stems from the loosening of the association between and WUE. The considerable difference in slopes of the regression lines (Fig. 4) suggests there were important weather differences between 1995 and 1996. Supporting evidence of weather differences comes from the comparison between the wet environment in 1995 and the early environment in 1996. These two experiments were replications in time of the same basic experiment (same plant materials grown for about the same length of time during about the same time of the year, producing about the same amount of DW). Yet, F2 plants in the early environment had 14% greater mean WUE than plants in the wet environment (Fig. 2).
In spite of the many reasons to break the relationship between and WUE in this study, the relationship persisted, although it was weaker than in previous studies in which Martin and Thorstenson (1988) and Martin et al. (1993) calculated that explained 45 to 66% of the variation in WUE in the F2 generation. In this study, calculated values of r2 revealed that explained 18 and 19% of the variation in WUE across environments in 1995 and 1996, respectively, and 5 to 14% in individual environments (Fig. 5). Condon et al. (1993) found that variation in the timing of growth broke the association between and WUE in wheat (Triticum aestivum L.), but the relationship could be reestablished if WUE was adjusted for transpiration-weighted vapor pressure differences over the season. In a study on Kentucky bluegrass (Poa pratensis L.), Ebdon et al. (1998) concluded that the association was broken at certain harvest times by variation in leaf temperature.
While WUE of L. pennellii was either greater or smaller than for the other plant materials (Fig. 2), was consistently the smallest in this species (Fig. 3). In our previous studies, L. pennellii always had greater WUE and smaller than L. esculentum (Martin and Thorstenson, 1988; Martin et al., 1993). The low values in all environments here (Fig. 3) combined with inconsistent WUE ranking (Fig. 2), suggest that L. pennellii had the greatest leaf WUE, but that whole plant WUE was under strong influence by factors beyond the leaf level, possibly factors associated with lack of environmental adaptation and stress.
There are contradicting reports in the literature concerning whether the association between WUE and DW is positive or negative. Three of the four environments in the current study showed a positive correlation; however, WUE explained only 6 to 27% of the variation in DW depending on environment (Fig. 6). It is nevertheless encouraging that plants with great WUE had large DW. (See also the discussion of DW in the paragraph below dealing with the simulated selection data in Table 1.)
The published literature contains reports of positive associations (Condon et al., 1987; Ehleringer et al., 1990; Virgona et al., 1990; Kirda et al., 1992; Morgan et al., 1993) between and DW, but also of negative or no association (Meinzer et al., 1991; Read et al., 1991, 1992; Al Hakimi et al., 1997; Jefferies and MacKerron, 1997). In the current study, explained 33% of the variation in DW across the two environments in 1995 and 9 and 12% in the early and late environments in 1996, respectively (Fig. 7). In general, the strongest positive correlations seem to have been observed under well-watered conditions, but there are exceptions (Sadiq et al., 1997). The correlation between and DW in our study was positive in 1995 and the early environment in 1996, but negative in the late environment in 1996 (Fig. 7), which may be related to stress from root restriction of large late-season plants.
Thus, large plants tended to have high whole plant WUE (Fig. 6), but judged by , only large plants in the late environment had high leaf WUE (Fig. 7). Apparently, in most situations other factors, such as WU perhaps, contributed more to DW accumulation than did leaf WUE. However, with the large plants in the driest and hottest part of late summer, leaf WUE became increasingly important.
Figure 5 suggests that the gain in WUE by selecting low should be fairly small (
, respectively). Simulated selection (Table 1) showed that selection of low- plants would probably produce about one third of the improvement in WUE compared with direct selection of WUE, if direct selection were practically possible. The high heritability of
and
suggests improvement of and WUE may be achieved through selection in this population. It should be cautioned, however, that similarly high h values might not be obtained in those environments where L. pennellii is better adapted and where practical breeding programs would most likely be located. The success of indirect selection of WUE using generally agrees with findings on wheat by Ehdaie et al. (1993), but in their study the efficiency of indirect selection was greater in dry than wet environments, while in the current study the opposite was suggested by the smaller difference in between extremes (Table 1) and the lower h values in the dry than the wet environment. Whereas direct selection for WUE should increase DW in the dry and the late environments and elsewhere have no effect on DW, indirect selection of low- plants should decrease DW in all environments except the late environment where DW would be unaffected (Table 1). This implies that while indirect selection of plants with the low- phenotype should succeed in increasing WUE, the resulting subpopulation of plants would differ from the subpopulation recovered by direct selection for WUE so that indirect selection might reduce DW. Low single-plant DW is often viewed unfavorably, many times perhaps unnecessarily so. First, low individual plant DW does not necessarily indicate low DW per unit land area. Smaller, but more water-use efficient plants may allow planting at higher stand density and could quite possibly increase DW per unit land area. Second, standard breeding practices that apply simultaneous selection for and DW, or backcrossing to the productive parent, should increase individual plant DW of early breeding materials, if so desired.
We conclude that the association between WUE and seems robust because it was maintained in the genetically highly variable F2 generation and under relatively adverse environmental conditions. Although selection of low captured only one third of the potential improvement in WUE, this study suggests that is a viable tool to improve crops, especially with the current lack of alternative means to rank WUE of field-grown plants. The reduced plant DW that would result from obtaining high WUE by selection of low- plants would not necessarily be a problem in agriculture, or it could be offset with standard breeding practices.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
Received for publication January 20, 1999.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
13C), water use efficiency, and biomass productivity of Lycopersicon esculentum, Lycopersicon pennellii, and the F1 hybrid. Plant Physiol. 1988;88:213-217.This article has been cited by other articles:
|
|
 |
|
  C. R. de Souza, J. P. Maroco, T. P. dos Santos, M. L. Rodrigues, C. M. Lopes, J. S. Pereira, and M. M. Chaves Impact of deficit irrigation on water use efficiency and carbon isotope composition ({delta}13C) of field-grown grapevines under Mediterranean climate J. Exp. Bot., August 1, 2005; 56(418): 2163 - 2172. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Ebdon and K. L. Kopp Relationships between Water Use Efficiency, Carbon Isotope Discrimination, and Turf Performance in Genotypes of Kentucky Bluegrass during Drought Crop Sci., September 1, 2004; 44(5): 1754 - 1762. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
