| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Centro Internacional de Agricultura Tropical (CIAT), A.A. 6713 Cali, Colombia
b Institut für Ökologie und Naturschutz, Universität Wien, Althanstraße 14, A-1090 Wien, Austria
c Center for the Application of Molecular Biology to International Agriculture (CAMBIA), GPO Box 3200 Canberra, ACT 2601, Australia
* Corresponding author (peter{at}cambia.org).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: AN, nutrient solution with adequate nutrient levels • AN + Al, nutrient solution with adequate nutrient levels containing Al • Alinorg, inorganic monomeric Al • LA, leaf area • LN, nutrient solution with low nutrient levels • LN + Al, nutrient solution with low nutrient levels containing Al • RDW, root dry weight • RL, total root length • SDW, shoot dry weight • [ ], concentration of an ion or an ion complex • { }, activity of an ion or an ion complex
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Many of these studies were conducted with solution culture rather than with soil to circumvent the problem that several soil properties change simultaneously when soil acidity is modified. Initially, solutions contained nutrient levels far in excess of those required for maximum plant growth rates (see critiques by Blamey et al. [1991] and Edmeades et al. [1995]). In such solutions, Al toxicity is alleviated as a result of physicochemical interactions between Al and other ions, including the formation of nontoxic complexes with OH-, SO2-4, and silicate ions; precipitation of Al as hydroxide or phosphate; and high ionic strength per se (Blamey et al., 1983; Blamey et al., 1991; Wheeler and Edmeades, 1995; Kinraide, 1997; Ma et al., 1997). In addition, high concentrations of divalent and, to a lesser extent, monovalent cations can ameliorate Al toxicity, presumably because they reduce cell-surface negativity (Kinraide and Parker, 1987; Kinraide et al., 1992; Kinraide, 1998).
Recognition of these interactions led to the use of low-ionic-strength nutrient solutions to avoid artifacts caused by unrealistic Al speciation (Blamey et al., 1983, 1991; Wheeler et al., 1992; Edmeades et al., 1995). These solutions were designed to simulate soil solutions of acid soils because the composition of these determines the bioavailability of nutrients (Pearson, 1971; Parker and Norvell, 1999). Nutrient concentrations, about one order of magnitude lower than those of earlier solution-culture techniques, are at the minimum levels required to achieve close-to-maximum plant growth rates (Asher, 1978, 1986; Edmeades et al., 1995; Wheeler, 1996). As a consequence of these more realistic conditions, Al3+ activities ({Al3+}) required for inhibiting root growth are at least one order of magnitude lower than those of earlier studies. They are comparable with values inferred from studies relating root-growth inhibition to {Al3+} measured in soil solutions (Pavan et al., 1982; Wright et al., 1989; Blamey et al., 1991).
Nutrient deficiencies associated with soil acidity may be particularly severe in South American savannas, because the soils of these areas tend to be highly weathered and more acidic than comparable regions on other continents (Eswaran et al., 1997). Hence, currently used low-ionic-strength nutrient solutions supporting close-to-maximum plant growth rates may not reflect the various nutritional constraints under which plants are exposed to Al in these soils. We hypothesized that an adequate assessment of Al resistance and edaphic adaptation of crops targeted for these areas requires a solution with low concentrations of those nutrients that typically limit plant growth on these soils.
Large areas of the South American savannas have been sown to pastures based on forage grasses of the Brachiaria genus. A field study in an area representing the highly weathered acid soils of the Colombian and Venezuelan savannas had previously shown that acid soil-adapted signalgrass produces a higher forage yield than less-adapted ruzigrass (CIAT, 1995). We therefore used these two species, as well as soils from this area, as a model to develop a solution-culture technique for rapidly evaluating the edaphic adaptation of crops targeted for these areas.
The initial objective of this multi-tiered study was to examine the nutrient status of the highly weathered acid soils of the Colombian savannas, and to design a nutrient solution that simulates chemical factors (Al toxicity, nutrient deficiencies) that limit plant growth in these soils. To test whether the designed solution was a realistic approximation to the chemical properties of these soils, we then examined whether its growth-inhibitory effect resulted in a similar growth difference between signalgrass and ruzigrass as had been observed in the field experiment mentioned above. Finally, to investigate whether interactions among the various stress factors must be taken into account for an accurate assessment of edaphic adaptation, we dissected the stress syndrome into Al toxicity and nutrient deficiency and examined whether they inhibited growth of the two grasses in a synergistic manner.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The characteristics of the sandy loam (clay loam) before fertilization, previously determined according to Salinas and García (1985), were as follows: 170 (370) g kg-1 clay and 650 (180) g kg-1 sand, pH 5.1 (5.0) at a soil/water ratio of 1:1, 9 (34) g kg-1 organic matter, 24 (72) mmol kg-1 total N, 2.0 (2.1) mg kg-1 Bray II-extracted P, 0.03 (0.06) cmolc kg-1 Bray II-extracted K, 0.7 (2.6) cmolc kg-1 KCl-extracted Al, 0.13 (0.21) cmolc kg-1 KCl-extracted Ca, and 0.08 (0.10) cmolc kg-1 KCl-extracted Mg. Aluminum saturation was 77 (89) % and the effective cation-exchange capacity 0.94 (2.97) cmolc kg-1.
Applying Fertilizer and Rewetting Soil Samples
About 4 mo after collecting and air-drying the soils, a 20-kg aliquot of each soil type was taken. Fertilizer, consisting of (in milligrams of element per kilogram of soil): 20.6 N (urea), 25.8 P (triple superphosphate), 51.6 K (KCl), 34.0 Ca (dolomitic lime), 18.0 Ca (triple superphosphate), 14.6 Mg (dolomitic lime), 10.3 S (elemental sulfur), 1.0 Zn (ZnCl2), 1.0 Cu (CuCl2), 0.05 B (H3BO3), 0.05 Mo (Na2MoO4 · 2 H2O) was applied to one-half of each sample (10 kg). This mixture is routinely used to raise the fertility of Colombian savanna soils in field studies; addition of 1 mg of element per kg soil corresponds to a surface-area-based application rate of 1.94 kg of element ha-1 (Rao et al., 1992, 1996). The four soil samples, two with and two without fertilizer, were rewetted to 90% (w/w) field capacity by adding water at 144 mL kg-1 of dry soil for the sandy loam or 225 for the clay loam.
Incubating Soil Samples and Extracting Soil Solutions
Ten 333.3-g aliquots were taken from each of the four rewetted soil samples and transferred into 0.02-mm thick polyethylene bags (Bartlett, 1965). The bags, in upright position with their tops rolled back, were arranged in a single layer and incubated at 28°C. On Day 1, 2, 4, 8, 16, 32 and 48, one bag per soil type and fertilizer level (four aliquots in all) was removed and weighed. Soil solutions were extracted by centrifuging each aliquot, in several rounds, at 2000 x g for 45 min (Aitken and Outhwaite, 1987). On Day 64, three bags of each soil type and fertilizer level (12 aliquots in all) were processed to quantify the variability among aliquots (extraction of replicate aliquots for all time points was not possible because the process was too time-consuming). The average loss of soil moisture after 64 d of incubation was 12.5% ± 4.7 (mean ± SD) of the initial water content.
Analyzing Soil Solutions
The electrical conductivity and pH of each soil solution were determined in a 2.5-mL aliquot immediately after extraction. The remainder was filtered (0.22 µm; polycarbonate), and the concentration of inorganic monomeric Al (Alinorg) was estimated, using a modified version of the short-term colorimetric method described by Kerven et al. (1989): 300 µL of soil solution were mixed with 50 µL of 0.1% (w/v) 1,10-phenanthroline, 0.5% (w/v) ascorbic acid, 20 µL of 0.11% (w/v) pyrocatechol violet, and 100 µL of 1 M imidazole-HCl (pH = 5.9). Absorbance at 585 nm was measured exactly 1 min after adding the imidazole buffer. Aluminum chloride was used for calibration.
The remaining aliquot of each soil-solution sample was stored at -80°C for later analysis of nutrient concentrations. Ammonium was determined colorimetrically (Kandeler and Gerber, 1988). Other cations, including Ca2+, Mg2+, K+, and Na+, were measured by atomic-absorption spectroscopy (Unicam 969 spectrophotometer; Unicam, Cambridge, UK). Anions, such as phosphate, NO3-, SO2-4, and Cl-, were quantified by HPLC, using a Dionex AS-11 column (Dionex, Sunnyvale, CA) and gradient elution (0.5–37.5 mM NaOH in 18 min at a flow rate of 2 mL min-1 and 30°C). Suppressed conductivity was used for detection.
Simulating the Ionic Speciation of Soil Solutions
The ionic speciation of each soil-solution sample was simulated with GEOCHEM 2.0 software using the thermodynamic constants for the formation of Al complexes compiled by Nordstrom and May (1989) (Parker et al., 1987). The measured ion concentrations (Alinorg in the case of Al) were used as input values.
GEOCHEM simulations were performed at the measured pH values (3.9–5.9), disallowing precipitation of Al(OH)3 and other solids and permitting CO2 to penetrate the solution at a partial pressure of 3 x 10-4 atm (using the program's options "pH fixed," "open system," and "ionic strength calculated"). The program computed the activities of ions and ion complexes, using the Davies equation and an ionic strength that had been corrected for ion pairing predicted based on ionic-strength-adjusted stability constants. Throughout this paper, the parentheses [ ] and { } denote concentration and activity of a given ion, respectively.
Preparing the Designed Nutrient Solutions
A low-ionic-strength nutrient solution simulating the chemical properties of the highly weathered acid soils from the Colombian savannas was designed based on the data obtained in the first part of this study (LN + Al, low nutrient levels plus Al). In addition, three other treatments were designed: an Al-free reference treatment containing sufficient nutrients for close-to-maximum plant growth rates (AN, adequate nutrients) and two treatments for exposing plants to Al toxicity (AN + Al) or nutrient deficiency (LN) separately (see Discussion and Table 1 for the design criteria).
|
|
|
|
|
Growth Chamber
Seedlings were grown in containers holding 110 mL of nutrient solution per individual plant. Nutrient solutions were prepared by adding an acidified Al-stock solution (16 mM AlCl3, 4 mM HCl) and a 50x concentrated nutrient mixture (containing all nutrients plus the remainder of the HCl required to adjust the pH to 4.2) to vigorously stirred deionized water (MilliQ, 18 M
). The solutions were changed daily. Their pH tended to increase slightly over a 24-h period; the highest value in Al-free solutions, measured at harvest, was 4.45 (LN). The highest value in Al-containing solutions was 4.38 (LN + Al), a pH at which the Al3+ activity is predicted to be 40 µM (= 93% of its original value at pH 4.20; GEOCHEM).
Growth conditions were 24°C, a 12-h diurnal cycle, and a photon-flux density of photosynthetically active radiation of 110 µmol m-2 s-1. Because growth was slow under these low-light conditions, groups of three individual plants were treated as an experimental unit. This experiment consisted of two independent harvests, each comprising six experimental units (18 seedlings) for each of the two grasses and four treatments.
Greenhouse
Seedlings were grown in containers holding 350 mL of nutrient solution per individual plant. Nutrient solutions were prepared and changed daily as outlined for the growth chamber. The highest pH value in Al-free solutions, measured at harvest, was 4.53 (LN). The highest value in Al-containing solutions was 4.33 (LN + Al), a pH at which the Al3+ activity is predicted to be 41 µM (= 95% of its original value at pH 4.20; GEOCHEM).
The greenhouse was located at CIAT headquarters. Typical diurnal maximum and minimum temperatures in the greenhouse were 36 and 19°C. The typical maximum photon-flux density of photosynthetically active radiation during the day was 1100 µmol m-2 s-1. This experiment consisted of two independent harvests, each comprising six individual plants for each of the two grasses and four treatments.
Measuring Plant Growth and Nutrient Uptake
Plants were harvested 20 d (growth chamber) or 13 d (greenhouse) after the seedlings had been transferred to the nutrient solutions. All plants from the growth chamber and half of the plants from the greenhouse were used to measure growth. Roots were stained in a solution containing 0.1% (w/v) methylene blue [3,7-bis(dimethylamino)phenothiazin-5-ium chloride] and 0.1% (w/v) neutral red [N8,N8 3-trimethyl-2,8-phenazinediamine monohydrochloride] for 24 h, and scanned with a flatbed scanner, which was also used to scan leaves. Root-image-analysis software (WinRHIZO, Regent Instruments Inc., Quebec, Canada) was used to determine total root length (RL) and leaf area (LA). The dry weights of roots (RDW) and shoots (SDW) were determined after drying plant material at 70°C for 2 d.
Half of the plants from the greenhouse, selected at random during harvest, were used to measure nutrient concentrations in tissues. Shoots and roots of individual plants were dried at 70°C for 2 d. They were then weighed and digested in a 2:1 mixture of 65% (w/w) HNO3 and 70% (w/w) HClO4 at 200°C for 2 h. The concentrations of Ca, Mg, K, and Al were determined, using atomic-absorption spectroscopy (Salinas and García, 1985).
Statistical Analysis
All growth data were log-transformed (Sokal and Rohlf, 1995) before being subjected to Model I ANOVA to test for interaction between the effects of Al toxicity and nutrient supply on plant growth. Relative growth parameters were computed using weighted jackknifing (Buonaccorsi and Liebhold, 1988), and compared with t tests, based on separate variance estimates if heteroscedasticity was detected with Levene's test. ANOVA of nutrient-concentration data was performed without data transformation, and differences between group means were tested for significance using the Spjotvoll/Stoline test.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Effect of Fertilizer Application
Nutrient concentrations in soil solutions from soil samples without fertilizer, extracted shortly after rewetting, were very low. Their ionic strength varied between 1.3 and 1.7 mM. Fertilizer application significantly increased the concentrations of cationic macronutrients (micronutrients were not measured). This effect was more pronounced in the sandy loam. The cation-activity ratio, that is, -log [{K+}/({Ca2+} + {Mg2+})0.5], was between 1.9 and 2.2 in soil solutions of samples without fertilizer (Moss, 1963). Fertilization decreased these values to between 1.2 and 1.3 (sandy loam) or between 1.5 and 1.6 (clay loam).
Fertilizer application also increased the levels of anions in soil solutions, except phosphate, the increase being, again, more pronounced in the sandy loam. In soil solutions extracted from the clay loam, SO2-4 levels were one to two orders of magnitude lower than for solutions from the sandy loam, even if fertilizer had been added.
Soil-Solution pH and Nitrogen Mineralization
The pH of soil solutions increased by 1 to 1.5 units during the first 16 d after rewetting, and dropped to a similar extent between Days 16 and 64. The amplitudes of pH fluctuations were larger in the sandy loam than in the clay loam (Fig. 1A)
. The acidification of soil solutions from Day 16 onward was clearly associated with the onset of nitrification, which led to a pronounced accumulation of NO-3, particularly in soil solutions from sandy loam + fertilizer (Fig. 1B). By contrast, ammonium concentrations in soil solutions remained virtually constant throughout the incubation of soil samples. They were higher in the soil samples + fertilizer, presumably because the urea in the fertilizer was partly mineralized (Fig. 1C).
|
|
The acidification associated with nitrification increased the solubility of Al. A plot of log {Al3+} versus pH suggested that Al solubility was controlled by a form of gibbsite that was slightly more soluble than the crystalline form, which often puts an upper limit to Al solubility in acid soils (Lindsay and Walthall, 1989). Only soil solutions extracted shortly after rewetting were significantly undersaturated with respect to crystalline gibbsite (Fig. 4) . There was no evidence that kaolinite decreased Al solubility below that of gibbsite.
The activities of H2PO-4 in soil solutions indicated that variscite in equilibrium with gibbsite imposed an upper limit to P solubility (Barber, 1984). Only shortly after rewetting, H2PO-4 activities exceeded significantly this solubility limit (Fig. 4).
Growth of Brachiaria Seedlings in the Designed Nutrient Solutions
Based on the composition of soil solutions analyzed in the first part of this study, a treatment was designed to simulate, in liquid culture, the chemical properties of the highly weathered acid soils of the Colombian savannas (LN + Al, low nutrient levels plus Al; Table 1).
Effect of the Treatment Simulating Chemical Properties of Savanna Soils
Dry matter production of signalgrass and ruzigrass under unstressed conditions was 152 and 180 mg plant-1 in the greenhouse and 52.8 and 85.8 mg (3 plants)-1 in the growth chamber, respectively (AN, adequate nutrient levels; Table 2). The effect of the LN + Al treatment, simulating soil solutions of the Colombian savanna soils, was evaluated by quantifying growth relative to the AN treatment. The less-adapted ruzigrass was considerably more affected than the well-adapted signalgrass, irrespective of whether plants were grown in the greenhouse or growth chamber (Fig. 5)
. Statistically significant interspecific differences were observed for the relative SDW and RDW, LA, and RL (P < 0.05 to P < 0.001 for the various growth parameters).
|
The analysis of variance of growth data from the plants grown in the greenhouse, however, indicated a statistically significant interaction between the effects of Al toxicity and nutrient deficiency in ruzigrass but not in signalgrass (Table 4). Corresponding data obtained from the plants grown in the growth chamber displayed essentially the same trend, although they were complicated by a slightly stimulatory effect of Al on root growth of signalgrass under low nutrient supply (data not shown). This interaction was associated with an alteration of the morphology and architecture of the root system of ruzigrass. Roots of plants simultaneously exposed to both stress factors were thicker and had more laterals close to the tip of the main axis, both typical symptoms of Al injury (Fig. 6) .
|
|
|
Effects on Nutrient Uptake
The mineral composition of shoots displayed a similar trend to that of roots (correlation coefficients for shoot vs. root values ranged from 0.52 for K to 0.95 for Mg; computed with data from both grasses and all treatments). The only exception to this trend was Al, which accumulated in the roots of both grasses without being translocated to shoots. The inhibitory effects of Al toxicity and nutrient deficiency on nutrient acquisition were therefore evaluated based on nutrient concentrations in shoots, which were unaffected by the contributions of ions passively adsorbed to cation-exchange sites in root cell walls. An ANOVA showed that a low supply of nutrients during growth reduced uptake of N, P, and K in both grasses to a similar extent (Table 5). In the case of Ca and Mg, this inhibitory effect was amplified by Al toxicity, as indicated by the statistically significant interaction between the effects of nutrient supply and Al toxicity (Table 5). No difference in nutrient acquisition between the two grasses was detected.
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Adding fertilizer increased the ionic strength of soil solutions from 1.3–1.7 mM to values similar or greater than those of soil solutions used to design low-ionic-strength nutrient solutions (5.4–13.4 mM) (Gillman and Bell, 1978; Edmeades et al., 1985; Bruce et al., 1989).
Rewetting Soils Stimulates Nitrification
The seasonal pattern of soil drying and rewetting induces microbially mediated fluctuations of pH and mineral N in tropical soils (Sánchez, 1976; Gigou et al., 1985; Sylvester-Bradley et al., 1988; Mueller-Harvey et al., 1989). Studies in rice fields in the Colombian savannas, for example, have shown that soil solutions become increasingly acidic over the course of the wet season (CIAT, 1994; Okada and Fischer, 2001).
After an initial rise in pH, perhaps caused by slow equilibration of the soils with the added water, we found a similar acidification of soil solutions, starting 2 to 4 wk after rewetting (Fig. 1A). This acidification was presumably caused by nitrification, well known for its acidifying effect, because it was associated with a substantial increase in the concentration of NO-3 (Fig. 1B). This suggests that rewetting the soil samples, which had been collected at the end of the wet season and subsequently air-dried, might have stimulated nitrification in a similar way as the onset of the wet season presumably does in the field.
The NO-3 produced during nitrification may not have been derived from ammonium, because ammonium levels in soil solutions did not decrease as nitrification set in (Fig. 1C). This finding is similar to the results of previous experiments with soil cores collected in the Colombian savannas (Sylvester-Bradley et al., 1988). Together, they are consistent with the suggestion that heterotrophic nitrifiers might have produced the NO-3 directly from organic substrates. Heterotrophic nitrification often dominates in acid soils because autotrophic nitrifiers are inhibited at low pH (Haynes, 1986).
Nitrification Affects Soil-Solution Composition
Before the onset of nitrification, cation levels in soil solutions were not correlated with that of NO-3 (see {NO-3} <1 mM; Fig. 3). Chloride appeared to balance most of the positive charges of cations because it was more abundant than NO-3 and SO2-4. Sodium was the most abundant cation in soil solutions from soils without fertilizer, but not in soils + fertilizer (Fig. 3). Dominance of Na has been reported for soil solutions of subsoils, but only occasionally for surface soils (Bruce et al., 1989).
The activities of other cations (before nitrification) were in the order of K+ 
Ca2+ > Mg2+, irrespective of soil type or fertilizer addition (Fig. 3). Yet cation-activity ratios in soil solutions from soils without fertilizer were higher than those from fertilized soils, indicating a comparatively low abundance of K relative to divalent cations in native soils (see Results).
The acidification associated with nitrification had a profound impact on the composition of soil solutions. Above a NO-3 activity of approximately 1 mM, cation activities increased in parallel to the activity of NO-3, the latter being a measure of the amount of protons produced by nitrification (Fig. 3). The stoichiometric relationships between the amounts of NO-3 and cations accumulating in soil solutions, and the approximately constant cation-activity ratios (see Results for a definition), together suggest that the protons produced during nitrification desorbed cations from pH-dependent cation-exchange sites in proportion to their abundance on the exchange phase (Fig. 2; Moss, 1963; Curtin and Smillie, 1995). A similar mechanism of cation release appears to exist in the field, because the progressive acidification of soil solutions in rice fields during the wet season, observed in a previous field study in the Colombian savannas, was also accompanied by a marked increase in cation concentrations (CIAT, 1994; Okada and Fischer, 2001).
These data suggest that nitrification may be an important factor affecting the soil-solution composition in these soils if levels of organic N and/or NH+4 are high enough. In contrast to our soil-incubation experiments, leaching and uptake of NO-3 by plants counteract the impact of seasonal peaks of nitrification on soil-solution composition in the field (Sánchez, 1976; Gigou et al., 1985; Sylvester-Bradley and Mosquera, 1985; Mueller-Harvey et al., 1989). In established pastures NO-3 salts are therefore not expected to accumulate to the same extent as in this study. The acidifying effect of nitrification, however, allowed us to establish that the solubilities of Al3+ and H2PO-4 were controlled by gibbsite and variscite (in equilibrium with gibbsite), respectively (Fig. 4; Barber, 1984; Lindsay and Walthall, 1989).
A Nutrient Solution Simulating Chemical Properties of Highly Weathered Acid Soils
One of the goals of this study was to design a nutrient solution that simulates soil solutions of highly weathered acid soils (LN + Al, low nutrient levels plus Al). We used the composition of soil solutions from soils without fertilizer before the onset of nitrification as a guideline to adjust the activities of Al3+ and macronutrients to values typical for the Colombian savanna soils. For micronutrients, the concentrations in the low-ionic-strength nutrient solution described by Blamey et al. (1991) were adopted. No attempts were made to adjust their concentrations to lower levels because maintaining such low levels would have required a flowing solution culture system. Manganese toxicity was not considered in this study because its occurrence depends on the soil's parent material, and other South American savanna soils have been shown to contain lower-than-world-average levels of Mn (de Sá e Melo Marques, 2000). Table 1 details the criteria used to design the LN + Al solution based on the results of the first part of this study.
The concentrations of most macronutrients in the LN + Al treatment are lower than in other low-ionic-strength nutrient solutions that have been used for Al-toxicity studies (Blamey et al., 1983, 1991; Wheeler et al., 1992; Edmeades et al., 1995). Plant growth is likely to be limited as a result of such low nutrient levels (Asher, 1978, 1986).
To accurately quantify the growth-inhibitory effect of the LN + Al treatment, growth has to be referenced against the plants' inherent growth potential, measured under unstressed conditions (AN, adequate nutrients without Al). Nutrient concentrations in soil solutions from soils + fertilizer were similar or greater than those in the low-ionic-strength nutrient solution designed by Blamey et al. (1991) to support close-to-maximum growth rates, while minimizing problems with Al speciation. Because it would be desirable to measure the effect of Al at adequate (but not excessive) nutrient levels, we did not adjust nutrient concentrations of the AN solution above their levels in that solution. Only the concentration of P was raised from 2 to 5 µM to compensate for depletion of such a small quantity in intermittently renewed solution culture (Table 2). Previous greenhouse experiments at variable P and N levels, but otherwise identical conditions, had confirmed that their concentrations in the AN solution resulted in 84 to 94% of maximum yield for the two grasses (Lasso, 1998).
To disentangle the effects of Al toxicity and nutrient deficiency, we also designed two solutions that exhibited either an identical activity of the principal phytotoxic Al species (Al3+; AN + Al) or identical nutrient concentrations (LN) as the LN + Al treatment (Tables 2 and 3).
The Growth Response of two Brachiaria Species in the Designed Nutrient Solution Mirrors the Difference in Forage Yield in the Field
To test the effectiveness of the LN + Al treatment in simulating the chemical properties of the Colombian savanna soils, we evaluated growth of two contrasting Brachiaria species relative to unstressed conditions ([LN + Al]/AN). Results showed that the acid soil-adapted signalgrass was significantly less affected by simultaneous Al stress and nutrient deficiency than was the less-adapted ruzigrass. This growth difference was similar to the differences in forage yield measured in a previous field evaluation at the Pista site, where the samples of clay loam had been collected for this study (CIAT, 1995).
Comparable results were obtained whether plants were grown in the greenhouse or growth chamber, suggesting that physiological responses to this treatment were reproducible between very different growth regimes (Fig. 5). Hence, this technique should be a robust method to identify Brachiaria genotypes adapted to infertile acid soils. Relative growth of B. brizantha cv. Marandú, a Brachiaria species of intermediate edaphic adaptation, was intermediate between the other two species in all experiments (data not shown). These data are consistent with the suggestion that the LN + Al treatment provides a realistic approximation to the stress factors that limit plant growth in the highly weathered acid soils of the Colombian savannas.
Nutrient Deprivation Affects Aluminum Resistance of the Less-Adapted Brachiaria Species
Having tentatively established the effectiveness of the LN + Al treatment in simulating chemical factors that limit plant growth in these savanna soils, we dissected the stress syndrome into two of its components. This was accomplished by evaluating the relative effects of nutrient deficiency and Al toxicity in separate treatments (LN/AN and [AN + Al]/AN, respectively).
A low supply of nutrients reduced growth in both signalgrass and ruzigrass to the same extent (see Results). A similar result was obtained for Al toxicity, except for a moderate difference in the dry weight and total length of roots of plants grown in the greenhouse (Fig. 7B). The pronounced growth difference between the two grasses under simultaneously Al-toxic and nutrient-deficient conditions, therefore, was the result of an interaction between the effects of the two stress factors on growth of less adapted ruzigrass (Fig. 5, Table 4). In principle, either inhibition of nutrient uptake by Al or an increase in Al susceptibility under nutrient deprivation could account for this interaction. The distinct root morphology and architecture of ruzigrass grown in the LN + Al solution, reminiscent of the effects of Al toxicity, suggested that its level of Al resistance was affected by the low nutrient concentrations in the medium (Fig. 6).
We therefore compared Al susceptibility of the grasses at the two different levels of nutrient supply. Only under nutrient-limited growth conditions did a clear interspecific difference in Al resistance appear (Fig. 7A). Again, B. brizantha cv. Marandú ranked between the other two Brachiaria species, in accordance with its intermediate level of edaphic adaptation (data not shown). This increase in Al sensitivity under low nutrient supply may help explain the poor persistence of ruzigrass in South American savannas. It may also explain why a root-elongation assay in simple salt solutions had revealed a marked difference in Al resistance between signalgrass and ruzigrass, although we could not detect this difference in a previous experiment with fertilized soil samples exhibiting varying degrees of Al saturation (Rao et al., 1996; Wenzl et al., 2001).
To identify a particular nutrient, whose deficiency could account for an increased sensitivity of ruzigrass to Al, we analyzed nutrient uptake by plants. There was no difference between the two grasses in the way nutrient deficiency and Al-toxicity inhibited nutrient acquisition (Table 5). This appears to indicate that the adverse effect of nutrient deficiency on Al resistance of ruzigrass was not due to an inefficient uptake of nutrients. Instead, it may have been the result of a less effective use of scarce nutrient(s) in physiological processes that prevent or alleviate Al phytotoxicity.
We therefore identified the nutrients whose tissue concentrations decreased most strongly when nutrient deficiency was superimposed on Al stress, because a lower tissue concentration is likely to affect the efficiency with which a particular nutrient can be employed to counteract Al phytotoxicity. As shown in Fig. 8, the uptake of P and Mg was affected the strongest. Differences in the way either of these nutrients is employed to counteract Al toxicity may therefore be responsible for the marked difference in Al resistance between the two grasses under nutrient-deficient growth conditions.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The poorer growth of less-adapted ruzigrass in this solution appeared to be the result of an insufficient capacity simultaneously to counteract Al toxicity and nutrient deficiency, even though these stresses could be successfully dealt with separately. Clearly, experimental isolation of individual stress components of highly weathered acid soils is a prerequisite for mechanistic studies. The data presented, however, highlight the need for a complementary holistic approach that conceptualizes adaptation to infertile acid soils as a coordinated physiological adjustment to several stress factors. Analysis of growth in the designed nutrient solution may serve to rapidly test this holistic aspect of a plant's adaptive capacity.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication June 12, 2002.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
