SURFACE APPLICATION OF LIME AMELIORATES SUBSOIL ACIDITY AND IMPROVES ROOT GROWTH AND YIELD OF WHEAT IN AN ACID SOIL UNDER NO-TILL SYSTEM

Crop root growth and grain yield can be affected by chemical modifications in the soil profile due to surface lime application. A field trial was carried out on a loamy dystrophic Typic Hapludox at Ponta Grossa, State of Paraná, Brazil, to evaluate root growth and grain yield of wheat (Triticum aestivum L. cv. CD 104, moderately susceptible to Al), about 10 years after surface liming (0, 2, 4, and 6 Mg ha) and three years after surface re-liming (0 and 3 Mg ha), in a long-term no-till cultivation system. Soil acidity limited wheat root growth and yield severely, probably as a result of extended water deficits during the vegetative stage. Surface liming caused increases up to 66% in the root growth (0–60 cm) and up to 140% in the grain yield. Root density and grain yield were correlated positively with soil pH and exchangeable Ca, and negatively with exchangeable Al and Al saturation, in the surface and subsurface layers.


INTRODUCTION
Soil acidity frequently affects agricultural production in Brazil.The low content of calcium (Ritchey et al., 1982) and aluminum toxicity (Pavan et al., 1982) affect root growth and the absorption of water and nutrients by plants, usually causing reduction of crop yields in acid soils (Sumner et al., 1986).
Liming is the most efficient practice in order to solve problems related to soil acidity.To control soil acidity in a no-till (NT) system, lime is broadcast on the surface without incorporation.Since the lime reaction is generally limited to the layer it was applied, acidity at the subsurface layers, in case of toxic levels of aluminum and/or calcium deficiency, can compromise root penetration and plant nutrition.Results of field studies show that the movement of lime to depth varies according to timing and rates of liming, lime application forms, soil type, weather conditions, addition of acidic fertilizers, and cropping systems (Oliveira & Pavan, 1996;Caires et al., 2000;Rheinheimer et al., 2000;Gascho & Parker, 2001;Conyers et al., 2003;Pires et al., 2003;Tang et al., 2003).
Wheat is economically one of the most important winter crops in agricultural activities in South Brazil.Although there are genotypes with distinct reactions to soil acidity (Muzilli et al., 1978;Costa et al., 2003;Tang et al., 2002), there is a lack of information on the implications of soil acidity and surface liming on wheat root growth and grain yield under NT cultivation.
Root growth and grain yield of wheat were evaluated in an acid soil after surface application of lime rates and re-liming in a long-term NT cultivation system.

MATERIAL AND METHODS
The experiment was carried out in Ponta Grossa, PR, Brazil (25º10'S, 50º05'W), on a dystrophic Typic Hapludox.At the beginning of the experiment, soil chemical and texture analyses of the 0-20 cm layer presented the following results: pH (1:2.5 soil: 0.01 mol L -1 CaCl 2 suspension) of 4.5; exchangeable Al 3+ , Ca 2+ , Mg 2+ , and K + contents of 6, 16, 10, and 1.4 mmol c dm -3 , respectively; total acidity pH 7.0 (H + Al) of 58 mmol c dm -3 ; P (Mehlich-1) of 9.0 mg dm -3 ; total organic matter of 33 g dm -3 ; base saturation of 32%; Al 3+ saturation of 18%; and 295, 240, and 465 g kg -1 of clay, silt, and sand, respectively.Prior to the establishment of the experiment, the field site had been used for grain cropping under the NT cultivation system during 15 years.
A randomized complete block design was used, with three replications in a split-plot arrangement.The main plots (8.0 m × 6.3 m) consisted of surface dolomitic lime at the rates of 0, 2, 4, and 6 Mg ha -1 .The lime rates were calculated to raise the base saturation in the topsoil (0-20 cm) to 50, 70, and 90%.The dolomitic lime used contained 176 g kg -1 Ca, 136 g kg -1 Mg, and 84% effective calcium car-bonate equivalent (ECCE), and was broadcast on the soil surface in July 1993.In June 2000, the main plots were divided in two subplots (4.0 m × 6.3 m) for the study of surface re-liming influence (196 g kg -1 Ca, 130 g kg -1 Mg, and 90% ECCE) at the rates of 0 and 3 Mg ha -1 .The reapplied rate was calculated to raise the base saturation in the topsoil (0-20 cm) to 65% (Caires et al., 2000) of the treatment 4 Mg ha -1 of lime made in July 1993 (pH 0.01 mol L -1 CaCl 2 of 4.6; CEC pH 7.0 of 110.8 mmol c dm -3 ; and 41% of base saturation).
Wheat, cv.CD 104 (moderately susceptible to the soil exchangeable aluminum, with tolerance to Al 3+ saturation between 5 and 20% at the 0-20 cm depth, according to IAPAR, 2003) was sown on June 14, 2003, with 0.17 m between rows and 140 kg of seed per hectare, for a population between 250 and 300 plants m -2 .During wheat sowing no fertilizer was applied due to the residual effect of the fertilizer applied (24 kg ha -1 P and 46 kg ha -1 K) to the prior crop (soybean).The plants emerged on June 21, 2003 and were counted on June 28, 2003.The initial average population was 255 plants m -2 , without effect of the liming treatments.Fertilizers were applied by top dressing (August, 05, 2003) at rates of 80 kg ha -1 N and 35 kg ha -1 K, as ammonium nitrate and potassium chloride, respectively.
Wheat plants flowered fully in September 13, 2003, 84 days after emergence, and the crop maturation occurred in October 25, 2003, 126 days after emerging.
The air temperature was adequate for the wheat to develop and there was considerable rainfall before sowing (55 mm) and before the plant flowering (56 mm).However, there was an extended water deficit during the vegetative development stage (Figure 1).Rainfall was 434 mm during the wheat crop cycle.
Samples of wheat roots were collected during the flowering period, in September 20, 2003, by means of a sampling tube of 3.5 cm diameter, at the depths of 0-10, 10-20, and 20-60 cm.Six sub-samples (3 from the sowing row and 3 between rows) were taken randomly in the sub-plot to form a composite sample.The roots were separated from the soil by dispersion in water through a 0.5 mm mesh sieve.Root length was estimated by the method of Tennant (1975), in a 1 × 1 cm grid.The average root radius was calculated by the following equation (Schenk & Barber, 1979): r = (mu πL -1 ) ½ , where mu is the wet mass and L, the root length.
After maturation, the wheat grain was harvested from 6.12 m 2 plots (middle 12 rows by 3 m in length).Grain was processed in screen for the amount of middling to be removed and measured for each subplot.The values of wheat middling obtained and its proportion to the production were used in the treatment evaluation.Grain yield was expressed at the 130 g kg -1 moisture content.
The results were submitted to variance and polynomial regression analyses.Regression equations were adjusted to the obtained data according to lime rates, adopting as criteria for the model choice the agronomic meaning and the magnitude of coefficients of determination significant at 5%.The effects of re-liming were compared through the F-test.

RESULTS AND DISCUSSION
Liming treatments caused varied soil acidity conditions for plant development, as intended (Table 1).Soil chemical analyses performed according to the methods recommended by Pavan et al. (1992), before sowing wheat, revealed increases of pH, exchangeable Ca 2+ and Mg 2+ , and base saturation, and also the reduction of exchangeable Al 3+ and Al 3+ saturation, in the surface and subsurface layers, with surface liming.Considering the rates and the re-liming, soil pH 0.01 mol L -1 CaCl 2 varied from 4.4 to 6.1, in the 0-5 cm layer, from 4.2 to 5.3, in the 5-10 cm layer, and from 4.2 to 4.8, from 10 to 60 cm depth; the exchangeable Al 3+ varied from 0 to 7 mmol c dm -3 , from 0 to 11 mmol c dm -3 , and from 2 to 10 mmol c dm -3 , respectively in at depths of 0-5, 5-10, and 10-60 cm; and the variation in the Al 3+ saturation was from 0 to 20% (0-5 cm), from 0 to 34% (5-10 cm), and from 5 to 35% (10-60 cm).
The traits of wheat root growth were not significantly influenced by the interaction between lime rates and re-liming.So, the results regarding the growth of the root system were analyzed separately by the means of the observations.
Increasing surface liming rate (x, in Mg ha -1 ) increased linearly wheat root length per unit surface area ( , in cm cm -2 ) to the depth of 60 cm ( = 88.36 + 8.46x, R 2 = 0.92).Re-liming also increased significantly the root length of wheat to the depth of 60 cm Table 3 -Total root length per unit surface area in the 0-60 cm layer, and relative root length of wheat in various layers as affected by surface liming application and surface re-liming at the rate of 3 Mg ha -1 .
(Table 3).The increase in the root length with liming was up to 66%.However, the relative wheat root length in the soil profile was not affected by surface liming.There was higher concentration of the wheat root length at 0-10 cm depth (62-66%), independently of liming.The relative length of wheat roots was about 75% at the 0-20 cm depth, which agrees with results of other studies (Rosolem et al., 1992;Caires et al., 2002a, Tang et al., 2002;2003).The correlation analysis between the traits of wheat root growth and the soil chemical attributes (Table 4) shows that the root growth was favored mainly by the increase of pH, exchangeable Ca 2+ and base saturation, and by the reduction of exchangeable Al 3+ and Al 3+ saturation in the soil, due to surface liming (Table 1).The exchangeable Al 3+ and the Al 3+ saturation were the soil attributes that presented the closest correlations with the root growth per unit surface area to the depth of 60 cm, showing that there was aluminum toxicity for wheat root growth.The aluminum in toxic concentrations causes reduction in root growth rate, but the reaction of genotypes to the aluminum in acid soils is variable (Muzilli et al., 1978;Costa et al., 2003;Tang et al., 2002;2003).These results differ substantially from those obtained for soybean (Caires et al., 2001) and corn (Caires et al., 2002b;Tissi et al., 2004), in similar conditions of soil acidity under NT.In these studies, concentrations of exchangeable Al 3+ in the soil from 8 to 11 mmol c dm -3 did not cause limitation to root growth for corn and soybean, both cultivated in the absence of water deficit under NT.Although plant genotypes differ in their susceptibility to soil acidity, Muzilli et al. (1978) observed practically that wheat cultivars were twice more tolerant to soil acidity (Al 3+ saturation) than the soybean cultivars.
Studies have shown smaller toxicity of aluminum in systems in which crop residues are left on the soil surface (Miyazawa et al., 1993;Franchini et al., 1999;Salet et al., 1999).The higher content of organic matter (Bayer et al., 2000;Rhoton, 2000) and the greater nutrient concentration on the soil surface (Rhoton, 2000) under NT may reduce aluminum toxicity through the formation of Al-organic complexes and the greater ionic strength of the soil solution (Salet et al., 1999).However, this low toxicity of alu-minum for root growth of plants under NT (Caires et al., 2001;2002b;Tissi et al., 2004) should only occur in the absence of water deficit.In this study, although the rain episodes occurred in adequate intensity soon after sowing and at the beginning of wheat flowering, rainfall was irregular during the crop cycle -long periods with water limitation were observed during the vegetative development stage (Figure 1).The aluminum phytotoxicity under NT seems to be quite dependent on the water regime occurring during the growth cycle of the plants in the field.In unfavorable conditions of rainfall, the phytotoxicity of Al 3+ due to high soil acidity is intensified and compromises severely root growth of plants with little tolerance to aluminum.
Wheat grain yield and middling proportion were significantly influenced by the interaction between lime rates and re-liming (Figure 2).Grain yield of wheat increased linearly with the application of lime on the soil surface, with or without surface reliming (Figure 2a).Re-liming increased grain yield of wheat only for the smaller rates of lime (0 and 2 Mg ha -1 ), when the soil presented higher acidity (Table 1).According to the adjusted regression equations, the increases in grain yields were 380 and 200 kg ha -1 per Mg of lime, corresponding to a yield increase of 130% and 40% with the highest lime rate applied (6 Mg ha -1 ), respectively without and with re-liming.
Re-liming at the rate of 3 Mg ha -1 , on plots that had not received prior liming, increased wheat grain yield by 72%.Wheat grain yield was 140% higher in the treatment that received 6+3 Mg ha -1 of lime than in the control treatment without liming.Surface application of lime rates reduced wheat middling proportion only in the absence of re-liming (Figure 2b).Reliming at the rate of 3 Mg ha -1 also reduced wheat middling proportion but only in the absence of prior liming, under more acid soil conditions.These results showed, unquestionably, that surface application of lime to control soil acidity under NT increased yield and quality of wheat grain.
The increase in wheat grain yield with surface liming was related significantly with the increase of soil pH, exchangeable Ca 2+ and base saturation, and with the reduction of exchangeable Al 3+ and Al 3+ saturation, at the four studied depths (Table 5).Wheat response to liming soils was observed where the Al 3+ saturation was more than 25% (Winkler & Sfredo, 1991).In another trial using a range of wheat genotypes, the increase in wheat grain yield was also related to the reduction in soil aluminum saturation (Costa et al., 2003).Therefore, the improvement in soil acidity conditions at the surface and in subsurface layers by surface liming application under NT had positive effects on wheat grain yield.
A series of studies have shown high wheat, corn and soybean grain yields in acid soils under NT and little response of these crops to surface liming (Pöttker & Ben, 1998;Caires et al., 1999;Rheinheimer et al., 2000;Caires et al., 2002a;2002b).The explanations for this behavior has been related to the following inherent factors to this cropping system: (i) pH increase and reduction of the Al 3+ content by water soluble organic compounds from vegetable materials that are left on the soil surface to serve as a mulch; (ii) reduction of the Al 3+ activity in solution due to the higher organic matter content and the higher nutrient concentration in the soil surface, and (iii) sufficient exchangeable cation concentrations, even in soils of high acidity, due to the increase in the cation exchange capacity because of the higher content of organic matter.However, these data were obtained of field experiments performed without the occurrence of water limitation.In this study, there was marked response of wheat to surface liming, probably due to an increased concentration of Al 3+ in soil solution as a result of extended water deficit during the vegetative development stage, showing that the response of the crops to liming under NT seems to depend on the water regime that occurs during the growing cycle of the plants.

ACKNOWLEDGEMENTS
To CNPq and to CAPES for providing scholarships to the authors.

Figure 1 -
Figure 1 -Maximum and minimum air temperature and rainfall throughout the development period of the wheat crop in 2003.

Figure 2 -
Figure 2 -Wheat grain yield (a) and middling proportion (b) as affected by surface liming application, without ( ) and with ( ) surface re-liming at the rate of 3 Mg ha -1 .Wheat middling proportion = (amount of middling/total grain production) × 100.Bars represent the least significant difference (LSD) by Tukey test at P = 0.05.*:significant at P < 0.05 and **:significant at P < 0.01.

Table 1 -
Soil chemical attributes for different depths before sowing wheat, about 10 yr after surface liming application and 3 yr after surface re-liming at the rate of 3 Mg ha -1 .

Table 2 -
Wheat root length density and mean radius as affected by surface liming application and surface re-liming at the rate of 3 Mg ha -1 .

Table 4 -
Coefficients of correlation between wheat root growth traits and soil chemical attributes for different soil layers.: non-significant, *: significant at P < 0.05, and **: significant at P < 0.01. ns

Table 5 -
Coefficients of correlation between wheat grain yield and soil chemical attributes for different soil layers.: non-significant, *: significant at P < 0.05, and **: significant at P < 0.01. ns