BORON EXTRACTION AND VERTICAL MOBILITY IN PARANÁ STATE OXISOL, BRAZIL(1)

The deficiency or excess of micronutrients has been determined by analyses of soil and plant tissue. In Brazil, the lack of studies that would define and standardize extraction and determination methods, as well as lack of correlation and calibration studies, makes it difficult to establish limits of concentration classes for analysis interpretation and fertilizer recommendations for crops. A specific extractor for soil analysis is sometimes chosen due to the ease of use in the laboratory and not in view of its efficiency in determining a bioavailable nutrient. The objectives of this study were to: (a) evaluate B concentrations in the soil as related to the fertilizer rate, soil depth and extractor; (b) verify the nutrient movement in the soil profile; (c) evaluate efficiency of Hot Water, Mehlich-1 and Mehlich-3 as available B extractors, using sunflower as test plant. The experimental design consisted of complete randomized blocks with four replications and treatments of five B rates (0, 2, 4, 6, and 8 kg ha-1) applied to the soil surface and evaluated at six depths (0–0.05, 0.05–0.10, 0.10–0.15, 0.15–0.20, 0.20–0.30, and 0.30–0.40 m). Boron concentrations in the soil extracted by Hot Water, Mehlich-1 and Mehlich-3 extractors increased linearly in relation to B rates at all depths evaluated, indicating B mobility in the profile. The extractors had different B extraction capacities, but were all efficient to evaluate bioavailability of the nutrient to sunflower. Mehlich1 and Mehlich-3 can therefore be used to analyze B as well as Hot Water.


INTRODUCTION
Knowledge on the soil dynamics of a nutrient is essential to understand its availability in the soil-plant system.Boron availability is related to soil characteristics such as pH, mineralogical composition, texture and organic material.Valladares et al. (1999) found a negative correlation between available B and increase in soil pH in 103 soil samples from Rio de Janeiro.Similarly, B adsorption reached a maximum at pH 7.0 and the concentration available in solution diminished from surface to sub-surface soil samples with different textures.With the sub-surface movement of B, there was a positive correlation between nutrient absorption and gibbsite concentration (Soares et al., 2008).Silva et al. (1995) evaluated B leaching in a dystropherric Red Latosol (typic Hapludox) under cotton, with an average clay content of 500 g kg -1 with successive annual applications over eight years; the authors verified that the nutrient was displaced as deep as 60 cm and the quantity extracted (Hot Water) was associated with the rate applied.
Levels of available B in soil can also be related to the extraction method.The results of extractors and methods used to evaluate the micronutrient availability vary greatly.This was mainly due to the chemical composition of the extractors as well as soil characteristics, which influence nutrient dynamics and extraction efficiency directly (Abreu et al., 1994).
There are various B extraction methods to evaluate plant availability.However, the results of some methods are better correlated with the plants.Among the extractors are Hot Water, proposed by Berger & Truog (1939), saline solutions (CaCl 2 , BaCl 2 ) proposed by Abreu et al. (1994), Mehlich-1 acid solutions, proposed by Mehlich (1953), HCl, proposed by Tedesco et al. (1985) and complexing solutions , according to Mehlich (1984).
The recommendations of fertilizers for various States of Brazil are normally based on the correlation and calibration of soil concentrations with relative plant yields.The search for experimentally tested and proven methods becomes necessary when reliable results are desired, since the success of fertilizing practices depends largely on precise soil chemical analyses (Sfredo et al., 1984).
The objectives of this study were to evaluate: soil B concentrations as related to the fertilizer rate applied at each soil depth and by extractor type; B movement in the soil profile; and the efficiency of the Hot Water, Mehlich-1 and Mehlich-3 by the correlation of B content in sunflower and B concentrations obtained by the different methods.

MATERIALS AND METHODS
The study was carried out on an experimental farm of the Brazilian Agricultural Research Center (Embrapa) in Londrina, State of Paraná.In a homogeneous area of approximately 2.0 ha, ten samples of the soil classified as eutropherric Red Latosol (Rhodic Eutrudox) were collected at different depths between the sunflower rows.The physical and chemical properties were determined for each depth prior to the experiment (Table 1).
Granulometric analysis was determined by sedimentation, using the pipette method with 1.0 mol L -1 sodium hydroxide as dispersant, following the Embrapa method (Embrapa, 1979).
The sunflower hybrid M 732 (Helianthus annuus L.), planted in September 2002 and harvested in January 2003, was used as test plant for analysis of absorbed B. The crop was fertilized at planting with 300 kg ha -1 of 5-20-20 NPK fertilizer, applied in furrow and 45 kg ha -1 of N as sidedressing (ammonium sulfate), applied 25 days after emergence (DAE).
Sunflower leaves were collected at the beginning of flowering, stage R 4 /R 5 , according to the identification method of the sunflower development phases described by Scheneiter & Miller (1981).Leaves were randomly collected from the crop, with an average of 20 leaves per hectare, making up a composite sample.Then the leaves were washed with distilled water, ovendried at 50 °C to constant weight and later ground in a Wiley mill.Leaf B concentrations were evaluated by dry digestion at 550 °C and quantified by inductively coupled plasma-optical emission spectrometry.
After harvesting the of sunflower, soil samples were collected and homogenized, dried at 30 °C and passed through a 2 mm sieve.The soil samples were collected with a core sampler, according to the treatments at different depths, between the sunflowers rows to determine B concentrations by the different extractors.
The variables were subjected to variance and regression analysis.For each depth, the regression equations were adjusted to the concentrations extracted and the applied rates.The identities of the extractors were verified according to Leite & Oliveira (2002) to zero, β 1 is equal to 1, and the value of r is close to 1.The identity was evaluated by Hot Water vs. Mehlich-1, Hot Water vs. Mehlich-3 and Mehlich-1 vs. Mehlich-3.In order to evaluate the efficiency of the extractors, regression equations were adjusted between the estimated concentrations for Hot Water, Mehlich and Mehlich-3 and the B concentrations in sunflower leaves.

Boron concentration in relation to soil depth
At each soil sampling depth, B concentrations were different, according to the fertilizer rates and extractor, with interaction between them (p < 0.01, Table 2).
Boron concentrations in soil estimated by Hot Water, Mehlich-1 and Mehlich-3 increased linearly with B rate applied to the soil surface at each depth evaluated, indicating that the nutrient was mobile in the soil (Figure 1).Boron movement in soil was closely related with the B soil concentration and rate applied, as observed by Rosolem & Biscaro (2007).These authors added 5 kg ha -1 B as boric acid to the surface and verified that B concentrations present in water percolated through the soil column were 10 times as high as in treatments without B application.
In acidic soil, B in solution occurs predominantly as H 3 BO 3 , a weak acid that has a very low dissociation rate that makes it highly leached in the soil (Tanaka et al., 1993;Silva et al., 1995;Quaggio et al., 2003).Thus, of the total B applied, a part was moved in the form of H 3 BO 3 .Silva et al. (1995) studied B leaching in an Ultisol under cotton and reported that the nutrient reached a depth of 60 cm after nine years of annual borax application in the sowing furrow.While B mobility is favored by acidic conditions, the movement of the element in deeper layers is related to the application time and rate, initial concentration of the element (Rosolem & Biscaro, 2007), quantity of percolated water (Patil et al., 1997;Communar & Keren, 2006, 2007) and magnitude of soil absorption.Azevedo et al. (2001) affirm that B absorption directly correlates with organic material content, specific clay surface, presence of kaolinite, and exchangeable Al.
Boron concentrations decrease with greater depth at all rates applied (Figure 1).Higher B concentrations, from the soil surface to a depth of 0.20 m, are associated with H 3 BO 3 fertilization, a form that moves easily.On the other hand, there are higher concentrations of organic material, which plays an important role in B adsorption and at the same time represents an important natural source of the nutrient (Elrashisdi & Connor, 1982;Oliveira Neto et al., 2003).Boron adsorption directly correlates with soil organic material, which is greater at the soil surface (Azevedo et al., 2001).Depressive effects are registered for organic material, in B extraction with Hot Water (Ferreira et al., 2001) as well as by acidic extraction (Lima et al., 2007).However, it is known that in conditions with more organic material, these effects will have a higher magnitude in the Hot Water extractor.

Table Results of variance analysis of B in soil according to fertilizer rate (borax) and extractors (Hot
Water, Mehlich-1 and Mehlich-3), at each sampling depth (**) significant at 1.0 % using SAS statistical program (SAS, 2000).
In the layer 0.20-0.40m, B concentrations were low.At this depth, Bataglia & Raij (1990) and Silva & Ferreira (1998) also reported that Mehlich-1 extracted less B than the Hot Water extractor in soil with low concentrations of the nutrient in a more sandy texture.With increasing depth, the contribution of organic material diminished, with increasing importance of the inorganic fraction, which increases forms of inorganic B. Under these conditions, the acidity of the extractors encouraged positively charged oxide formation, increasing adsorption of B (Azevedo et al., 2001).
While acidic conditions encourage B mobility, B movement in depth will be a function of applied rate, application date, initial concentration of the element in the soil (Rosolem & Biscaro, 2007), and quantity of water percolated (Patil et al., 1997;Communar & Keren, 2006, 2007).
Boron moves in soil due to diffusion processes and by descending flow, moving along with water in percolation.The movement is therefore controlled by physical processes characterized by the gradient of water concentration and B concentration (in function of the rate applied).Still, B movement in soil is mainly controlled by physical-chemical processes, which are determined by the adsorption/desorption phenomenon.The desorption rate is positively related to the rate of water percolation (Patil et al., 1997;Communar & Keren et al., 2006, 2007) and inversely with the concentration of clay and organic material content in the soil (Zerrari et al., 2001).The lower B concentrations below 0.20 m reflect the interaction of factors that influence its movement.
When correlated, the coefficients between Hot Water, Mehlich-1 and Mehlich-3 extractors were high (Figure 2).They differed in extraction capacity according to B rates and depths, according to the extractor identity test proposed by Leite & Oliveira (2002); except when relating Mehlich-1 vs. Mehlich-3 and Hot Water vs. Mehlich-3, in the layer 0.10-0.15m and 0.15-0.20 m, respectively (Table 3, Figure 1).The results did not agree with those reported by Abreu et al. (1997), which refer to the similarity of extraction of the three extractors.However, it is worth remembering the difference in criteria used by the authors, since the criteria proposed by Leite & Oliveira (2002) to establish identities of the extractors are more rigorous.
Boron concentrations found in soil were within the critical limits for the nutrient (0.40 to 2.30 mg dm -3 ),   The extractor correlation coefficients observed down to the depth of 0.20 m (Table 3) were high (> 0.90) and significant, at 1.0 %.From this depth downwards, coefficients diminished, with no registered significance between the Mehlich-1 and Mehlich-3 extractors at the two depths evaluated (0.20-0.30m and 0.30-0.40m) nor between Hot Water and Mehlich-3 at 0.30-0.40m.However, considering the identity criteria proposed by Leite & Oliveira (2002), the extractors Mehlich-1 and Mehlich-3 were shown to be different from Hot Water extraction.Therefore, extractor use viability criteria should complete the correlation between concentrations extracted and plant growth measurements.

Relationship between B in soil B absorbed by sunflower
When evaluating the effect of B rate applied to the soil on B concentration in sunflower leaves, a significant and positive interaction was verified.There was a quadratic effect (p < 0.001) with a peak at the B rate of 7.73 kg ha -1 resulting in to 68 mg kg -1 of B in leaves (Figure 3).However, the foliar B concentration in the control was 47 mg kg -1 , a value considered appropriate for sunflower cultivation (Sfredo et al., 1984;Blamey et al., 1997).
Pearson's correlation coefficients for soil B concentration at all depths and sunflower leaf B concentrations were highly significant, indicating strong dependence between the variables (Table 4).Alvarez V. (1995) registered that an extractor is considered adequate when variations in concentrations extracted correspond to variations in quantities absorbed.In this context, the extractors proved to be sensitive in evaluating B concentrations available to sunflower at the different depths according to the rate applied.
Other studies register differences between acidic and Hot Water extractors in evaluating available B. Renan & Gupta (1991) verified greater correlation between B concentrations obtained with acidic extractors and B concentration in leaves of four plant species, compared to B extracted with hot water.The correlation between B extracted by acidic solution and B foliar concentration was better than that for B extracted by Hot Water, according to Ponnamperuma et al. (1981) in rice, and Souza Lima et al. (2007) in maize.Bataglia & Raij (1990) verified lower correlation coefficients for B extracted by Mehlich-1 and B absorbed by sunflower when compared to results for Hot Water; similarly, Ribeiro & Tucunango Saraiba (1984) obtained a correlation between B in the soil and B in sorghum of 0.65 for Hot Water and 0.58 for Mehlich-1.
According to the identity criteria proposed by Leite & Oliveira (2002), the extractors are different, although they proved to be effective to analyze B available to sunflower.However, Walworth et al. (1992) concluded that there were advantages in using the Mehlich-3 extractor, for allowing the simultaneous extraction of Ca, Mg, K, P, B, Fe, Cu, Mn, and Zn.
No significant correlation was observed between Mehlich-1 and Mehlich-3 extractors, at 0.20-0.30m and 0.30-0.40m depths and between Hot Water and Mehlich-3 extractors at 0.30-0.40m depth (Table 3).Despite this fact, significant correlations were registered at 1.0 % between B extracted and B absorbed at all depths, for all extractors (Table 2), demonstrating that Hot Water, Mehlich-1 and Mehlich-3 extractors are appropriate for evaluating plant-available B. Thus the fundamental aspect in the decision about which extractor to use is the correlation between concentrations extracted and concentrations in plant.This shows that the extraction methods studied here can be used to quantify B in sunflowers for the soil tested.

CONCLUSIONS 1 .
Soil B concentrations extracted by Hot Water, Mehlich-1 and Mehlich-3 increased linearly according to B rate applied to the soil surface at each depth, indicating that B was mobile in the soil.2.The extractors differed in extraction capacity, according to B rate and depth, with greater extraction at the surface.3.Hot Water, Mehlich-1 and Mehlich-3 were efficient in evaluating B available to sunflower grown on eutropherric Red Latosol (Rhodic Eutrudox).

Figure 3 .
Figure 3. Boron concentration in sunflower leaf tissue as related to the B rate applied to soil.

Table 4 . Correlation coefficient between boron concentrations in soil and sunflower, according to depth and extractor
** in the columns, significant correlation at 1.0 % probability.