EXTRACTION METHODS AND POTASSIUM AVAILABILITY FOR SOYBEAN IN THE PARANA STATE’S SOILS, BRAZIL

: Evaluating the potassium (K) availability to the plants can be performed by different methods. In general, multielement methods are employed and they are used for K due to increase operational of soil testing laboratory. The efficacy of ammonium acetate, Mehlich-1, Mehlich-3 and ion exchange resin soil K extractants in determining the K availability for soybean [ Glycine max L. (Merrill)], in different Parana’s soils, were investigated in this study. Twelve soil samples, collected from the upper layer 0–0.20 m, were cultivated with soybean for 42 days in greenhouse conditions. The average amount of available soil K extracted in test soils was found, displayed in decreasing order: ammonium acetate > Mehlich-3 > Mehlich-1 = ion exchange resin. The correlation coefficients between the K extracted content by ammonium acetate, Mehlich-1, Mehlich-3 and ion exchange resin and the K accumulated amount in the plants were, respectively, 0.90, 0.77, 0.88 and 0.89. The efficacy of ammonium acetate, Mehlich-3 and ion exchange resin extractants in assessing plant-available K in soils from Parana State were similar. The Mehlich-1 was the least efficient in estimating plant-available K in Parana’s soils.


INTRODUCTION
The evaluation of potassium (K) availability in soil is accomplished by the estimating their amount in readily available pools to plantsi.e., soil solution K and exchangeable K. However, in many situations the predicting plant available soil K has proven to be a difficult task because the plants typically remove more K from the soil than is present in these two pools (ROSOLEM et al., 2012;GARCIA et al., 2008;KAMINSKI et al., 2007;MOODY;BELL, 2006). When solution and exchangeable K are reduced to low levels by plant uptake, non-exchangeable K (fixed pools) can be released from the soils to replenish the plant available pool of K ions (ROSOLEM et al., 2012).
Various extractants and methods have been used for assessing the portion of soil K that is available to plants. The first purpose of each extractant is the determination of available nutrients for plants. Two parameters are very important, when an extractant is utilized: (i) evaluating the nutrient labile form and (ii) extracting methods that are fast and economic. The neutral 1 mol L -1 ammonium acetate (NH 4 OAc) solution, which extracts both solution and exchangeable K is the most common extractant in determining available K (NACHTIGALL; RAIJ, 2005). The method is simple, fast, low cost and well accurate. However, for increasing the efficiency of soil testing laboratory, extractants that extract several elements simultaneously are more attractive. In Brazil, the Mehlich-1 (TEDESCO et al., 1995) and the ion exchange resin (RAIJ et al., 2001) soil K extractants are the methods used to assess K availability to plants.
The selection of an extractant to determine the availability of a nutrient in the soil is based on the standard extractanti.e., the plant. Martins et al. (2004) used 1 mol L -1 NH4OAc, Mehlich-1, and boiling nitric acid (HNO 3 ) at different concentrations (0.25, 0.5, 1.0, 2.0 and 4.0 mol L -1 ) in extracting available K for wheat in soils of Campos Gerais, Paraná, Brazil. They found that 1 mol L -1 NH4OAc and boiling 4 mol L -1 HNO3 extractants were good indices for K availability. Slaton et al. (2010) compared Mehlich-3 and 1 mol L -1 HNO 3 and found that Mehlich-3-extractable K is an excellent predictor of soil K availability for soybean grown on silt loams in eastern Arkansas. Bortolon et al. (2010) found that Mehlich-1, Mehlich-3, and ion exchange resin extractants were good indices for available K estimation for soybean and wheat in soils of southern Brazil. On the other hand, Hosseinpur and Zarenia (2012) showed that 1 mol L -1 NH4OAc, 0.1 mol L -1 BaCl 2 , 0.1 mol L -1 HCl, and boiling 1 mol L -1 HNO 3 can not be used as available K extractants, whereas 0.1 mol L -1 HNO 3 , Mehlich-1, and 0.01 mol L -1 CaCl 2 extractants would be suitable as soil testing methods for determining available K for common bean in Iran soils.
Many researchers suggested various extractants for available K determination according to research sites (MEDEIROS et al., 2010;SAMAVATI, 2008;ZBÍRAL;NEMEC, 2005;WANG et al., 2004;MARTINS et al., 2004). Therefore, it is necessary to assess the ability of K extractants to predict plant-available K in a wide range of soils and plants.
In Paraná State, there have been few studies on methods for assessing K availability, thus making it difficult to define the most appropriate method for soil conditions in Paraná. However, it is essential to conduct research relating the amount extracted by the different extractants and the amount of K accumulated by plants using different extractants in soil testing laboratories.
The objectives of this study were to assess and compare the efficacy of ammonium acetate, Mehlich-1, Mehlich-3, and ion exchange resin extractants in determining potassium availability for soybean in different Paraná soils.

MATERIAL AND METHODS
The experiment was carried out in an experimental greenhouse, localized in the Universidade Estadual do Oeste do Paraná, in Marechal Cândido Rondon, Paraná, Brazil (24º31' S, 54º01' W and altitude of 420 m), where the environmental conditions were: minimum and maximum mean air temperature of 18 and 36ºC, respectively; mean air relative humidity of 65%.
Soil samples were collected from the upper layer 0-0.20 m at twelve sites in Paraná State, covering different regions and characteristics ( Table 1). The physic and chemical properties of soils (Table 2) were determined by adopting standard procedures (EMBRAPA, 1999). Lime (CaO 25%, MgO 12% and EEC 96%) was applied before the experiment to raise base saturation up to 70% for clay soils, 50% for sandy soils and 60% for medium texture soils (Raij et al., 1997). The soils were then moistened to reach 80% water retention capacity and incubated for 25 days. The range of base saturation was 55-74%, of pH in CaCl 2 5.6-6.4 and of organic matter 7-32 g dm -3 . Afterwards, 7.5-L subsamples of each soil were transferred to 8-L plastic pots.
Soybean [Glycine max (L.) Merrill, cv. CD 215] was sown, and seven days after seedling emergence, they were thinned to two plants per pot. Fertilization was performed with macro and micronutrients in solution four days after plant emergence, as recommended for soybean crop grown under greenhouse as described by Novais et al. (1991), with modification. The only exception was the K, which was omitted the application this nutrient.
The total amounts of fertilizer were: 40 mg dm -3 of N (ammonium sulfate), 160 mg dm -3 P (triple superphosphate), 15 mg dm -3 of S (calcium sulfate), 5 mg dm -3 of Cu (copper sulfate), 5 mg dm -3 of Zn (zinc sulfate) and 2 mg dm -3 of B (boric acid). At 15 and 30 days after emergence, the application of 40 mg dm -3 N as urea solution was also performed. Soil water content was monitored daily and kept around 80% of the soil water retention capacity.   KNUDSEN et al., 1982). PBC K : potential buffering capacity of K [in (mmol c kg -1 )/(mmol L -1 ) ½ ] determined as previously described by Mielniczuk (1978).
Forty days after plant emergence, the shoots were harvested, washed and oven-dried at 65 °C for 72 h, and then weighed. The plant material was ground, digested in nitric-perchloric acid, and the K concentration analyzed by a flame photometer (MALAVOLTA et al., 1997). The total K amount in the above-ground plant parts was calculated from the K concentration and dry matter data.
Before starting the experiment, the soil from each pot was sampled, oven-dried at 50 °C for 48 h and ground to pass through a 2.0 mm mesh screen.
The experiment was arranged in a randomized block design with four replicates (an individual pot containing two plant represented one replicate). The K content extracted by NH 4 OAc, Mehlich-1, Mehlich-3 and resin and the accumulated amounts in the shoots of soybean plants were subjected to analysis of variance (ANOVA); the effect of contents was partitioned for each extractant by linear regression analysis. Correlation analyses between K extracted by four extractants were performed. In addition to the selection of better extractants, simple correlation coefficients were calculated between K extracted by chemical methods and amount accumulated by plants. All analyses were performed using SigmaPlot version 11.0 software for Windows (Systat Software, Inc., San Jose, CA, USA).

RESULTS AND DISCUSSION
The amounts of K extracted using the different soil tests are shown in Table 3. Soil K amounts extracted with 1 mol L -1 NH 4 OAc were, on average, higher than the K amounts extracted with Mehlich-1 and resin (Table 3). Potassium amounts extracted by NH 4 OAc solution were 34% higher than the amounts extracted with Mehlich-1 (Table 3). These results confirm those reported by other authors (BORTOLON et al., 2010;SAMAVATI, 2008;ZBÍRAL;NEMEC, 2005;WANG et al., 2004). The agitation time of soils (15 min for NH 4 OAc and 5 min for Mehlich-1) may be one of the factors that resulted in greater extraction of the K by NH 4 OAc solution, by enabling higher contact time and ion exchange between K ion of soil exchange sites and the NH 4 + ion of extraction solution (HABBY et al., 1990).
The lowest K amounts extracted by the ion exchange resin (Table 3) may be associated with the extraction mechanism of the method. The acid extractors (Mehlich-1 and Mehlich-3) and neutral (NH 4 OAc) displace most of the soil exchangeable K considered available for correlation studies while in the extraction with resin occurs the removal by K exchange in the solution and that can move to the solution (TEDESCO et al., 1995). This difference in the extraction mechanism provides different levels of K extracted among the methods. The lower K extraction of the resin was also reported by other studies (MEDEIROS et al., 2010;BORTOLON et al., 2010;GIANELLO, 2005).
Soil K amount extracted with NH 4 OAc was similar to the amount extracted with Mehlich-3 (Table 3), confirming the results obtained by Bortolon et al. (2010), for soils of Rio Grande do Sul, Brazil. The similarity between these two extractors is due to the fact of extract similar K pools, mainly due to the formation of NH 4 OAc during compounding of Mehlich-3 solution (MEHLICH, 1984).  The highest amount of K extracted with Mehlich-3 compared to the Mehlich-1 (Table  3) may be attributed to the composition and higher pH (pH 2.5) of Mehlich-3 solution. This solution contains ammonium acetate, formed from the addition of all reagents that comprise the method; thereby increasing the K levels extracted (MEHLICH, 1984). These results are similar to those reported by other authors (BORTOLON et al., 2010;SAMAVATI, 2008;ZBÍRAL;NEMEC, 2005;WANG et al., 2004).
Correlation coefficients for the amounts of K extracted by the different methods are shown in Figure 1. Potassium extracted by Mehlich-1 (common method in Paraná State) was significantly correlated to the K extracted by 1 mol L -1 NH 4 OAc (r = 0.85), Mehlich-3 (r = 0.87) and ion exchange resin (r = 0.79) (Figure 1a, 1c and 1e). On the other hand, the correlation coefficients between the K amounts extracted with NH 4 OAc and Mehlich-3 and ion exchange resin were, respectively, 0.90 and 0.83 (Figures 1b and 1d), while the correlation between Mehlich-3 and ion exchange resin was 0.79 (Figure 1f). This high degree of association and significance among extractants indicates that all the studied methods used to evaluate the K availability to plants, although each extractant has its efficiency and accuracy. Bortolon et al. (2009) also showed the efficiency of Mehlich-1, Mehlich-3 and resin extractants for determination of K availability in soils of Rio Grande do Sul, Brazil under notillage system. However, due to the greater convenience of operation in routine laboratories, mainly due to the ease of obtaining clear extracts by decantation, thus eliminating the filtration of soils suspensions, and the low cost of analysis (ROSSI; FAGUNDES, 1998), Mehlich-1 is more suitable for the evaluation of available K in soil.
The relationships between extracted K by four extractants and amount of K taken up by soybean plants are shown in Figure 2. Correlation coefficients between the K concentrations extracted from the soil and the K amount accumulated in the soybean plants were 0.90 (p < 0.01) for 1 mol L -1 NH 4 OAc (Figure 2a), 0.78 (p < 0.01) for Mehlich-1 (Figure 2b), 0.87 (p < 0.01) for Mehlich-3 (Figure 2c), and 0.89 (p < 0.01) for resin ( Figure  2d). The NH 4 OAc, Mehlich-3 and ion exchange resin extractants were the most efficient in estimating K availability to for soybean in different soils of Paraná State, whit wide variation in physical and chemical properties. Adequate to assess the availability of nutrients to plants method should be efficient for a broad range of soil types and crops. According to Alvarez Venegas (1996), an appropriate method to assess nutrient availability to plants should be effective for a broad range of soil types and different types of crops.
Mehlich-1 had the lowest correlation coefficients between the K concentration in the soil and that accumulated by the soybean plants, indicating that this method was the least effective in assessing K availability in soils of Paraná, Brazil. Martins et al. (2004) also observed the low efficiency of Mehlich-1 in assessing K availability for wheat in soils of Campos Gerais, Paraná, Brazil. These authors verified that 1 mol L -1 NH 4 OAc and boiling 4 mol L -1 HNO 3 extractants were good indices for K availability. Slaton et al. (2010) compared Mehlich-3 and 1 mol L -1 HNO 3 and found that Mehlich-3-extractable K is an excellent predictor of soil K availability for soybean. However, in study performed with soil samples from the Rio Grande do Sul, Bortolon et al. (2010) found similar correlation coefficients between the K extracted with NH 4 OAc, Mehlich-1, Mehlich-3 and resin and the amount accumulated by maize plants (0.83 to 0.86) and soybean (0.84 to 0.94). Hosseinpur and Zarenia (2012) showed that 0.1 mol L -1 HNO 3 , Mehlich-1, and 0.01 mol L -1 CaCl 2 extractants would be suitable as soil testing methods for determining available K for common bean in Iran soils. In general, the high efficiency of 1 mol L -1 NH 4 OAc, Mehlich-1, Mehlich-3 and ion exchange resin extractants in determining the K availability for the majority of crops cultivated in Brazil is usually reported in the literature (Medeiros et al., 2010;Bortolon et al., 2009;Schlindwein and Gianello, 2005).

CONCLUSIONS
The efficacy of ammonium acetate, Mehlich-3 and ion exchange resin extractants in assessing potassium availability for soybean in soils from Paraná State was similar.
The Mehlich-1 extractant was less efficient in estimating the soybean-available K in the soils studied.