Abstract

For a long time, the importance of potassium (K) fertilization in Ethiopian soils has been incorrectly perceived to be unnecessary due to the misconception that K reserves in the soils were sufficient. This study was conducted with the aim of assessing four soil testing methods (NH₄OAc, CaCl₂, Mehlich-3 [M-3], and Schofield-Woodruff [ΔF]) to determine plant available K in Ethiopian highland clay soils. Sixty composite (0-20 cm), geo-referenced surface soils were collected from 20 districts in 11 zones of Ethiopia’s central highlands. While M-3 and NH₄OAc K extraction methods seem quite similar, but over-estimate soil K availability, the CaCl₂ method mainly identifies the soluble K fraction, under-estimating exchangeable K. The ΔF method shows considerable agreement with the CaCl₂ method but has additional sensitivity to exchangeable K. ΔF results indicate that 60% of the soils sampled could be considered as sufficient K suppliers, depending in the crop species, while the rest are poor K suppliers. Preliminary mineralogical examination revealed a dominance of illite/smectite in most samples but the ratio between the two minerals, which might determine K fixation and release rates, is yet unknown. In the absence of any alternative proven chemical method to evaluate soil K availability, direct measurements of crops’ K consumption should be integrated with simultaneous soil-K tests. Also, clay mineralogy should be further investigated on a local basis in order to determine and understand actual and possible dynamics of soil K status and availability, to establish a sufficient basis for practical recommendations.

Keywords: Illite; Mehlich-3; smectite; soil-K availability; soil-K extraction methods.

Introduction

The intensive fertilization approach emphasizes the need for fertilizer inputs to replace crop nutrient removal and to maintain soil nutrient reserves. In Ethiopia K fertilization was deemed to be unnecessary due to the misconception that K reserves in the soils were sufficient and, on the whole, in a form available to plants. Moreover, crops’ response to K fertilization was inconsistent or insignificant (Murphy, 1968). In addition, the exchangeable K content of most agricultural soils exceeded the universally accepted critical level, as set by the index based on ammonium acetate extraction method, at 0.25 cmol kg^-1.

Analyses of both past and recent information on K status in different woredas (districts) of Ethiopia, however, show that there has been a gradual decline in K status due to continuous mining, leaching, and soil erosion (Wassie, 2009). In its national soil fertility survey initiative, the Ethiopian Soil Information System (EthioSIS, 2013-2016) found K deficiency in key areas that have Vertisols, Nitisols, and other soil types. This has also been supported by crop K response demonstrations. At the same time, crop response to K fertilizer has emerged in many highland Vertisols, despite soil analysis results that show K levels higher than the critical level of 195 ppm, adopted by EthioSIS.

Potassium’s significance to plant nutrition is well recognized despite its complex and dynamic nature in soils (Zörb et al., 2014). Long-term intensive cropping in the absence of K inputs, adversely affects K supply to crop plants and consequently reduces crop yields (Swarup and Ganeshmurthy, 1998). Next to nitrogen (N), crops absorb K in greater amounts than any other nutrient. It is indispensable in nearly all processes required to sustain adequate plant growth and reproduction. Potassium plays a basic role in a series of fundamental metabolic and physiological processes in the plant. Its accumulation rate during early growth stages precedes N accumulation. Therefore, its supply to plants seems to be decisive for N utilization. In turn, K significantly affects plant growth rates and governs the degree of realization of yield potential (Grzebisz et al., 2012).

Soil K can be categorized into four main fractions: K in soil solution; exchangeable K; non-exchangeable K, which is fixed but potentially available; and K in the mineral matrix (Hoagland and Martin, 1933). Soil K availability to plants and microbes declines according to its chemical phase and location in the soil, as follows: soil solution > exchangeable K > fixed K (non-exchangeable) > mineral K (Sparks and Huang, 1985; Sparks, 1987; Sparks, 2000). According to Barbagelata (2006), these four categories give a general representation of the potential sources for plant-available K, but no distinct boundaries exist among them. The bulk of soil K is confined to the solid mineral soil phase (Sparks and Huang, 1985), while the exchangeable and the non-exchangeable K comprise a small portion of total soil K, located mostly at the soil solid-solution interphase. There are equilibria and kinetic reactions between the four soil K categories that affect the level of soil solution K at any particular time, hence determine the level of readily available K for plants. Although exchangeable K is widely used to evaluate soil K status and to predict K availability to crops (Krauss, 2003; Samaadi, 2006), such predictions have proven to be a difficult task due to the complexity of the dynamic equilibrium among the various forms of soil K (Barbagelata, 2006).

Potassium availability to plants is related in many ways to the structure and morphology of soil minerals, particularly clay (Zörb et al., 2014). Clay minerals comprise significant diversity of composition, structure, and consequent chemical and physical traits (Barton and Karathanasis, 2002). Thus, K sorption and desorption in soil are largely influenced by the amount and proportions of different clay mineral types. Potassium is readily adsorbed by 2:1 smectite clay minerals, thus plants require a higher dosage of K fertilizer than on other clay minerals, such as 2:1:1, 1:1, oxide, and alophane. Nursyamsia et al. (2008) suggested that of the 2:1 clay mineral types, beidelite or smectite has the highest fixation capacity. Bajawa (1987) showed that K fixation declines in the order of smectite > vermiculite > hydrous mica = chlorite = haloysite. The contradiction between the complex, fuzzy dynamics of K soil status, on the one hand, and the need to quantify the soil’s ability to supply K for current and future crops, on the other hand, calls for careful methodologies of soil testing. The relevance of total soil K as an indicator for plant nutrition, in many cases, therefore is quite small. However, quantifying K in soil solutions, and estimating the rates at which K is released from the exchangeable K pool, can provide enhanced diagnoses of the readily available K pool, thus supporting decisions regarding fertilizer application.

A wide range of soil extraction methods claim to quantify the readily available K soil fraction, however, each method holds advantages and drawbacks that are derived from the nature of the local soil. The most common soil test procedure at the global level is the use of neutral 1N ammonium acetate (NH₄OAc) extraction on air- or oven-dried soil samples (Cox et al., 1999). NH₄OAc extracts mainly soil solution K, exchangeable K, and a portion of interlayer K. This method uses a neutral salt solution to replace the cations present in the soil exchange complex. Therefore, the K concentration determined by this method is referred to as “exchangeable” for non-calcareous soils and “exchangeable plus soluble” for calcareous soils. Though this method mostly reflects the fertilizer K requirements of plants, there is some evidence that the NH₄OAc method is not sensitive enough for Vertisols. Cox et al. (1999) also claim that while 1N NH₄OAc soil test values work well for some soils, this approach is not reliable enough for soils with appreciable proportions of non-exchangeable interlayer K⁺, such as smectite mineral soils. The situation is even more problematic under intensively cropped agricultural systems (Bansal et al., 2002).

A single soil extraction with 0.01 M CaCl₂ appears to be the simplest, most inexpensive, and environmentally-friendly method, the results of which display the least variability among laboratories, compared to some other methods (Houba et al., 1996). Water and weak salt solutions extract K ions in the soil solution that are in equilibrium with those on the exchangeable complex. It is assumed that this method extracts the most readily available K from the exchangeable phase. Together with K in soil solution, the results effectually represent the plant’s available soil K pool, providing more precise estimates, compared to methods extracting the total exchangeable K.

The concept of a nutrient potential as a measure of soil K status was first suggested by Schofield (1947). This method uses an indirect measure of the energy input required by a plant to remove nutrients from the soil. Woodruff (1955) related classical thermodynamics to soil exchangeable K⁺ and calcium and magnesium (Ca²⁺+Mg²⁺) release to the soil solution for determining the free energy of K-Ca exchange equilibria in soils. According to Woodruff (1955), the energy of exchange is a measure of the chemical potential of K in the soil relative to the chemical potential of Ca in the same soil. The ability of a soil to supply K to plants is characterized by both the total amount of nutrient present (quantity, Q) and the energy level at which it is supplied (potential, P). The K⁺ potential (ΔGK) is a free energy measure of the soil’s nutrient availability, expressed as a ratio of the relative activity and exchange between K⁺ and Ca²⁺+Mg²⁺ (Keene et al., 2004). The Schofield-Woodruff (ΔF) method classifies soils according to their K supplying power as follows: soils with high K supplying power: ΔF > -2,000 cal mole^-1; medium K supplying power: -3,500 < ΔF < -2,000 cal mole^-1; and, poor K supplying soils: ΔF < -3,500 cal mole^-1.

An alternative method is Mehlich-3 (M-3) (Mehlich, 1984), which has become very common, almost ‘universal’, as it suits a wide range of soils (Zbiral and Nemec, 2000) and is relatively low cost. M-3 was developed as a multi-element (phosphorus [P], K, Ca, magnesium [Mg], sodium [Na], copper [Cu], zinc [Zn], manganese [Mn], Boron [B], aluminum [Al], and iron [Fe]) soil extraction and is widely used in agronomic studies to evaluate soil nutrient status and to establish fertilizer recommendations, mainly for P and K in humid regions. Several authors (Beegle and Oravec, 1990; Gartley et al., 2002; Wang et al., 2004) showed that both neutral 1N NH₄OAc and M-3 methods remove almost the same amount of K from the soil.

The M-3 soil test is widely accepted and employed throughout Africa, including Ethiopia. The debates that have recently been raised regarding the actual profile of K availability in selected Ethiopian clay soils, and the crucial consequences it might have on fertilization policy and on locally recommended practices, necessitates the reassessment of K extraction methods that are being employed. The objective of the present study, therefore, is to examine and compare various common methods and establish a starting point for further investigation, aiming to adopt the most appropriate method to evaluate K availability of selected Ethiopian highland clay soils.

Materials and methods

Soil sampling

Sixty geo-referenced composite surface (0-20 cm) soil samples were collected from 20 districts located in 11 zones within Ethiopia’s central highlands (Fig. 1). These areas were Awi zone (Dangila district), East Gojam zone (Aneded, Huleteju Enense districts), East Showa zone (Adea, Gimbichu districts), Gurage zone (Cheha, Enemore Ener districts), Hadiya zone (Limo district), North Shewa zone (Basona Worena, Kimbibit, Kuyu, Moretena Jiru, Siyadeberna Wayu districts), Sidama zone (Hagere Selam district), South West Shewa zone (Becho district), West Arsi zone (Arsi Negele), West Gojam zone (Bure, South Achefer, Yilmana Densa districts) and West Shewa zone (Jeldu district). From each district, three sub-districts were selected based on K⁺ levels found (high, medium and low K) in a previous study, using the M-3 method. The present samples, however, were not taken from exactly the same sampling points as the previous study.

Sample preparation, analysis, and K extraction methods

Samples were air-dried, gently crushed and sieved using a 2 mm diameter sieve for analysis. Particle size was determined using a laser diffraction technique at Ethiopia’s national soil testing center. Water and chemical extraction was conducted in the laboratories at Gilat Agricultural Research Center, ARO, in Israel. Water content was determined both from saturated paste and air-dried soil, and electrical conductivity (EC) and pH was measured from saturated paste extract. Cation exchange capacity (CEC) and exchangeable K percentage (EPP) were calculated after measuring Ca, Mg and Na from the NH₄OAc extract. Potassium, Ca, Mg and Na in all extracts were determined using atomic absorption spectroscopy (AAS).

Neutral ammonium-acetate extraction

Soil samples were dried and crushed to pass through a 2 mm sieve. Two grams of soil was extracted with 1:10 soil-solution ratio of neutral 1M NH₄OAc at pH 7 after shaking for 15 minutes (Helmek and Sparks, 1996).

Calcium chloride extraction

Air-dried soil samples were extracted for 2 hours with a 0.01 M CaCl₂ solution at 20°C, at a 1:10 ratio of sample to extracting solution, respectively. After measuring the pH in the settling suspension, the concentrations of nutritional and polluting elements were measured in the clear filtrate (Houba et al., 1996; Simonis and Setatou, 1996; Salomon, 1998).

ΔF method

This is based on determining the exchange energy of K (ΔF), with the prevalent divalent cations (Ca²⁺+Mg²⁺) (Schofield, 1947). The range of ΔF values is usually between -2,000 and -4,000 cal mole^-1, where the upper value (-2,000) indicates K sufficiency, and the lower (-4,000) K deficiency. The free energy of replacement (ΔF) is calculated using the formula proposed by Woodruff (1955)where R is gas constant, and T is the absolute temperature (°K):

This concept simulates the measure of the energy a plant must invest to remove K from the soil, and thus can represent K availability to plants.

M-3 method

The M-3 extracting solution is comprised of 0.2 M acetic acid (CH₃COOH), 0.25 M ammonium nitrate (NH₄NO₃), 0.015 M ammonium fluoride (NH₄F), 0.013 M HNO₃, and 0.001 M ethylene di-amine tetra-acetic acid (EDTA). A soil sample of 2.5 g is mixed with the extracting solution at the ratio of 1:10, respectively, and shaken for 5 minutes (Mehlich, 1984).

Clay mineralogy

Clay mineralogy analysis was conducted for 11 selected soil samples in the Ministry of National Infrastructures, Energy and Water Resources Geological Survey laboratory in Israel. These samples were disaggregated and passed through a 2 mm sieve. The clay fraction was collected from thin suspensions according to Stokes Law after carbonate minerals and salts were removed from samples by diluted HCl acid or buffered acetic acid, repeatedly washed, and treated by a low-intensity ultrasonic treatment for a few minutes. Clay suspension was pipetted onto glass slides and analyzed after air-drying, glycolation (at least 8 hours at 60°C and cooling overnight), and heating for 2 hours to 550°C) (Moore and Reynolds, 1989). The mineralogical composition of the clay fraction was analyzed by X-ray diffraction.

Results and discussion

Analytical results for selected physical and chemical soil properties are summarized and presented in Table 1. Soil pH (H₂O) values varied from 4.68 to 8.07, normally categorized from very acidic to moderately alkaline (Bruce and Rayment, 1982). The lowest value (4.68) was observed in South Achefer district in west Gojam zone, while the highest one (8.07) was found in Dangila district in Awi zone.

EC, measured form saturated soil paste extract (ECe), ranged from non-saline (0.11 dS m^-1, at Hulet Ej Enese) to moderately saline (5.73 dS m^-1, at Yilmana Densa), based on Richards (1954) classification of soil salinity rates.

Soil clay content, determined using laser diffraction, ranged from 25% at Basonaworena district to 83% at Limo and Moretina Jiru districts. While there is a wide range of clay contents across districts (30-80%, as an average), the intra-district variation was much smaller, with several exceptions at Aneded, Kuyu and Siyadeberna Wayu (Table 1).

While soil clay content has appreciable positive effects on soil water retention at both situations - air-dried and saturated paste (Fig. 2) - its significance determining soil fertility, and in the present study (soil K availability), is questionable.

Each of the three common methods in use to evaluate soil K availability, namely CaCl₂, NH₄OAc, and M-3, showed a very poor linkage between the estimated level of soil K availability and soil clay fraction, even at the wide range presented here (Fig. 3). Furthermore, the fourth method (ΔF), provided a slight indication that the greater the soil clay fraction, the energy required by plants to extract K tended to decrease (Fig. 3). Interestingly, the NA₄sub>OAc and M-3 methods obtained quite similar results, which were roughly two-fold higher than the K values obtained using the CaCl₂ method.

Indeed, direct pair-comparisons between the soil K extraction methods revealed a very high correlation between M-3 and NH₄OAc, with a coefficient of 0.85 (Fig. 4A). These results indicate that M-3 is somewhat more stringent than NH₄OAc at forcing K ions out from some clay minerals. However, together with their close chemical nature, these two methods appear to have similar chemical mechanisms of mining K out of clay minerals. This is in agreement with previous findings (Beegle and Oravec, 1990; Gartley et al., 2002; Wang et al., 2004), including on Ethiopian soils (Mamo et al., 1996). Unequivocally, the K extraction ability of the CaCl₂ method is about 35% and 43% below those of the M-3 and NH₄OAc methods, respectively (Fig. 4B, Fig. 4C). It remains unclear, however, which of the three methods provides a better indication of K availability for plants. While M-3 and NH₄OAc might appear too stringent, releasing a portion of the non-exchangeable K, CaCl₂ might be too mild in representing the soluble K, with only a slight depiction of exchangeable K (Houba, et al., 1996; Cox et al., 1999).

Testing the ΔF method against either M-3 or NH₄OAc revealed an absence of any correlation between them (Fig. 5A, Fig. 5B). In fact, there was a complete discrepancy between the methods in evaluating soil K status, and hence soil fertility. This result suggests intrinsic differences in the mode of action between the methods. While M-3/NH₄OAc methods might indiscriminately pull out K⁺ from relevant as well as irrelevant soil phases, the ΔF method seemingly offers a finer approach, which possibly distinguishes between clay minerals corresponding to their chemical affinity to K⁺. Thus, soils obtaining relatively low K status according M-3 might significantly differ in their K availability, according to the ΔF method, and vice versa. Excluding a few exceptions, a much better concurrence was found between ΔF and the CaCl₂ methods (Fig. 5C), which may indicate the dominance of the soluble K fraction, the most available one, in the results of ΔF.

Sorting the 60 Ethiopian soil samples by their K supplying potential, as determined using the ΔF method revealed that about 60% of them would be considered medium, while the rest are poor K suppliers (Table 2). Note that this sorting method significantly discriminates between neighboring soils, in some cases sending them to distant locations on the list. No linkage between ECe, soil pH or CEC and ΔF can be observed, indicating for the small relevance of these measures (at their detected range here) to the soil K supplying potential. As already mentioned and clearly noticed in Table 2, the relationships between soil total K, as determined by the M-3 or NH₄OAc methods, and ΔF, are absolutely coincidental as long as additional information is provided.

The mineral composition of clay fractions in each soil may provide the information required to interpret differences in K status and availability among them. The dominant minerals (>50%) in most of the 11 samples examined were interstratified illite and smectite (Table 3). Unfortunately, the ratio between the two in each sample was not resolved with the methods employed. While both minerals belong to the 2:1 clay class, they significantly differ in their ability to bind and release K ions (Barton and Karathanasis, 2002). Compared to illite and other clay minerals, smectite has a significantly higher tendency to fix K (Brady, 1984). Furthermore, many clay soils possess a relatively rapid dynamic transformation between the two minerals, which is influenced by temperature, moisture, pH, soil-K status, and plant roots (Scherer et al., 2003). Plant roots display remarkable ability to rapidly absorb K from the rhizosphere (Hinsinger, 2015; Adamo et al., 2016), thus strongly affecting the dynamic transformation between illite and smectite. Crops grown in the absence of K fertilization induced a rapid transformation of illite to smectite, accompanied by accelerated K fixation (Tributh et al., 1987). This situation is quite typical of many arable clay soils in Ethiopia’s central highlands; decades of inadequate K supply may have affected the clay mineral composition, favored increased K fixation and reduced K availability to crops (Abiye et al., 2004).

Concluding remarks

M-3 and NH₄OAc K extraction methods seem quite similar in their ability to partially extract non-exchangeable K, in addition to soluble and exchangeable K. Thus, both methods provide an over-estimate of soil K availability. The CaCl₂ method, on the contrary, mainly identifies the soluble K fraction, under-estimating exchangeable K. The ΔF method shows considerable agreement with the CaCl₂ method but has additional sensitivity to exchangeable K. Nevertheless, in the absence of any alternative proven chemical method to evaluate soil K availability (Wang et al., 2016) to compare with, the ΔF results only provide indications, not reliable recommendations. A possible solution appears to be integrating chemical methods with biological tests, namely, direct measurements of K consumption by crops with simultaneous soil-K tests, as suggested by Affinnih et al. (2014), and more recently by Li et al. (2016). In addition, clay mineralogy should be further investigated on a local basis in order to determine and understand actual and possible dynamics of soil K status and availability, to establish a sufficient basis for practical recommendations.

Acknowledgements

Gratitude is given to the late Prof. Tekalign Mamo, and to the Agricultural Transformation Agency (ATA), Ethiopia. Special thanks are given to Prof. Dr. Uri Yermiyahu, Mr. Isaac Zipori, and staff of the Center for Fertilization and Plant Nutrition (CFPN) at the Gilat Research Center for Arid and Semi-Arid Agricultural Research, Agricultural Research Organization (ARO), Volcani Center, Israel, and to Mr. Eldad Sokolowski, IPI Coordinator
for SSA/Ethiopia. This study was supported by ATA, Ethiopia, and the International Potash Institute (IPI), Switzerland.

References

• Adamo, P., P. Barré, V. Cozzolino, V. Di Meo, and B. Velde. 2016. Short Term Clay Mineral Release and Re-Capture of Potassium in a Zea mays Field Experiment. Geoderma 264:54-60.
• Affinnih, K.O., I.S. Salawu, and A.S. Isah, 2014. Methods of Available Potassium Assessment in Selected Soils of Kwara State, Nigeria. Agrosearch 14(1):76-87.
• Bajawa, M. 1987. Comparative Ammonium and Potassium Fixation by Some Wetland Rice Soil Clays as Affected by Mineralogical Composition and Treatment Sequence. J. Agron. Crop Sci. 158(1):65-68.
• Bansal, S.K., C.S. Rao, N.S. Pasricha, and P. Imas. 2002. Potassium Synamics in Major Benchmark Soil Series of India under Long-Term Cropping. 17^th World Congress of Soil Science, Bangkok, Thailand, 14-21 Aug 2002.
• Barbagelata, P.A. 2006. Evaluation of Potassium Soil Tests and Methods for Mapping Soil Fertility Properties in Iowa Corn and Soybean Fields, Iowa State University. Retrospective Theses and Dissertations. p. 53-62.
• Barton, C.D., and A.D. Karathanasis. 2002. Clay Minerals. In: Lal, R. (ed.). Encyclopedia of Soil Sciences. Marcel Dekker Inc., NY, USA. p. 187-192.
• Beegle, D., and T. Oravec. 1990. Comparison of Field Calibrations for Mehlich-3 P and K with Bray-Kurtz P1 and Ammonium Acetate K for Corn. Communications in Soil Science and Plant Analysis 21(13-16):1025-1036.
• Brady, N.C. 1984. The Nature and Properties of Soils. Ninth Ed. Mac Millan Publ. Co., New York.
• Bruce, R. C., and G.E. Rayment. 1982. Analytical Methods and Interpretations Used by the Agricultural Chemistry Branch for Soil and Land Use Surveys. Queensland Department of Primary Industries. Bulletin QB8 (2004). Indooroopilly, Queensland.
• Cox, A.E., B.C. Joern, S.M. Brouder, and D. Gao. 1999. Plant-Available Potassium Assessment with a Modified Sodium Tetraphenylboron Method. Soil Sci. Soc. Amer. J. 63:902-911.
• Gartley, K.L., J.T. Sims, C.T. Olsen, and P. Chu, P. 2002. Comparison of Soil Test Extractant Used in Mid-Atlantic United States. Communication in Soil Science and Plant Analysis 33:873-895.
• Grzebisz, W., K. Pepliñski, W. Szczepaniak, P. Barlog, and K. Cyna. 2012. Impact of Nitrogen Concentration Variability in Sugar Beet Plant Organs throughout the Growing Season on Dry Matter Accumulation Patterns. J. Elementology 17(3).
• Helmek, P.A., and D.L. Sparks 1996. Lithium, Sodium and Potassium. In: Sparks, D.L. (ed.). Methods of Soil Analysis. Part 3. Chemical Methods. Soil Sci. Soc. Amer., Madison, WI. p. 517-550.
• Hinsinger, P. 2015. What Root-Mineral Interactions are Occurring in the Rhizosphere and Play a Key Role in Potassium Biogeochemistry and Plant Nutrition? Frontiers in Potassium Science: Developing a Roadmap to Advance the Science of Potassium Soil Fertility Evaluation. Summary Report, 28.
• Hoagland, D.R., and J.C. Martin. 1933. Absorption of Potassium by Plants in Relation to Replaceable, Non-Replaceable, and Soil Solution Potassium. Soil Sci. 36(1).
• Houba, V.J.G., T.M. Lexmond, I. Novozamsky, and J.J. Van der Lee. 1996. State of the Art and Future Developments in Soil Analysis for Bioavailability Assessment. Science of the Total Environment 178(1):21-28.
• Krauss, A. 2003. Assessing Soil Potassium in View of Contemporary Crop Production. Regional IPI-LIALUA Workshop on Balanced Fertilization in Contemporary Plant production. Kaunas-Marijampol, Lithuania, September 30 - October 1, 2003.
• Keene, A., M.D. Melville, and B.C.T. Macdonald. 2004. Using Potassium Potentials to Examine Nutrient Availability in an Acid Sulfate Soil Landscape, Northern Australia. In: SuperSoil 2004: Proceedings of the Third Australian and New Zealand Soils Conference [CD-ROM]. Sydney, Australia: Regional Institute Ltd.
• Li, T., H. Wang, Z. Zhou, X. Chen, and J. Zhou. 2016. A New Grading System for Plant Available Potassium Using Exhaustive Cropping Techniques Combined with Chemical Analyses of Soils. Nature Scientific Reports 6. DOI: 10.1038/srep37327.
• Mehlich, A. 1984. Mehlich-3 Soil Test Extractant: A Modification of Mehlich-2 Extractant. Communications in Soil Science and Plant Analysis 15:1409-1416.
• Metson, A.J. 1961. Methods of Chemical Analysis for Soil Survey Samples. New Zealand Department of Scientific and Industrial Research. Soil Bureau Bulletin No.12. In: Hazelton, P.A, and B.W. Murphy (eds.). Interpreting Soil Test Results: What Do All the Numbers mean? Second Edition. New South Wales, (NSW) Department of Natural Resources, Collingwood, Australia. CSIRO Publishing. p. 168-175.
• Moore, D.M., and R.C. Reynolds. 1989. X-Ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford University Press, Oxford.
• Murphy, H.F. 1968. A report on the fertility status and other data on some soils of Ethiopia. Experiment Station Bulletin No. 44. College of Agriculture, Haile Sellasie University, Dire Dawa, Ethiopia.
• Nursyamsia, D., K. Idrisb, S. Sabihamb, D.A. Rachimb, and A. Sofyanc. 2008. Dominant Soil Characteristics Influencing Available Potassium on Smectitic Soils. Indonesian Journal of Agriculture 1(2):121-131.
• Richards, L.A. 1954. Diagnosis and Improvement of Saline and Alkaline Soils. USDA, Washington, USA.
• Salomon, E. 1998. Extraction of Soil Potassium with 0.01M Calcium Chloride Compared to Official Swedish Methods. Communications in Soil Science and Plant Analysis
  29:2841-2854.
• Samaadi, A. 2006. Potassium Exchange Isotherms as a Plant Availability Index in Selected Calcareous Soils of Western Azarbaijan Province, Iran. Turk. J. Afric. 30:21-222.
• Scherer, H.E., H.E. Goldbach, and J. Clemens. 2003. Potassium Dynamics in the Soil and Yield Formation in a Long-Term Field Experiment. Plant Soil Environment. Agricultural Department of the University of Bonn, Germany
  49(12):531-535.
• Schofield, R.K. 1947. A Ratio Low Governing the Equilibrium of Cations in the Soil Solution. Proc. Intern. Congr. Pure Appl. Chem. 3:257-261.
• Simonis, A.D., and H.D. Setatou. 1996. Assessment of Available Phosphorus and Potassium in Soils by the Calcium Chloride Extraction Method. Commun. Soil Sci. Plant Anal. 27:685-694.
• Sparks, D.L., and P.M. Huang. 1985. Physical Chemistry of Soil Potassium. In: Munson, R.D. (ed.) Potassium in Agriculture. p. 201-276. Amer. Soc. Agron., Madison, WI.
• Sparks, D.L. 1987. Potassium Dynamics in Soils. Advances in Soil Science 6:1-63.
• Sparks, D.L. 2000. Bioavailability of Soil Potassium, D-38-D-52. In: Sumner, M.E. (ed.) Handbook of Soil Science. CRC Press, Boca Raton, FL.
• Swarup, A., and A.N. Ganeshmurthy. 1998. Emerging Nutrient Deficiencies under Intensive Cropping System and Remedial Measures for Sustainable High Productivity. Fertilizer News 43(7):37-50.
• Mamo, T., C. Richter, and B. Heiligtag. 1996. Comparison of Extractants for the Determination of Available Phosphorus, Potassium, Calcium, Magnesium and Sodium in Some Ethiopian and German Soils. Communications in Soil Science and Plant Analysis 27(9-10):2197-2212.
• Tributh, H., E.V. Boguslawski, A.V. Lieres, D.Steffens, and K. Mengel. 1987. Effect of Potassium Removal by Crops on Transformation of Illitic Clay Minerals. Soil Science
  143:404-409.
• Wang, H.Y., W. Cheng, T. Li, J.M. Zhou, and X.Q. Chen. 2016. Can Nonexchangeable Potassium be Differentiated from Structural Potassium in Soils? Pedosphere 26:206-215.
• Wang, J.J., D.L. Harrell, R.E. Henderson, and P.F. Bell. 2004. Comparison of Soil-Test Extractant for Phosphorus, Potassium, Calcium, Magnesium, Sodium, Zinc, Copper, Manganese, and Iron in Louisiana Soils. Communication in Soil Science and Plant Analysis 35:145-160.
• Wassie, H. 2009. On Farm Verification of Potassium Fertilizer Effect on the Yield of Irish Potato Grown on Acidic Soils of Hagere Selam, Southern Ethiopia. Ethiopian Journal of Natural Resources 11.2:207-221.
• Woodruff, C.M. 1955. The Energies of Replacement of Calcium by Potassium in Soils. Soil Sci. Amer. Proceed. 19:167-171.
• Zbiral, J. and P. Nemec. 2000. Integrating of Mehlich-3 Extractant into the Czech Soil Testing Scheme. Communication in Soil Science and Plant Analysis 31:2171-2182.
• Zörb, C., M. Senbayram, and E. Peiter. 2014. Potassium in Agriculture-Status and Perspectives. J. Plant Physiol. 171:656-669.