Potassium and magnesium in some Kenyan soils: their mineral sources and release to Ca-resin

Five soil types from Kenya were examined for their ability to release K and Mg to crops by extracting them with a calcium-saturated cation exchange resin, and for the mineral sources of K and Mg. Amounts and rates of release of exchangeable, short- and long-term reserves of K and Mg were determined. This was done for K before and after one, two and 10 wetting and drying (WD) cycles, designed to mimic the effect of the local climate on K release. The order of decreasing K and Mg contents of the soils was Muguga > > Katumani > Kampi-ya-Mawe (KYM) ∼ Ithookwe ∼ Kwale. In the case of K, wetting and drying increased the exchangeable fraction in all of the soils. The Muguga, Ithookwe and KYM soils have the capacity to release a great deal of K in the long-term, but at very different rates; the Kwale soil has few reserves and fixes K on WD.
The K and Mg contents of the soils were closely related to the mica contents of the clay + silt fractions. Exchangeable and non-exchangeable K was released from soil micas primarily by ion exchange, with some contribution to non-exchangeable K from dissolution. Release of exchangeable Mg was primarily by ion exchange, whilst that of non-exchangeable matrix Mg was by dissolution of trioctahedral mica in the clay + silt and the sand fractions of the soils. Response would be expected only to K fertilizers and only with the Kwale soil. However, all the other soils should, in the interests of good husbandry, receive maintenance dressings of K and Mg at convenient times in the crop rotation.     

Effects of silicate weathering and lessivage on K‐content in forest soils derived from Pleistocene sediments

The potassium (K) content of soils developed from Pleistocene calcareous till, glacial sand and loess in NW Germany was investigated in order to characterize stores of K in feldspars (K_feldspar) and mica/illite (K_mica/illite) as well as changes as a function of soil depth. From each horizon, up to seven sand, six silt and three clay fractions were separated. K_feldspar and K_mica/illite were quantified by means of chemical composition and estimation by IR‐spectroscopy. On account of distinct differences in mineralogical composition between different particle size fractions, K‐content of the bulk soil < 2000 μm and the proportion of K_mica/illite and K_feldspar are clearly related to grain size distribution of the sample. Generally, the K‐content of particle size fractions of a soil derived from calcareous till is significantly higher than that of a soil from glacial sands. K_mica/illite of clay and silt fractions increases with depth, reflecting greater mica/illite weathering at the soil surface, whereas K_feldspar shows no noticeable change. Illite accumulates by lessivage in Bt horizons. On a whole‐soil basis, the Bt horizons of Luvisols derived from loess and calcareous till contain more K_mica/illite than either the A or the C horizons. By comparing the K‐content in the different particle size fractions with soil depth, the highest rate of change is found for soils derived from glacial sand. Gains in K in the silt fractions of soils from calcareous till and glacial sand result from weathering of feldspar sand grains. Additionally, decomposition of feldspar‐containing rock fragments of gravel size, and aeolian sedimentation, may also have contributed to these gains.     

Mathematical models to study the kinetics of potassium release from swell-shrink soils of Central India in relation to their mineralogy

Employing four mathematical models (first-order, parabolic-diffusion, Elovich and zero-order), kinetics of potassium desorption from eight soils with and without cropping were studied to evaluate their ability in explaining K release from soils. The decline in the soil test K in cropped soils over original soils was drastic in easily desorbable forms compared to that of strongly held forms like 3 M H₂SO₄ K. Results showed that parabolic diffusion as well as first-order kinetic equation explained the K release data well for both original and K depleted (cropped) soils. Elovich and zero-order equations were not suitable to describe the kinetic data. However, zero-order equation explained K release data better in case of K-depleted soils as compared to original soils. Soils with higher initial K contents registered higher release rate constants. Over the entire period of cropping the range of release rate (b) decreased from 1.26 to 1.53 × 10^?2 to values ranging from 1.12 to 1.30 × 10^?2 h^?1. In contrast, the first-order equation, parabolic diffusion showed higher b values for cropped soils as they represent the diffusion gradient. Mica and its biotite content in both silt and clay fractions showed significant correlation (r) with b values. Similarly with the rate of K release, clay content of soils maintained significant r whereas the silt content did not.     

Mineralselektive K-Freisetzung aus Böden mit Oktodecylammonium-Ionen (nc = 18)

Mineral-selective K release from soils by octodecylammonium- ions (n_c = 18) n_c18-releasable potassium was determined on soils of five different areas. n_c18-releasable K is the potassium which is displaced by octodecylammonium ions (ODA). Because the main sources of this potassium are trioctahedral micas, it is called briefly biotite-K. The investigated soils are from different areas in Bavaria (loess and clay stone derived soils, alluvial soils); partly the locations are K fertilizer trials. Soils of known age are locations from the Niederrhein terraces (Holocene soils). In the 0.2–2 μm (coarse clay) and 2–6 μm (fine silt) fractions the biotite-K content lies between 12 and 40% of total K. At given fractionation the highest amounts of K were released from the coarse clay of the soils. In the loess soils rich in silt the biotite-K pool in the medium and coarse silt fractions was also considerable. Clay soils showing poor K supplying power (K fertilizer trials) in the field are characterized by very low biotite-K quantities in all silt fractions. In the Holocene soils the age of which varies between 4 000 and 8 000 years the ODA releaseable values are greater than in the loess soils. In all fractions of the young Holocene soils the released K is greater than in the same fractions of the old Holocene soils. Calculated to a soil depth of 60 cm the K release in kg/ha reaches from 25 000 (young soil form) to 17 000 (old soil form).     

Exploitation of potassium by various crop species from primary minerals in soils rich in micas

We investigated the question of whether exchangeable K⁺ is a reliable factor for K⁺ availability to plants on representative arable soils (Aridisols) rich in K⁺-bearing minerals. Five soils with different textures were collected from different locations in Pakistan and used for pot experiments. The soils were separated into sand, silt, and clay fractions and quartz sand was added to each fraction to bring it to 1 kg per kg whole soil, i.e., for each fraction the quartz sand replaced the weight of the two excluded fractions. On these soil fraction-quartz mixtures wheat, elephant grass, maize, and barley were cultivated in a rotational sequence. Growth on the sand mixture was very poor and except for the elephant grass all species showed severe K⁺-deficiency symptoms. Growth on the mixture with silt and clay fractions was much better than on the sand fraction; there was no major difference in growth and K⁺ supply to plants whether grown on silt or clay, although the clay fraction was rich and the silt fraction poor in exchangeable K⁺. On both these fractions the plant-available K⁺ supply was suboptimal and the plants showed deficiency symptoms except for the elephant grass. This plant species had a relatively low growth rate but it grew similarly on sand, silt, and clay and did not show any K⁺ deficiency symptoms, with the K⁺ concentration in the plant tops indicating a sufficient K⁺ supply regardless of which soil fraction the plants were grown in. The reason for this finding is not yet understood and needs further investigation. It is concluded that on soils rich in mica, exchangeable K⁺ alone is a poor indicator of K⁺ availability to plants and that mica concentrations in the silt and clay fraction are of greater importance in supplying crops with K⁺ than exchangeable K⁺.     

POTASSIUM RESERVES IN A 'HARWELL' SERIES SOIL

A sample of Harwell soil containing 36 percent fine clay (< 0.3 μm) and 14 per cent coarse clay plus fine silt (0.3–5μm) was separated into fractions, and the K-supplying power of soil and fractions measured by cropping with ryegrass, exchange with Ca resin and double-label isotopic exchange with ⁴²K and ⁴⁵Ca ions. Mineralogical examination of the fractions coupled with the cropping experiments showed that the K-supplying power of the soil to ryegrass can be explained by the presence of a zeolite, clinoptilolite-heulandite, in addition to the clay minerals, mica, and interstratified illitic smectite, commonly found in a glauconitic clay-rich soil. The 0.3–5 μm fraction, containing much zeolite, has an exchange diffusion coefficient for K ions to Ca resin of 1.8 × 10^?16 cm²sec^?1 compared with a value of 5.7 × 10^?20 for the < 0.3μm fractions in which interstratified illitic smectite is the dominant mineral. Isotopic exchange shows that all Ca ions in fractions < 50μm are isotopically exchangeable. In fractions coarser than 20μm, some of the K ions in felspar and mica were not exchangeable within the duration of the experiments.     

Decomposability of organic matter in particle size fractions from field soils with straw incorporation

Soils from two field experiments on straw disposal were fractionated according to particle size using ultrasonic dispersion and gravity-sedimentation in water. Samples of whole soils, clay. silt and sand-size fractions were held for 49 days at 20°C and the CO₂ evolution measured on 14 dates by gas chromatography.Recovery of soil solids. C and N was 99, 98 and 93%, respectively. Most of the soil C and N was in the clay (<2μm). (loamy sand, 50% C and 56% N; sandy loam. 65% C and 68% N), the silt (2–20 μm) having smaller proportions (loamy sand, 41% C and 38% N; sandy loam. 29% C and 27% N). The sand fraction (20–6000 μm) accounted for 4–7% of the organic matter, and 1–2% of the C was water soluble. Straw incorporation generally increased the C and N content of whole soils and size fractions.The decomposition rate constants were higher for the sandy loam than for the loamy sand soil. For both soils, the decomposability of the organic matter decreased in the order: sand > clay ⩾ whole soil > silt. Straw incorporation increased the decomposition rate of whole soil and sand organic matter. whereas the effect of straw on clay and silt respiration was small.Between 58 and 73% of the respiration was from clay, 21–25% from silt and 6–19% from the sand size fraction.     

Zinc sorption–desorption by sand,silt and clay fractions in calcareous soils of Iran

Zinc sorption–desorption by sand, silt and clay fractions of six representative calcareous soils of Iran were measured. Sand, silt and clay particles were fractionated after dispersion of soils with an ultrasonic probe. Zinc sorption analysis was performed by adding eight rates of Zn from 6 to 120 μmol g^?1. For the desorption experiment, samples retained after the measurement of Zn sorption were resuspended sequentially in 0.01 M NaNO₃ solution and shaken for 24 h. Results indicated that Zn sorption by soil fractions increased in the order clay > silt > sand, and correlated negatively with CaCO₃ content and positively with cation exchange capacity (CEC) and smectite content. Results indicated that for all fractions, the Langmuir equation described the sorption rates fairly well. In contrast to sorption, Zn desorption from soil fractions increased in the order sand > silt > clay, and correlated positively with CaCO₃ content, CEC and smectite content. Results showed that parabolic diffusion and two constant equations adequately described the reaction rates of Zn desorption. In general, for all soils studied, the coarser the particle size, the less Zn sorption and more Zn desorption, and this reflects much higher risk of Zn leaching into groundwater or plant uptake in contaminated soils.     

Potassium release in some saline soils of Israel

Potassium release from samples of two saline and three nonsaline soils was studied by measuring the amounts of the cation in 0.01 N CaCl₂ and water extracts. The factors which affected the release were: the distribution of potassium between the silt and clay fractions, the total potassium content of the sample and its salinity. Soil samples with a distribution ratio above unity (i.e., with potassium mainly in the silt fraction) and a high total potassium content released K at a higher rate than those with a distribution ratio below unity and a high total potassium content. The effect of salinity was of minor importance in comparison with those of the two other factors.     

Adsorption‐desorption of zinc in mollisols and their relationship with uptake of fertilizer‐applied zinc by rice

Abstract

Zinc (Zn) adsorption in mollisols conformed to the linear form of Freundlich equation. The log K values were positively correlated with silt, clay, and carbonate contents and soil pH, but negatively correlated with sand content. Sequential desorption of adsorbed Zn in 0.05M Ca(NO₃)₂, 0.1M Mg(NO₃)₂, 0.005M DTPA, and 0.1M HCl revealed that weakly and specifically bound fractions of added Zn, which could easily equilibrate with soil solution, were low and decreased with silt and carbonate contents and soil pH. Weakly bound fraction increased with sand content. Strongly bound and complexed fraction of applied Zn was the maximum and increased with clay, soil organic carbon and carbonate contents and specific surface area, but decreased with sand content. The mineral bound fraction of applied Zn was intermediate and increased with silt, clay and carbonate contents, and soil pH and specific surface area. Zinc uptake due to added Zn fertilizer by rice crop (Y) negatively correlated with log K, but positively related to Zn content in the equilibrium soil extract and Zn desorbed in 0.05M Ca(NO₃)₂. Both log K and l/n values together explained 59.5% of the total variation in Y, while Zn content in the equilibrium soil extract, Zn desorbed in 0.05M Ca(NO₃)₂, 0.005M DTPA and 0.1M HCl collectively accounted 79.6% of the total variation in Y.     

Nitrogen in particle size fractions of soils incubated for five years with 15N-ammonium and 14C-hemicellulose

Four soils with a range of clay and silt contents were incubated for 5 a with ¹⁵N-labelled (NH₄)SO₄ and ¹⁴C-labelled hemicellulose and then fractionated according to particle size by ultrasonic dispersion and sedimentation. The distribution of labelled and native N between clay, silt and sand fractions was determined and elated to previous results on the C distributions. Between 29% and 48% of the added N was found in organic form. The ¹⁵N atom percentage excess decreased in the order: clay > whole soil > silt > sand. For both clay and silt, the enrichment factor for labelled and native N decreased with increasing fraction weight. Clay enrichment was higher for labelled than for native N, the converse being true for silt. The distribution of whole soil labelled organic N was: clay 77–91%, silt 4–11%, and sand <0.5%. Corresponding values for native N were 69–74%, 16–22%, and 1–2%, respectively. All soils had higher proportions of labelled than of native N in the clay, the converse was true for the silt. The C/N ratio of the native silt organic matter was higher and that of clay organic matter lower than whole soil C/N ratios. Differences between the C/N ratio distributions of native and labelled organic matter were small. The relative distribution of labelled N and C was very similar confirming that the turnover of C and N in soil organic matter is closely interrelated.     

The distribution of native and labelled carbon between soil particle size fractions isolated from long-term incubation experiments

Four soils with 6, 12, 23 and 46% clay were fractionated according to particle size after incubation for 5–6 years with ¹⁴C labelled straw, hemicellulose or glucose: 6–23% of the ¹⁴C was still present and the amount increased with increasing content of fine particles. clay fractions contained 66–84% of the ¹⁴C and the silt fractions accounted for 4–19%. <2% was found in the sand fractions and 4–9% was water soluble. The distribution of the native C was: clay, 46–68%; silt, 20–31%; sand, 2–7%. The clay fractions had higher relative proportions of ¹⁴C than of native C, the reverse being true for the silt fractions. This distribution pattern was not directly related to soil clay content or to kind of organic amendment. The C enrichment factor of clay and silt fractions (per cent C in fraction/per cent C in whole soil) increased with decreasing fraction size for both native and ¹⁴C. However, clay enrichment factors were higher for ¹⁴C than for native C, whereas silt enrichment factors were lower. A soil (9% clay) that had been incubated in the field for 18 years with ¹⁴C labelled straw was also analysed. Labelled C content at sampling was 9% of the initial value. In contrast to the other soils the distribution of labelled and native C was similar in the clay and silt fractions, which contained 55% and 33% of the whole soil C, respectively. The results indicate that clay-bound organic matter may be important in mediumterm organic matter turnover, whereas silt-bound organic matter may participate in longer-term organic matter cycling.     

Forms and distribution of potassium in particle size fractions on talc overburden soils in Nigeria

The potassium status of soils developed over talc overburden in a southern Guinea savanna of Nigeria was evaluated using exchangeable, acid extractable, total and residual potassium values in particle-size fractions. Soil samples collected from genetic horizons of six profile pits at Kampe Forest Reserve were separated into sand, silt and clay fractions. Exchangeable K, acid-extractable K, total K and residual K were determined in these fractions. Reserved K values were similar to those of mobile K, but lower than total and residual K, whereas exchangeable K showed the lowest values. Total K was >25 cmol kg^?1 in all the profiles; reserved K ranged from 9.26 to 24.45 cmol kg^?1 and mobile K ranged from 5.12 to 29.57 cmol kg^?1. Exchangeable K accounted for <1% of total K and ranged from 0.20 to 0.50 cmol kg^?1. In most cases, the clay fraction of the soils had the highest values for all potassium forms, followed by the silt fraction, while the sand fraction had the lowest values for these forms of potassium.     

THE DISTRIBUTION OF LABELLED SUGARS IN SOIL PARTICLE SIZE FRACTIONS AS A MEANS OF DISTINGUISHING PLANT AND MICROBIAL CARBOHYDRATE RESIDUES

Soils of the Countesswells and Insch series incubated with ¹⁴C labelled glucose or plant materials have been separated into clay (< 2 μm), silt, (2–20 μm), fine sand (20–250 μm) and coarse sand (>250μm) fractions and the distribution of individual labelled and unlabelled sugars was determined in each fraction. Both soils contained about 10–15 per cent clay, 18–23 per cent silt and about 60 per cent fine and coarse sand. For all soil samples the concentrations of sugars were usually greatest in the clay, slightly less in the silt, with values in the sand fractions being five or ten times lower, except when fresh plant material was present. In ¹⁴C glucose amended Insch soil, 55 per cent of the radioactivity in sugars (predominantly hexoses) occurred in the clay, 36 per cent in the silt, 3 per cent in the fine sand and 6 per cent in the coarse sand after 28 days incubation. For the Countesswells soil the values were 55, 42, 2 and 1 per cent respectively. In ¹⁴C ryegrass amended soil before incubation. 77 per cent of the radioactivity in sugars (predominantly glucose, arabinose and xylose) was in the coarse sand. After one year's incubation this had fallen to 59 per cent. In soil amended with ¹⁴C cereal rye straw the distribution of radioactivity in sugars after four years incubation was: clay, 21 per cent; silt, 43 per cent; fine sand, 21 per cent; coarse sand, 4 per cent. These distributions were compared with that of the naturally occurring sugars: clay, 31–42 per cent; silt, 40–43 per cent; fine sand, 3–11 per cent; coarse sand, 12–20 per cent.     

Release Kinetics of Nonexchangeable Potassium by Different Extractants from Soils of Varying Mineralogy and Depth

Abstract

Nonexchangeable potassium (K) release kinetics of six major benchmark soil series of India as affected by mineralogy of clay and silt fractions, soil depth and extraction media was investigated. The cumulative release of nonexchangeable K was greater in smectitic soils (353 mg K kg^?1 at 0‐ to 15‐cm depth and 296 mg K kg^?1 at 15‐ to 30‐cm depth, averaged for 2 soils and 3 extractants) than in illitic (151 mg K kg^?1 at 0‐ to 15‐cm depth and 112 mg K kg^?1 at 15‐ to 30‐cm depth) and kaolinitic (194 mg K kg^?1 at 0‐ to 15‐cm depth and 167 mg K kg^?1 at 15‐ to 30‐cm depth) soils. Surface soils exhibited larger cumulative K release in smectitic and illitic soils, whereas subsurface soils had larger K release in kaolinitic soils. Among the extractants, 0.01 M citric acid extracted a larger amount of nonexchangeable K followed by 0.01 M CaCl₂ and 0.01 M HCl. The efficiency of citric acid extractant was greater in illitic soils than in smectitic and kaolinitic soils. Release kinetics of nonexchangeable K conformed fairly well to parabolic and first‐order kinetic models. The curve pattern of parabolic diffusion model suggested diffusion controlled kinetics in all the soils, with a characteristic initial fast rate up to 7 h followed by a slower rate. Greater nonexchangeable K release rates in smectitic soils, calculated from the first‐order equation (b=91.13×10^?4 h^?1), suggested that the layer edge and wedge zones and swelling nature of clay facilitated the easier exchange. In contrast to smectitic soils, higher release rate constants obtained from parabolic diffusion equation (b=39.23×10^?3 h^?1) in illitic soils revealed that the low amount of exchangeable K on clay surface and larger amount of interlayer K allowed greater diffusion gradients, thus justifying the better fit of first‐order kinetic equation in smectitic soils and parabolic diffusion equation in illitic soils.     

Comparison of Four Equations to Describe the Kinetics of Lead Desorption from Soils

The rate of Pb desorption was investigated from clay (Silty clay, Torrifluvent), CaCO₃-rich (Sandy clay, Calciorthid), and sandy (Sandy loam, Quartzipsamment) soils at two different temperatures. Lead has not been released from CaCO₃-rich soils which suggests irreversible Pb sorption by the soil. The desorption was quite hysteretic from sand and clay soils. The total amount of Pb released from the clay soil exceeded that released from the sandy soil. The lower Pb desorption associated with the sandy soil is probably due to its higher calcium carbonate content relative to the clay soil. The kinetics of Pb desorption were evaluated using the Elovich, modified Freundlich, parabolic diffusion, and first order equations. The first order and parabolic diffusion equations adequately described the kinetics of Pb desorption from clay and sandy soils under isothermal conditions. The choice of first order and parabolic equations among others investigated was based on the goodness of fit and the more scientific theoretical assumptions of the equations. The apparent Pb diffusion rate coefficient (D_d) and desorption rate coefficient (k_d) values from the clay and sandy soils increased with increasing temperature.     

Importance of the textural fraction and its mineralogic characteristics in the potassium contents of different argentine soils

Abstract

Our purpose was to determine the influence of the textural fraction and its mineralogical characteristics on the contents of the exchangeable (Ki) and nonexchangeable (Knoi) forms of soil potassium (K). Contributions of each textural fraction on Ki were as follows: 73–80% for clay, 8–18% for silt, and 2–10% for sand. The clay fraction relative to the contribution to Ki was not the same for all the soils studied. Thus, not only the quantity of clay must be taken into account, but also the cation exchange capacity (CEC) of the involved minerals. For Knoi, the total contribution showed a greater dispersion, the corresponding figures being as follows: clay 35–68%, silt 25–52%, and sand, 4–25%. Therefore, clay was not always the main contributor. In some soils, silt was the most important fraction due to the presence of muscovite.     

Cation exchange capacities of organo-mineral particle-size fractions in soils from long-term experiments

Soils with and without organic manuring from 10 long-term manurial experiments in East Germany were fractionated into organo-mineral particle-size separates by ultrasonic disaggregation and sedimentation/decantation. The cation exchange capacities (CECs) buffered at pH 8.1 were determined for the size fractions fine+medium clay, coarse clay, fine, medium and coarse silt, sand, and for the total soil samples. In the samples from nine field experiments the CECs decreased with increased equivalent diameters (fine+medium clay: 489–8 13 mmol_c kg^?1, coarse clay: 367–749 mmol_c kg^?1, fine silt: 202–587 mmol_c kg^?1. medium silt: 63–345 mmol_c kg^?1, coarse silt: 12–128 mmol_c kg^?1 and sand: 10–156 mmol_c kg^?1. The CECs varied with genetic soil type, mineralogical composition of the <6.3-μm particles, and the C and N contents of the size fractions. In a pot experiment examining the role of various organic materials in the early stages of soil formation, the clay-size fractions had the largest CECs (85–392 mmol_c kg^?1), followed by the medium-silt (1 9-222 mmol_c kg^?1) and fine-silt fractions (23–192 mmol_c kg^?1). The effect of organic amendments on CEC was in general: compost>fresh farmyard manure = straw + mineral fertilizer = mineral fertilizer.     

The nature of smectites and associated interstratified minerals in soils of the Gharb plain of Morocco

The mineralogies of ‘Tirs’ (Typic Pelloxererts), and ‘Debs’ (Typic Haploxerolls and Typic Xerochrepts) soils of the Gharb plain in north-western Morocco are investigated, with special attention given to the determination of the nature of the smectitic phase using the lithium test (Li test) and the alkylammonium method. The sand and silt mineralogy of Tirs soils is dominated by quartz with small amounts of feldspars and kaolinite. The sand and silt fractions of Dehs soils also contain significant amounts of mica, chlorite, and interstratified phyllosilicates. The clay minerals of Tirs soils are predominantly a high-charge smectite. The estimated interlayer charge for this phase is 0.61 mol(c)/O₁₀(OH)₂ and the fraction of tetrahedral charge varies from 38 to 44%. Although the percentage tetrahedral charge is less than 50%, the smectitic phase behaves as beidellite with the Li test. Dehs clays are more heterogeneous, consisting of smectite, vermiculite, illite, kaolinite, chlorite, and interstratified illite/smectite and illite/vermiculite. The Li test and the alkylammonium method demonstrate that a high-charge smectite or vermiculite is interstratified with illite. A low-charge montmorillonite is also present both in Tirs and in Dehs soils. The high-charge beidellitic phase is believed to be a transformation product of mica, whilst the low charge montmorillonite is thought to be inherited from the parent material.     

Influence of prolonged arable cropping on lignin compounds in sandy soils of the South African Highveld

The preservation of plant residues is important for sustainable arable cropping. Lignin is a marker for plant residues in soils. We have investigated influences of the length of cultivation on the dynamics of lignin. Composite samples were taken from the top 20 cm of soils that have been cropped for periods varying from 0 to 98 years in each of three different agro‐ecosystems in the Free State Province of South Africa. Lignin‐derived phenols were determined in the <2 µm (clay), 2–20 µm (silt), 20–250 µm (fine sand) and 250– 2000 µm (coarse sand) size separates. With increasing length of cultivation, the concentration of such phenols decreased to 36% of that in the grassland. The lignin contents as proportions of the total carbon did not change during cultivation, suggesting that there was no selective enrichment of lignin moieties as C was lost as a result of cultivation. The loss rate constants of lignin concentrations in particle‐size fractions increased in the order clay (0.17 year^?1) ≤ silt (0.18 year^?1) < fine sand (0.20 year^?1) < coarse sand (0.22 year^?1). Increasing ratios of phenolic acids to aldehydes in bulk soil, silt and fine sand fractions with increasing length of cultivation indicated that side chains were being oxidized. The ratios in the silt fraction, however, decreased after 10–20 years. We attribute this to a loss of lignin together with silt by wind erosion, resulting in a rejuvenation of lignin compounds in the remaining silt‐sized pools of C.