| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Chemistry and Biochemistry, South Dakota State Univ., Brookings, SD 57007 USA
b Dep. of Soil Science, Univ. of Saskatchewan, 51 Campus Drive, Saskatoon, SK, Canada S7N 5A8
shangchao{at}hotmail.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
13C) signatures of forest- and sorghum-derived C were used to quantify C losses and gains in organo-mineral fractions separated by particle size, and further by density (for sands and silts) and magnetic susceptibility (for clays). Nearly 50% of original C was in the silt-sized fraction, mostly in microaggregates of intermediate density; 30% was held by clays, particularly those of intermediate magnetic susceptibility; and 20% was of sand-size, low-density, often recognizable plant residues. The 13C values in the forest soil showed the more humified SOM to be associated with finer, denser, and less magnetic fractions. After cultivation, total C content was 28% lower, with 59% of this reduction in the silts, 28% in the sand, and 19% in the clays. Loss of forest-derived C amounted to 45%. The sand fraction lost 54% of its forest C, the silts 45% (mostly from intermediate density fractions), and the clays 23% (mostly from intermediate magnetic fractions). Gains in sorghum-derived C amounted to 32% of C in the sand fraction, 12% in the silts (relatively evenly distributed among densities) and 13% in the clays (mostly in the nonmagnetic fraction). Thus, losses of forest C and gains of sorghum C occurred in different organo-mineral fractions, indicating that there were no unique active fractions corresponding with the concept of C pools with defined turnover characteristics used in models of organic matter turnover.
Abbreviations: SOM, soil organic matter • 13C, carbon-13 natural abundance
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Various physical fractionation methods have been used to obtain SOM fractions with meaningful turnover characteristics (Christensen, 1992). Particulate organic matter (i.e., the coarse) light fraction is sensitive to cultivation and management practices such as tillage, fertilization, and crop rotation (Tiessen and Stewart, 1983; Bremer et al., 1994; Cambardella and Elliott, 1994; Gregorich et al., 1996). Using particle-size, density, and high-gradient magnetic field fractionations, Shang and Tiessen (1998) characterized the SOM in two semiarid soils under native vegetation from northeast Brazil and found that about one-half of the total SOM was stabilized in silt-sized microaggregates. These aggregate-protected organic materials were less decomposed and less stable than mineral-associated SOM in temperate soils. Cultivation caused a 22% loss of sand-sized organic matter and the breakdown of silt-size microaggregates (2.0–2.4 g cm-3 density) with a concomitant loss of C in an Oxisol (Shang and Tiessen, 1997). However, fractionation and C analysis provided information on only the net changes in C, and not on the mineralization of original forest C and incorporation of new crop C.
The use of 13C has allowed major advances in SOM turnover studies since it permits calculation of separate organic C pools derived from old and new vegetation, as long as the two types of vegetation have contrasting photosynthetic pathways (C3 vs. C4) (Balesdent et al., 1987; Mariotti, 1991; Skjemstad et al., 1994; Cadisch and Giller, 1996). Once a chronology of land use is established, the decay rate of organic matter derived from previous vegetation can be calculated (Vitorello et al., 1989; Veldkamp, 1994; Gregorich et al., 1995, 1997). Desjardins et al. (1994) reported that 46 to 49% of total organic C in an eastern Amazonian forest soil (0–10 cm) was derived from pasture 10 yr after forest-to-pasture conversion. Feller et al. (1991) examined 13C changes in particle-size fractions of a Brazilian forest soil and found that C turnover rates decreased with decreasing particle size.
The aim of this study was to show the roles of different organo-mineral fractions in the turnover and stabilization of SOM in a semiarid tropical soil. A research site with a history of vegetation replacement from a C3 forest to a C4 sorghum monoculture in semiarid northeast Brazil was chosen. The spatial variability of the site, SOM losses, and nutrient transformations under cultivation were documented (Tiessen et al., 1992), and nutrient limitations as a result of cultivation and SOM depletion were reported (Salcedo et al., 1997) previously. The C loss from the top 20 cm of soil during cultivation amounted to 3 mg g-1 or 0.8 kg m-2. We here report details on the turnover and stabilization of SOM in this semiarid tropical soil as controlled by its association with mineral phases, using physical fractionation (Shang and Tiessen, 1998) in combination with 13C measurements.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
A widely used approach to studying the effects of cultivation is to compare soil samples from cultivated lands with those from an adjacent uncultivated land, because randomized trials are usually not available for comparisons of long-term processes (Tiessen et al., 1992; Veldkamp and Weitz, 1994). We chose two plots from a mosaic of fields representing a shifting cultivation sequence on which Tiessen et al. (1992) had evaluated spatial variability and cultivation effects along transects crossing six different fields. The two plots reexamined here are (i) a field cropped for 12 yr with sorghum and (ii) an adjacent mature forest area, which had not been cultivated. The 50 by 60 m cropped plot was fertilized and had been limed once at the beginning of cultivation. Variability of data for the two plots used here are given in Table 1 . All soil properties examined had a low within-plot variation except exchangeable Al and Fe in the sorghum field, which were affected by liming (Tiessen et al., 1992). Following the approach of Veldkamp and Weitz (1994), we found that if a precision of ±10% around the mean at 90% confidence level is intended, the numbers of samples required to calculate organic C are six and two for the forest and sorghum sites, respectively (10 and 3 at 95% confidence level). In order for the subsequent fractionation and analysis to represent the means of the fields with at least 95% confidence we prepared composites of 15 (150 g each) and 8 (100 g each) samples for the cultivated and native plots, respectively. Soil was sampled from both plots along transects to a depth of 20 cm (Tiessen et al., 1992). Soil samples were air-dried and passed through a 2-mm sieve before compositing.
|
The silt-size fraction was freeze-dried and fractionated sequentially into four density fractions (<1.8, 1.8–2.0, 2.0–2.4, and >2.4 g cm-3) using sodium metatungstate solutions. The density fractionation was replicated three times, and yields between replicates were compared to verify the accuracy of the separation. The three replicates of each density fraction were combined for isotopic and chemical analyses. The mean of two measurements for combined samples was reported (The average analytical error of C analysis was 2.3% of the mean).
By using high-gradient magnetic separation, we fractionated clays into four fractions, M<0.25, M0.25–0.83, M0.83–1.38 and M>1.38, based on their magnetic susceptibility. Ten grams of clay was washed twice with 1 M NaCl by centrifugation to maximize dispersion, sonified, and diluted to a 0.5% suspension with deionized water. The clay suspension was passed first through the highest magnetic field strength (1.38 T) to collect the M1.38 fraction. Materials that were not retained in the magnetic field were least susceptible magnetically. Clay retained by the magnet was further fractionated. The field strength was then set to 0.25 and 0.83 T to recover the M0.25, M0.25–0.83, and M0.83–1.38 fractions. The average of two replicates for each soil clay is reported.
Chemical and Carbon-13 Natural Abundance Analysis
Soil samples were analyzed for organic C by dry combustion (CR-12, Leco Corp., St. Joseph, MI). This acid Oxisol contained no natural carbonates, and none of the applied lime remained. We detected no inorganic C. Total N and P were determined by H2O2–H2SO4 digestion and autoanalysis (Thomas et al., 1967). Exchangeable cations (Ca, Mg, K, Al, and Fe) were extracted by molar NH4Cl and analyzed by atomic absorption.
The natural 13C abundance (13CV-PDB) of soils and soil fractions was measured on an isotope ratio mass spectrometer (Europa Scientific 20-20, Crewe, UK). For clays and clay fractions, a single determination for each sample was carried out and the average of two replicated samples was used. For the remaining samples, an average of two determinations on 13C abundance was obtained (The average analytical error of 13CV-PDB is 0.45%). A test of eight measurements of 13CV-PDB on one SOM sample gave a standard deviation of ±0.17
To estimate the proportion (f) of C4 (sorghum)-derived C in a soil fraction from the sorghum field soil, the following equation was used:
 |
is the 13C value of a soil sample or a SOM fraction from the sorghum site, F is the 13C value of the corresponding forest soil sample or fraction, S is the 13C value of sorghum residues used as reference. The 13C value for sorghum plants (aboveground composite, harvested at ground level) was -11.4. The various parts and residues of sorghum may differ in 13C value; however, Desjardins et al. (1994) showed that the differences had no significant effect on the subsequent calculations. Subsequently the concentration (g kg-1 soil) of C4–C and C3–C in the cultivated soil or soil fractions and thus the loss of C3–C since cultivation can be calculated. The term (S - F) as introduced above contains some assumptions and errors. As discussed below, stabilization of organic matter in the soil causes a slight positive shift in 13C values from those of the incoming plant residues (Mariotti, 1991). The appropriate reference value for S is therefore that of the SOM fraction, as used in the forest soil. The equivalent value for the sorghum field is not available since the stabilization of sorghum-derived organic matter has not proceeded to steady state. We therefore used the value of sorghum residue, which is slightly more negative than the appropriate soil value would have been. To estimate how large this error might be, we repeated the calculations using the average 13C value of 13 tree foliage and twig samples from the forest, which was -27.2, and thus slightly more negative than the -26.5 of the whole native soil. By using the wrong reference for both situations we arrived at turnover rates and proportions (not shown) that were <5% different from the values reported, and showed slightly greater replacement of soil C with new (C4) C. |
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Distribution of Soil Organic Carbon
The composite sample of forest soil had a C concentration of 13.0 g kg-1, slightly higher than the mean of the eight samples. For the cultivated soil, the C content of the composite was equal to the mean (Tables 1 and 2)
. The distribution pattern of total soil C among particle-size fractions was similar between the forest and cultivated soils. The silt-size fraction contained the greatest proportion of C (nearly 50% of total C), followed by the clay (
30%) and sand fractions (21%). More than one-half the sand-size C was in the LF1 (<1.8 g cm-3), which consisted mainly of plant residues with high C concentration and wide C/N ratio (Table 2). The LF2, recovered by swirling in water and separated subsequently in heavy liquid, contained fine sand, charcoal, and some soil aggregates.
|
Density fractionation of silt-size materials resulted in a low-C, heavy fraction (>2.4 g cm-3) containing primary minerals, a fraction of aggregates with 2.0 to 2.4 g cm-3 density, and lighter fractions (<1.8 and 1.8–2.0 g cm-3) consisting of partially decomposed organic matter with a high C concentration and high C/N ratio (Table 2). The latter were protected by aggregation and Fe and Al complexation (Shang and Tiessen, 1998). The 2.0 to 2.4 g cm-3 fraction of cultivated soil yielded a lower mass than that of the forest soil, and the difference accounted for most of the mass loss from the whole silt fraction. The previous study on a soil from the same area showed that further dispersion of this intermediate density fraction liberated light (<1.8 g cm-3) organic matter and clay particles (Shang and Tiessen, 1998). The aggregates of the cultivated soil, compared with their forest counterpart, had a lower stability and were easily dispersed when treated by ultrasonic vibration.
The organic matter associated with clay is believed to be well-decomposed and strongly bound to clay surfaces since organic matter and clay particles could not be physically separated by ultrasonic vibration and high density flotation (Shang and Tiessen, 1998). By separating clays according to magnetic susceptibility, we found that the M0.25–0.83 and M0.83–1.38 fractions were C enriched (Table 2). For the cultivated soil clay, the C concentrations in M0.25–0.83 and M0.83–1.38 fractions were about three times that in the nonmagnetic fraction (M1.38). In the cultivated soil, the M1.38 had 49 g kg-1 more mass than its forest counterpart. This nonmagnetic fraction appears to have accumulated materials derived from the destabilization of microaggregates in sand- (HF) and silt-size fractions during cultivation. This indicates that primarily clay particles with low magnetic susceptibility were involved in the formation of 2.0 to 2.4 g cm-3 silt-sized aggregates in the undisturbed soil. The M1.38 fraction contained primarily kaolinite associated with fine Fe oxide particles (Shang and Tiessen, 1998). There was also a possible mass shift from the three fractions with higher magnetic susceptibility to the nonmagnetic fraction (Table 2) since the three fractions together weighed 11g kg-1 (30%) less than the sum of corresponding forest fractions. Liming raised soil pH, altering the chemistry of Fe, Mn, and Al and probably affecting the magnetic susceptibility and aggregate stability of the soil.
There were variations in 13C values of up to 2.3 among particle-size fractions and their subfractions of the native soil (Table 3)
. These variations are due to isotopic fractionation during microbial respiration and selective decomposition of plant components that have differing 13C abundance (Balesdent et al., 1987; Melillo et al., 1989). Three trends are apparent: 13C values become more positive (and less like those of plant residues) (i) with decreasing particle size, (ii) with greater density in the silt fractions, and (iii) with lesser magnetic susceptibility in the clay fractions. Hence the most decomposed organic materials may be found in heavy, nonmagnetic, and small size fractions. The lightest (<1.8 g cm-3) silt-size fractions of the forest soil, with a 13C value of -26.3, contained partially decomposed plant residues (Shang and Tiessen, 1998). A more positive 13C value of -25.5 was found in the >2.4 g cm-3 materials in which organic matter was well-humified and/or of microbial origin, as indicated by a low C/N ratio (Table 2). Coarse SOM had similar 13C values to plant residues from which it derived in studies by Skjemstad at al. (1994) and Gregorich et al. (1995). Variation in 13C values among clay subfractions, in combination with C/N ratios, (Table 2) indicated that the magnetic fractions contained more preserved organic matter, whereas decomposed organic matter of microbial origin was more concentrated in the nonmagnetic fraction (M1.38).
|
13C value of sand-size organic matter (LF1) was more positive than that of whole soil or the plants (Table 3). Melillo et al. (1989) showed an increased 13C value of incubated plant materials and suggested that it was caused by an initial loss of fractions low in 13C.
Losses and Gains of Carbon
An error was introduced by C losses that occurred during the fractionation, and that amounted to 11% of total soil C in the forest and 7% in the cultivated soil sample (Table 2). In addition to experimental error during fractionation and summing of fraction contents, some loss was probably water-soluble C. Forest soils may have a larger water-soluble C pool (Boyer and Groffman, 1996).
Based on the sum of three particle-size fractions, the 12 years of cultivation resulted in a net C loss of 3.2 g kg-1 soil (Table 3) or 28% of total soil C. Twenty-two percent of this loss was from the sand fraction, 69% from the silt size, and 9% from the clay fraction. The picture was complicated by a mass shift between fractions, which would have been accompanied by C transfers and may have obscured losses in some of the finer fractions (Shang and Tiessen, 1997). In an attempt to correct for this we assumed that the amount of clay (17 g kg-1 soil) transferred from the silt-sized microaggregates upon cultivation contained 19 g C kg-1 clay, the same C concentration as the clay fraction (Table 2). This would give at least 0.32 g C kg-1 soil that was transferred from silts to clays upon ultrasonic dispersion of the cultivated soil, but not the native soil where silt-sized aggregates were more stable. With this correction, the distribution of total C loss would become 59% from the silt-size fraction and 19% from the clay fraction. Since silts contained 50% and clays 30% of the soil's C, this loss distribution indicates that C in clays is more stable than that in silt fractions.
After 12 yr, 45% of total forest-derived C was lost from this surface soil in semiarid NE Brazil (Table 3). This is comparable to the 46% loss reported by Desjardins et al. (1994) for the top 20-cm layer 10 yr after deforestation followed by pasture in the humid climate of the eastern Amazon.
In the Amazonian soil, between 46 and 52% of total soil C was reincorporated as new C4–C during the first 10 yr under pasture. The input of C4–C amounted to only 20% (1.76 g C kg-1 soil) of soil C during 12 yr of sorghum cultivation in the semiarid soil (Table 3). This slower C accretion may be due to the effects of tillage on residue stability, or the lower productivity of the semiarid site.
We also examined concurrent gains and losses in SOM fractions using the 13C signature of organic matter (Table 3). The sand-size LF1 had the highest percentage (35%) of C4–C or new C among fractions, and because of its large C content, this fraction alone contained 28% of total C4–C (0.37 g C kg-1 soil) in the cultivated soil. Shang and Tiessen (1997) found that the coarse organic matter in a cultivated soil was more oxidizable than that in a forest soil. The newly incorporated C could be responsible for the higher lability of organic matter in the cultivated soil since it was derived from easily decomposed sorghum residues.
The sand-size LF1 lost 46% of its C3–C upon cultivation; however, forest-derived organic matter still comprised 65% (0.69 g kg-1) of the total C in this fraction. This suggests that nutrient cycling through coarse organic matter in the cultivated soil is linked largely to mineralization of newly incorporated organic materials, and that the use of total LF as an index for organically mediated soil fertility (Chan, 1997; Barrios et al., 1998) may need refinement in some soils.
The C4–C comprised 12% of the C in whole silt and 13% of the C in clay fraction (Table 3). Among the silt-size density fractions, the incorporation of new C in the lighter (<2.0 g cm-3) fractions accounted for 50% (0.25 g kg-1 soil) of total C4–C in the whole silt fraction. The percentage C3–C losses from these lighter (<1.8 and 1.8–2.0 g cm-3) fractions (18 and 35%, respectively), however, were lower than those of the whole silt fraction and of the whole soil (Table 3). Organic matter in these fractions was closely associated with Fe and Al compounds within aggregates (Shang and Tiessen, 1998). Such material has been described as "occluded LF" by Golchin et al. (1994). The C4–C percentage in the 2.0 to 2.4 g cm-3 fraction was highest among density fractions, and at the same time the reduction in old C3–C in the fraction was greatest (65%). Part of this loss was linked to the breakdown of microaggregates, indicating that this fraction was more dynamic in C turnover. With 55% of C3–C loss (nearly 40% of total loss in the silt-size fraction) and only 13% of new C4–C, the >2.4 g cm-3 fraction showed the lowest retention and restitution of C. X-ray diffraction on the same density fraction of another soil sample from the same area showed that primary minerals with low surface activity predominated in its mineral phase (Shang and Tiessen, 1998).
In the clay, the greatest increase in C4–C content occurred in the least magnetic (M1.38) fraction, whereas the contribution of the three magnetic fractions to the total C4–C pool was very small. The M1.38 fraction contained 23% of total soil C4–C (0.30 g C kg-1 soil). Given its low C/N ratio (Table 2), the incorporated new C in the M1.38 fraction is likely to be of microbial origin. The greatest decrease in C3–C was seen in the intermediate (M0.25–0.83) fraction. The constant C3–C in the M1.38 fraction was probably affected by a mass gain from silt and sand-sized aggregates, which were more easily dispersed in the cultivated than the forest soil. The low percentage of C3–C loss and C4–C gain in the M0.25 clays indicated a passive organic pool (Table 3). This fraction represented only a small portion of total soil C. It had the highest dithionite-citric-bicarbonate–extractable Fe and Al (Shang and Tiessen, 1998), highest C/N ratio, and lowest 13C abundance among the four clay subfractions. The organic matter thus appeared to be plant residues, stabilized by sesquioxides and relatively untransformed by microbial processes.
In both silts and clays, the loss of C3–C and gain of C4–C occurred in different subfractions. Major C losses occurred from the relatively humified, heavy (>2.0 g cm-3) silt-sized fractions, and from the intermediate magnetic clays, while the gains occurred in coarse and light fractions as well as the nonmagnetic clays. Only the 2.0 to 2.4 g cm-3 fraction, containing 10% of total soil C, showed both gains and losses. The results indicate that sequestration and mineralization occur in different fractions and thus by different mechanisms and pathways.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
13C analysis indicated that this reduction was the result of a 45% loss of forest C and a concomitant accretion of sorghum C amounting to near 20% of the original C content. The loss rate of forest derived C is therefore very high, giving the native soil C a half-life of little more than 12 yr. Separation of organo-mineral complexes showed that this mineralization rate is the composite of a spectrum of mineralization rates observed for the different organo-mineral fractions, which lost between 14 and 66% of forest C. The greatest amounts were lost from the intermediate density silt-size aggregates and moderately magnetic clays. These results provide physical evidence for the coexistence of different C decomposition rates, which have been inferred from changes in decomposition kinetics and curve-splitting (Jenkinson and Rayner, 1977). Most investigators have also realized that mean residence times for soil C determined by radiocarbon dating are not a singular property of the soil's C but are the average of a spectrum of mean residence times of different C fractions. The present data illustrate some of this spectrum. Simulation models of SOM transformations take account of differences in C transformations by incorporating pools of different stability (Jenkinson and Rayner, 1977; Parton et al., 1988). Here we have shown that C decomposition and accretion as a result of land use change occur in different fractions of SOM. This indicates that there may be no unique pools of specific stability. One of the problems of modeling SOM turnover is that it has not been possible to match conceptual pools with measurable fractions—the challenge of modeling the measurable or measuring the modelable (Motavalli et al., 1994). The present analysis and data indicate that such a match may be impossible because stabilization and destabilization of SOM can follow different pathways, involving different fractions. The large C loss from intermediate density silt-sized fractions shows that SOM stabilized in organo-mineral complexes of this Oxisol is not resistant to degradation. This is in contrast to temperate soils and is a step toward explaining the lack of old, stable C in tropical surface soils (Tiessen et al., 1994).
|
ACKNOWLEDGMENTS |
|---|
Received for publication October 7, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
13C method in soil organic matter studies. Soil Biol. Biochem. 1994;26:153-160.This article has been cited by other articles:
|
|
 |
|
  Y. L. Zinn, R. Lal, J. M. Bigham, and D. V. S. Resck Edaphic Controls on Soil Organic Carbon Retention in the Brazilian Cerrado: Texture and Mineralogy Soil Sci. Soc. Am. J., June 8, 2007; 71(4): 1204 - 1214. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Dou and F. M. Hons Tillage and Nitrogen Effects on Soil Organic Matter Fractions in Wheat-based Systems Soil Sci. Soc. Am. J., September 20, 2006; 70(6): 1896 - 1905. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zach, H. Tiessen, and E. Noellemeyer Carbon Turnover and Carbon-13 Natural Abundance under Land Use Change in Semiarid Savanna Soils of La Pampa, Argentina Soil Sci. Soc. Am. J., August 3, 2006; 70(5): 1541 - 1546. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. Olk and E. G. Gregorich Overview of the Symposium Proceedings, "Meaningful Pools in Determining Soil Carbon and Nitrogen Dynamics" Soil Sci. Soc. Am. J., April 19, 2006; 70(3): 967 - 974. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. G. Gregorich, M. H. Beare, U. F. McKim, and J. O. Skjemstad Chemical and Biological Characteristics of Physically Uncomplexed Organic Matter Soil Sci. Soc. Am. J., April 19, 2006; 70(3): 975 - 985. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Solomon, F. Fritzsche, J. Lehmann, M. Tekalign, and W. Zech Soil Organic Matter Dynamics in the Subhumid Agroecosystems of the Ethiopian Highlands: Evidence From Natural 13C Abundance and Particle-Size Fractionation Soil Sci. Soc. Am. J., May 1, 2002; 66(3): 969 - 978. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
