| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a FCA, Universidad Nacional del Litoral, Av. P. Kreder 2805, CP:3080-Esperanza(SF), Argentina
b ESALQ, Universidade de Sao Paulo, Av. Padua Dias 11, CEP:13418-900, Piracicaba(SP), Brazil, Bolsista CNPq
c Institute of Soil Science and Plant Cultiviation, 24-100 Pulawy, Poland
* Corresponding author (apisilva{at}esalq.usp.br)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CS, clay + silt • F, friability • OM, organic matter • TS, tensile strength
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Tensile strength and F are influenced by several factors such as water content (Utomo and Dexter, 1981), dispersible clay (Kay and Dexter, 1992; Barzegar et al., 1994), concentration and composition of soil solution (Rahimi et al., 2000), wetting and drying cycles (Dexter, 1988b; Kay and Dexter, 1992), clay content and mineralogy (Kemper et al., 1987; Guérif, 1990, Ley et al., 1993), soil OM (Casuarano, 1993; Perfect et al., 1995; Rahimi et al., 2000), and cementing materials (Kay and Dexter, 1992; Kay and Angers, 1999). The influence of these factors on soil TS and soil F depends on climatic conditions, management practices, and soil composition (Kay and Dexter, 1992; Kemper et al., 1987; Macks et al., 1996; Watts and Dexter, 1997).
In tropical landscapes, a soil sequence from well-drained hilltops to poorly drained valleys is a frequent pattern. Variations of topography are indicative of changes in solum depth and degree of chemical weathering (Anjos et al., 1998) and, in some cases, of the presence of different soil classes. In tropical soils, especially Oxisols, the clay fraction is composed of several minerals, mainly kaolinite and gibbsite and Fe or Al oxides (Muggler et al., 1999).
Oxides are important for soil aggregation. Several authors have emphasized the aggregating capacity of sesquioxides (Pinheiro-Dick and Schwertmann, 1996; Igwe et al., 1999; Muggler et al., 1999). Aluminum and Fe oxides play an important function as binding agents between mineral particles having a strong influence on the soil structure by increasing the strength of failure zones (Kay and Angers, 1999). Nevertheless, some researches have shown that Fe oxides may or may not influence soil aggregation in Oxisols (Deshpande et al., 1968; Muggler et al., 1997, 1999), and sometimes the effects of the sesquioxides may be mainly because of the interaction between OM and oxides or clay content (Guérif, 1990; Bartoli et al., 1992a). Organic matter has been mentioned as an aggregating or disaggregating agent depending on its chemical composition and presence of other cementing materials (Golberg et al., 1990). In spite of the relevance of the subject, information on the effects of intrinsic soil properties on the TS and F of tropical Oxisols is limited.
This research hypothesized that TS and F are strongly influenced by soil intrinsic properties in some Oxisols that have a wide range of inherent soil properties. Therefore, the objective was to examine the effects of soil texture, OM and crystalline and poorly crystalline oxides, and their interactions on TS and F for natural aggregates of some Oxisols.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The experimental design was a 1410 m-long transect, consisting of 25 point locations spaced 
50 m. The transect covered three soil types that progressively vary their physical and chemical properties from the top to the bottom of the landscape. The soils were classified as sandy, kaolinitic Typic Hapludox; loamy, kaolinitic Typic Hapludox; and very clayey, kaolinitic Rhodic Hapludox; respectively. All soils are in the isothermic family.
Twenty-five soil blocks (15 by 20 by 15 cm) were collected with a shovel from the top layer of a catena (from 22°37'25'' S lat., 48°47'04'' W long. to 22°37'37'' S lat., 48°46'16'' W long.).
At the laboratory, the samples were air-dried and the soil blocks were broken up carefully into their constituent aggregates by hand. Aggregates between 12.5 and 19 mm were selected by sieving. This size range was chosen partly because the aggregates are easy to handle and measure, and partly because these aggregates must be fragmented by tillage to form an ideal seedbed which is typically composed of aggregates of 1- to 5-mm diam. (Dexter, 1988a). Since the aggregates were obtained from blocks by progressively breaking them down, at each stage the sample could have been mechanically stressed. However, it has been shown that this method of sample preparation has no significant effects on the structure or strength of air-dried samples (Grant et al., 1990).
For the measurements of soil properties, 35 aggregates were separated from each soil sample and crushed in a total of 875 individual tests (i.e., 25 transect locations by 35 aggregates).
The indirect tension test was carried out using an electronically controlled loading frame, which applied a constant strain rate of 0.03 mm s-1 until the aggregate failed; i.e., until the formation of a continuous tensile crack which runs approximately between the polar diameters. Before the test, each aggregate was weighed. After that, each aggregate was place in the more stable position and loaded progressively, across a diameter, between a fixed lower plate and an upper parallel mobile plate that was assembled to an electronic load cell of 20 kg capacity. The electrical output was recorded by a data acquisition system. After each test, the aggregates of each set were oven dried at 105°C to calculate the water content.
The TS was calculated as suggested by Dexter and Kroesbergen, (1985):
 | [1] |
Where 0.576 is the proportionality constant, P is the applied force at failure (N), and D2 is the effective diameter (m). Equation [1] was developed from the theory for the loading of spherical samples of linearly elastic material. Because of its simplicity it has been used to estimate the TS of natural aggregates of sand loam soils, silty loam soils, clayey soils, and heavy clayey soils (Ley et al., 1993; Macks et al., 1996; Watts and Dexter, 1998; Chan et al., 1999).
On the assumption that aggregate density is constant, the effective diameter of each aggregate was calculated following Watts and Dexter (1998):
 | [2] |
Where M is the mass of an individual aggregate (g), Mo is the mean mass of the aggregates in the population (g), and Dm is the mean diameter (mm). The mean diameter of all aggregates in each set was assumed to be the mean of the sieve sizes used to select them.
The soil F was calculated through the coefficient of variation method as proposed by Watts and Dexter (1998):
 | [3] |
Where 
Y is the standard deviation of measured values of TS, Y is the mean of measured values of TS and n is the number of replicates. The second term is the standard error of the coefficient of variation. This method has the advantages that it can be used on aggregates of a single size and that the results are easy to calculate (Watts and Dexter, 1998; Macks et al., 1996).
After the TS measurements, the aggregates of each soil sample were ground, passed through a 2-mm sieve and mixed. The individual samples were used for physical and chemical analyses. Particle-size distribution involved dispersion of samples with sodium hexametaphosphate [HMP, (NaPO3)n], and shaking the soil sample for 16 h (overnight) on a reciprocating shaker at 120 reciprications per minute, and determining the suspended particle content using the pipet method (Gee and Bauder, 1986). The classification of USDA was used to establish the particle-size limits. Carbon content was determined through oxidation with potassium dichromate; Fe(III) oxides (Fed and Ald) were extracted with the Na-dithionite-citrate bicarbonate system (Mehra and Jackson, 1960); and the poorly crystalline oxides (Feo and Alo) were extracted with NH4 oxalate (Schwertmann, 1964). Silicon was coextracted with Feo (Sio). Fe, Al, and Si were measured by atomic absorption spectroscopy.
Stepwise multiple regression analyses (stepwise selection procedure, SAS Institute, 1991) were carried out to evaluate the relationship between soil intrinsic properties and TS–F. First, the stepwise procedure was used to select the significant variables. Then, all possible interactions were computed and included as additional variables in the stepwise procedure. In cases where the interactions terms were statistically significant, their interpretation was conducted following the procedure used by Aiken and West (1991). The significance level for variable selection was P < 0.05. Multicollinearity was assessed using the variance inflation factor according to Neter et al. (1989), to avoid including highly correlated independent variables into the model. Tensile strength and F were tested for normality using Shapiro-Wilk statistic (SAS Institute, 1991).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Utomo and Dexter (1981) measured TS and F of sandy loam soils and proposed a soil classification based on the F values obtained from the volume dependence method. Recently, Macks et al. (1996) used this classification to evaluate the suitability of soil structural condition for crop establishment of a wide range of soils including Oxisols. The volume dependence method estimates the F values (F') as being the slope of the straight line that relates the logarithm of TS with the logarithm of sample size (aggregate volume). This method produces F values that are smaller than those determined from the coefficient of variation method (F). Chan et al. (1999) found a mean ratio of F/F' of 2 for clayey soils. Therefore, we doubled the values of F' to compare them with our F value. The soil F classification used in this paper is presented in Table 2. According to the mean value of F, the soils may be classified as "friable". Out of the 25 samples, 10 were "very friable" and 15 were "friable". Larger values of F indicate that the larger clods have a smaller strength than the smaller aggregates and may be more readily broken into smaller stronger units, producing a suitable aggregate-size distribution with a few passes of tillage implements. A "friable" soil condition implies that a soil requires a minimum of tillage to produce a good seedbed of small aggregates providing optimum conditions for the germination and establishment of plants (Macks et al., 1996). Those larger values can be attributed to the same factors that influenced the TS (Dexter and Kroesbergen, 1985).
|
Effect of Soil Texture and Organic Matter Content on Tensile Strength
The influence of texture and C content in TS was assessed through multiple-regression analysis. The results are given in Table 3. Clay + silt content and OM had the strongest effect on the TS.
|
 | [4] |
Where TS is tensile strength, CS is clay + silt content, OM is organic matter content, and (CS x OM) is the interaction (CS) x (OM). The units of these quantities are given in Table 1.
There was a very significant (p < 0.0001) interaction between the soil mineral fraction and the soil organic fraction. These results agree with those reported by Perfect et al. (1995). The effect of CS content on TS depends on the OM content and vice-versa, as shown in Eq. [5] and [6].
 | [5] |
 | [6] |
To illustrate the interaction, values of CS content and OM were chosen to generate simple regression lines by substituting those values in Eq. [4]. Following Aiken and West (1991), the selected values were: one standard deviation below the mean (CS = 38%, OM = 20.78 g kg-1), at the mean (CS = 51%, OM = 27.00 g kg-1), and one standard deviation above the mean (CS = 64%, OM = 32.22 g kg-1). The results are given in Fig. 1 and 2 .
|
|
Both clay and silt content affected soil strength. Several studies have found a positive effect of clay content on TS (Kemper et al., 1987; Guérif, 1990; Bartoli et al., 1992a). However, these authors did not find a positive correlation between silt content and TS. On the other hand, Ley et al. (1993) and Kay and Angers (1999) suggested that dispersible clay and silt are among the major factors controlling the TS. Clay and silt fractions seem to act as cementing material between large particles, especially in weakly structured tropical soils with clay content varying from 6 to 45% (Ley et al., 1993). The small soil particles can be mobilized from higher to lower energy positions, being deposited mainly at the points of contact or near-contact of the larger particles as wet soil dries. In those new places, the mineral colloids appear to play an important role in bridging between and cementing the larger particles. Generally the lower the free energy, the greater the strength (Dexter, 1988a).
Figure 2 shows a significant positive relationship between TS and OM for CS = 64% and CS = 51%, and essentially no relationship between TS and OM for CS = 38%, as demonstrated by the significance levels for the t-tests for linear equations slopes (CS = 38%: t = -0.20, p < 0.8417; CS = 51%: t = 2.42, p < 0.0246; CS = 64%: t = 7.79, p < 0.0001). These results may explain why some researchers found a positive relationship between TS and OM (Bartoli et al., 1992a; Casuarano, 1993; Rahimi et al., 2000), whereas others found no relationship (Utomo and Dexter, 1981; Watts and Dexter, 1998) or negative correlation (Guérif, 1990; Watts and Dexter, 1997).
The influence of OM on failure zones (i.e., surfaces of weakness) is dependent on the nature of failure zones, which in turn depends on soil texture, the forms of the OM, and its spatial distribution (Kay and Angers, 1999). Clay particles of Oxisols are rigid and less mobile because their surface charges are low (Bartoli et al., 1992b). When the CS content is elevated, the OM may contribute to increased TS values, because OM can stabilize microaggregates by being incorporated in the small pore spaces between the domains and other clusters (Dexter, 1988a). On the other hand, when the CS content is small and OM content is high, the OM may decrease the TS. If the organic anions are no longer than the edge of the clay particles, their specific adsorption can compensate the positive charges at the adsorption sites, providing excess negative charges to the particles. Kemper et al. (1987) explained that if a larger portion of the mineral surfaces are coated by organic anions these substances will reduce the contacts of mineral particles and decrease the TS.
Effect of Soil Texture and Organic Matter Content on Friability
The model adjusted for soil TS was also tested for soil F; i.e., F = f (soil texture, C content) (F = 3.13, P > F = 0.054). The results showed that F was not correlated either with soil texture or with OM. These results differ from those obtained by Macks et al. (1996) and Watts and Dexter (1998) who found a positive correlation between F and C content, probably because they studied soils with a wider range of C content (from 11 to 59 g kg-1) than we did (from 20 to 44 g kg-1). In the other hand, Macks et al. (1996) found a poor correlation between texture and F. The larger textural and structural heterogeneity could partly explain our data. Soil is friable because of the distribution of flaws or microcracks within it that directly depend on the quantity of large soil particles. The presence of large, nonshrinking sand grains can induce differential shrinkage of the clay matrix in response to drying, resulting in an increase in the number of microcracks. In contrast, when the clay fraction is predominant few flaws are formed, because the clay matrix can shrink isotropically and the clay particles can remain in contact with their neighbors (Towner, 1988). On the other hand, the shape of individual soil aggregates also affects the values obtained for F (Dexter and Watts, 2000). At high sand contents, aggregates appear to be more spherical than clay aggregates, which have prismoidal shape (Perfect et al., 1995). Similarly, Dexter (1985) found that aggregates were more spherical when they had lower clay contents and when they had higher OM contents. The wide range of soil texture in this study may be responsible for some shape variability, which may contribute to the high strength heterogeneity of the aggregates.
Effect of Oxides and Organic Matter Content on Tensile Strength
In some Oxisols, soil strength has also been associated with the presence of Fe and Al oxides (Pinheiro-Dick and Schwertmann, 1996; Muggler et al., 1999). To examine what soil constituents of CS fraction were responsible for the TS, multiple regression analyses were carried out including the different oxides forms (Oxalate Fe, Al and Si, and dithionite Fe and Al). The results are given in Table 4.
|
 | [7] |
Where TS is tensile strength, Feo is oxalate Fe content, and OM is organic matter content.
Tensile strength was significantly influenced by Feo and OM. Regression analyses showed that a larger proportion of the variance in TS was explained by poorly crystalline Fe oxides (Partial R2 = 89%), while OM only accounted for 3% of TS variation. The Feo had a stronger effect than OM on TS. The cementing effect of OM has been mentioned by several authors (Bartoli et al., 1992a; Casuarano, 1993; Muggler et al., 1999; Rahimi et al., 2000). However, Igwe et al. (1999) indicated that in tropical soils, which have high amounts of Fe or Al oxides, the contribution of OM to soil strength can be reduced. The aggregating capacity of Fe oxides was emphasized by Pinheiro-Dick and Schwertmann (1996), Barral et al. (1998), Muggler et al. (1999), and Igwe et al. (1999). It was suggested that oxides can act in two ways: (i) aggregates seem to be formed through an attraction between positively charged Fe-oxide particles and negatively charged matrix particles, mainly clay sized, and (ii) oxides can exist as coating on the surface of mineral particles bridging different soil sized fractions (Kay and Angers, 1999; Muggler et al., 1999). In both ways, oxides may alter the surface charges of soil minerals, increasing the bonds between mineral and organic particles. Consequently, the soil strength increases.
Multiple regression analyses showed that poorly crystalline Fe oxides were the soil constituents of the CS fraction that most contributed to the soil TS. The similar values of correlation coefficients obtained for the two models (Eq. 4 and 7) seem to corroborate this affirmation. Similar results were found for tropical soils by Pinheiro-Dick and Schwertmann (1996) and Muggler et al. (1999). These authors reported that Fe oxides appeared to act as aggregating agents between mineral particles, preferentially between silt particles. In fact, this behavior of the oxides explains why clay and silt fractions entered together in the first model (Eq. 4). Our results emphasize the importance of considering the silt fraction as another factor that affects the strength of tropical soils.
Effect of Oxides on Friability
To assess if some specific form of the oxides affected soil F, multiple regression using the stepwise selection procedure was carried out. The results are given in Table 5.
|
 | [8] |
Where F is friability, Feo is oxalate Fe content, Sio is oxalate Si content, and Fec is crystalline Fe content.
Friability was negatively correlated with Fe- and Si- oxalate, but was positively related with crystalline Fe forms. The negative relationship between Feo- and Sio and F may be associated with the fact that these soil constituents act as cementing agents, increasing soil strength and decreasing soil F (Pinheiro-Dick and Schwertmann, 1996). The difference between dithionite- and oxalate- extractable Fe is an estimation of Fe crystalline forms (Golberg et al., 1990). According to Barral et al. (1998), these Fe forms favour formation of coarser particles. Pinheiro-Dick and Schwertmann (1996) also mentioned that crystalline Fe oxides, like hematite and goethite, may be associated with gibbsite, kaolinite, and quartz forming 100 to 200 µm aggregates in.
These microaggregates, which are highly stable in tropical soils (Muggler et al., 1999), combine to form the next hierarchical order, the soil aggregates. Compound particles of lower hierarchical order have a higher internal strength than particles of higher hierarchical order (Dexter, 1988a). These concepts explain why aggregate TS decreases with increasing aggregate size. The high mean value of F obtained in this study indicates that TS decreases with aggregate volume (Utomo and Dexter, 1981), with the crystalline Fe forms being the major factor affecting soil F.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication December 11, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Markgraf and R. Horn Scanning Electron Microscopy-Energy Dispersive Scan Analyses and Rheological Investigations of South-Brazilian Soils Soil Sci. Soc. Am. J., May 16, 2007; 71(3): 851 - 859. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E.-J. Park and A. J. M. Smucker Erosive Strengths of Concentric Regions within Soil Macroaggregates Soil Sci. Soc. Am. J., October 27, 2005; 69(6): 1912 - 1921. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Blanco-Canqui, R. Lal, L. B. Owens, W. M. Post, and R. C. Izaurralde Mechanical Properties and Organic Carbon of Soil Aggregates in the Northern Appalachians Soil Sci. Soc. Am. J., August 4, 2005; 69(5): 1472 - 1481. [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 | |||
.
