Plastic limits of agricultural soils as functions of soil texture and organic matter content.
The plastic limits of soil define a range of soil moisture content in which soil has a plastic consistency. The limits are termed the lower plastic limit, often just referred to as the plastic limit (PL), and the upper plastic or liquid limit (LL). The plastic limits are also called the Atterberg consistency limits in tribute to the work of Atterberg (1911). In the plastic consistency range, i.e. at soil moisture contents between PL and LL, soil can be moulded or sheared without forming cracks (e.g. Campbell 1991). Soil at water contents below PL exhibits brittle behaviour and eventually fractures upon loading, while soil above LL behaves as a liquid and can flow under its own weight.
The range of moisture content over which a soil shows plastic properties is given by the plasticity index (PI) (e.g. McBride 2008):
PI = LL - PL (1)
Soil plasticity and the plastic limits have been explained by two approaches (see Campbell 1991 ): the water film theory, and the theory of critical state soil mechanics. The water film theory suggests that when soil is in the plastic consistency range, the clay particles, each of which is surrounded by a water film, slide over each other when soil is deformed. When the application of stress (the moulding forces) overcomes the soil water tension (cohesive forces), then soil undergoes plastic deformation with no elastic recovery. Upon further wetting, the water films coalesce and the water tension becomes less negative, so that the soil is capable of viscous flow when it reaches LL. As the soil is dried, the water tension becomes more negative and the thickness of the water films decreases, thereby increasing the inter-particle strength, and the clay particles tend to form more stable arrangements (micro-aggregates). The mean thickness of water films as a function of water potential has been discussed by Dexter and Richard (2009). According to this theory, water films at PL may be expected to range in thickness from 9 to 12 molecular layers of water. At a certain water content, corresponding to PL, sliding of individual clay particles over each other no longer occurs, and mechanical energy inputs cause the stable micro-aggregates to roll over each other, resulting in the onset of cracking. Values of water potential at the Atterberg limits found in various studies were summarised by McBride (1989); water potential at the plastic limits was affected by soil texture, and in the range 160-11 220 hPa for PL and 3-178 hPa for LL (see table 1 in McBride 1989). McBride (1989) investigated consistency limits for 290 soil horizons with clay contents ranging from 5 to 94%, and presented mean water potentials of 694-1051 hPa for PL and 30-98 hPa for LL.
The critical state soil mechanics theory suggests that there exists a line, termed the critical state line, which describes the conditions when soil is undergoing pure shear (e.g. moulding without a change in volume) during loading. Soil that is prepared for either the LL or PL test is in such a condition, and hence may be described by a point on the critical state line. The plastic limit was plotted on water content--suction curves for natural (undisturbed) and for continuously moulded London clay by Croney and Coleman (1954). They showed that PL was at the point where the curve for undisturbed soil joined the critical state curve, and that this occurred at a pore-water suction of 6300 hPa for that soil. The test procedures for the plastic limits are essentially dynamic shear tests (McBride and Bober 1989). Therefore, it may be speculated that PL and LL each correspond to a fixed undrained shear strength (see Campbell 1991 ). In fact, it has been found that the undrained shear strength at PL is ~100 times that at LL (Wroth and Wood 1978). On the basis of the ratio of undrained shear strengths at the plastic limits, it is possible to define the slope of the critical state line on a plot of the logarithm of mean normal pressure v. specific volume, and hence directly link soil consistency to soil strength (Wroth and Wood 1978; Campbell 1991).
The Atterberg consistency limits of soils are used primarily in classifying cohesive soil materials for engineering purposes (e.g. McBride 2008). They have been also related to the shear strength, bearing capacity, compressibility, swelling potential, and specific surface of soils (see McBride 1989). Hence, the plastic limits are important soil properties that can yield information on soil mechanical behaviour.
The lower plastic limit is widely used as a reference water content in tillage research. Gravimetric soil water contents, w, are compared with PL by using the ratio w/PL. Ideally, it would be better to express the wetness of the soil by the matric water potential. However, this requires much work, and it is easier and more realistic to express the wetness of the soil by w/PL. The optimum water content for tillage (e.g. Dexter and Bird 2001) is equal to the gravimetric water content at ~0.7-0.9 PL (Ojeniyi and Dexter 1979a; Muller et al. 2003; Barzegar et al. 2004; Keller et al. 2007), which is consistent with findings that friability reaches a maximum at a water content slightly below PL (Utomo and Dexter 1981a; Watts and Dexter 1998). When tillage is attempted at water contents above PL, soil is simply remoulded without any breakup or crumbling. Consequently, a soil with good workability usually has a water content at field capacity that is smaller than the water content at its PL.
The plastic limits and indices derived from them (e.g. in combination with actual soil water content) are used especially in soil engineering as a measure of soil mechanical properties and soil strength (e.g. Lang et al. 1996). Despite experimental evidence of the relationship between soil consistency and soil compressibility (see McBride and Bober 1989), the plastic limits are not used in (agricultural) soil compaction research to the same extent as they are used in foundation engineering. However, McBride and Bober (1989) suggested that the plastic limits could be obtained from uniaxial compression tests.
Following the above-mentioned water film theory, it is obvious that plasticity is related to the specific surface area (SSA, surface area per unit mass of soil) of the soil particles, and hence to the soil particle size distribution and, especially, clay content. This has been shown by several authors, including McBride and Bober (1989), de Jong et al. (1990), and Seybold et al. (2008). Furthermore, it has been shown that soil plasticity is influenced by clay mineralogy (e.g. Cerato 2001) and the nature of exchangeable cations (e.g. Schjonning 1991). Also, organic matter content (OM) has been found to influence soil plasticity and the plastic limits. For example, Archer (1969) found a strong positive correlation between PL and OM on a medium silt loam in England. However, the PI (Eqn 1) has been reported to either increase or decrease with increasing OM (see Campbell 1991). Organic matter plays an important role in agricultural (top)soils (e.g. Dexter et al. 2008a), and is therefore not negligible in the context of plasticity of agricultural soils. When we talk about OM, we are referring to OM that has decomposed and humified in situ and is in molecular forms, not to fleshly added OM that may be in fibrous or particulate forms and may be located differently in the soil matrix.
The objective of this paper was to study the plastic limits of agricultural soils covering a wide range of soil types in many countries as functions of soil texture and OM content. The resulting pedotransfer functions will then enable values to be estimated for cases where they have not been measured, and these estimated values can then be used to predict the suitability of different soil water contents for different machinery operations on different soil types.
Materials and methods
The samples were collected from agricultural fields in different locations in Australia, France, Germany, Israel, Poland, Sweden, the Netherlands, the UK, and the USA (Table 1) over the last 40 years. The samples therefore represent soils with different geological history, with different climatic conditions, and under different soil managements, and they cover a wide range of soil texture and OM content. Most of the data used here have been included in previous publications (see Table 1), where they were included only as part of the soil characterisation and to provide reference water contents; they have not been analysed or discussed previously.
Most of the soils were collected from the tilled layer. Some examples of subsoils were also collected and are listed as such in Table 1. The samples were measured shortly after collection.
The effects of drying during storage on plastic limits of topsoils are probably small because these soils will have air-dried and re-wetted many times in the past. However, subsoils may not have dried intensively during their history and therefore could have been changed significantly by air-drying. That is why the samples were stored moist (at the water contents at which they were collected) as much as possible.
The PL was determined according to the Casagrande plastic limit test, i.e. as the gravimetric water content at which a rolled thread of freshly moulded soil with a diameter of 3 mm just begins to crack (British Standard 1377 1975). Typically, nine replicate measurements were made for each soil.
The LL was determined mostly using the traditional Casagrande liquid limit apparatus (British Standard 1377), although some measurements were done using the drop-cone penetrometer test (mass of the cone plus shaft 80 g, cone angle 30[degrees]) (British Standard 1377 1975). We point out that all measurements of PL and LL were done either by the second author (A.R.D.) or, in some cases, by people trained by him, and therefore the operator variability is minimum.
We calculated PI according to Eqn 1. Unfortunately, clay mineral composition was not analysed for most soils. We determined the activity, A, as defined by Skempton (1953):
A = PI/clay (2)
The activity is an indicator of the dominant clay type present in a soil sample (e.g. Campbell 1991).
Soil texture and organic matter content
Soil particle size was analysed using the pipette or the hydrometer method. Soil OM content was mostly measured using the wet combustion method. However, for the French and the Swedish soils, the content of OM was calculated from the loss-on-ignition (by combustion at 1000[degrees]C) according to the ISO 10694 standard (ISO 10694 1995) or to Ljung (1987), respectively. Correction for carbonate carbon was not necessary for these soils. Soil texture and OM, together with values of PL and LL, are presented in Table 1.
Complexed and non-complexed clay
Dexter et al. (2008a) proposed that soil stability is related to the amount of complexed clay rather than to the total soil clay content. Therefore, we calculated the amount of complexed clay, CC, given as (Dexter et al. 2008a):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (3)
and the amount of non-complexed clay, NCC:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (4)
where OC is the soil organic carbon content and clay is the soil clay content in % by weight, and n is a real number in that 1 g of organic carbon is complexed with n grams of clay. Organic carbon content was calculated from OM by assuming a constant mass ratio of OM/OC = 1.724 kg/kg (Dexter et al. 2008a). Those authors found that 1 g of OC is the maximum that can be complexed with 10g of clay, giving the value of n = 10. Therefore, we adopted the value n = 10 in this paper.
Relationships between the plasticity parameters (i.e. PL, LL, or PI), and soil textural classes (clay, silt, and sand content) and soil OM content, and complexed and non-complexed clay were investigated by fitting linear regression models using the ordinary least-squares method. Factors were only included in the models if they were significant at P<0.1. The predictive performance of the multiple regression models was assessed by calculating the 95% prediction interval. For this we used the reg procedure of the SAS package (SAS Institute Inc., Cary, NC, USA). Basic assumptions of the linear modelling were verified as follows. We tested our data for homoscedasticity by plotting the residuals against the predicted values. Normal distribution was checked by displaying the data in a quantile--quantile-plot. For this, we used the gplot and the univariate procedure of the SAS package. Only models that fulfilled these postulates (homoscedasticity, normal distribution) are presented in this paper.
Results and discussion
The experimental procedures for determination of PL and LL are somewhat subjective and arbitrary, and for that reason prone to variability (McBride 2008). Therefore, it is important to evaluate the quality of the data. We did that for our data by expressing the test reproducibility in terms of the coefficient of variation, CV.
Unfortunately, we no longer had the data of all replicate measurements from all soils. However, the data of replicate measurements of PL were available for eight soils with clay content in the range 21-58%. For these soils, we found that the CV for PL was on average 2.9%. There were no effects of clay (P = 0.95), OM (P = 0.52), or PL (P = 0.45) on the CV of PL.
For LL, only raw data of the four French soils (Villamblain, Boigneville A and B, and Faux Perche; Table 1) were available. We calculated the confidence interval corresponding to one standard error to the regression line of gravimetric water content (w) v. logarithm of number of blows (N) at N = 25 (British Standard 1377 1975), and then expressed the test reproducibility in terms of the CV as we did for PL. For the four French soils, we found that the CV for LL was on average 2.5%.
Effects of clay content
As mentioned elsewhere, soil plasticity is related to the SSA of the soil panicles. Because small particles have a higher SSA than large particles, we found that the plastic limits were much more strongly correlated with clay content than with either silt or sand content. The PL as a function of clay content is shown in Fig. 1a. The relationship between PL and clay (Fig. 1a) was found to be:
PL = 21.28 (0.812) + 0.004 (0.0004) [clay.sup.2]; [R.sup.2] = 0.578, P < 0.0001 (5)
Here, and in the following equations, values in parentheses are the standard error. The soil in Fig. 1a that had a very low PL (13.7%) in relation to its clay content (33%) was the 'Condobolin' soil from New South Wales (Australia). This soil had no visible structure. The reason for the low PL could be that this soil consisted of micro-aggregated clay, and therefore may have been 'sub-plastic' (Butler 1955; McIntyre 1976). Butler (1955) described sub-plastic soils as soils for which 'the consistence properties suggest less clay then they actually contain'. In south-eastern Australia, sub-plasticity has been observed in calcareous aeolian clay; this clay has presumably been transported as aggregates of size range silt to fine sand, rather than as primary particles (Butler 1956; McIntyre 1976). Hence, sub-plastic soils feel, and in many ways behave, like sands, although they are, in reality, strongly micro-aggregated clays. In this case, sub-plastic soils are less plastic than expected from the particle size distribution (Blackmorc 1976; McIntyre 1976). Hence, sub-plasticity results in a reduction of the activity of the clays, A (Eqn 2) (Blackmore 1976). Further, Blackmore (1976) showed that, for sub-plastic soils, plasticity, i.e. PI (Eqn 1), increased with increasing intensity of mixing before plasticity tests, because more clay is released with increasing mixing intensity.
Other soils with low PL in relation to their clay content were 'Urrbrae B' (PL 19.7%, clay 46.0%), 'Hanslope 2' (PL 23%, clay 51.4%), and 'Hohe Warte P' (PL 30.0%, clay 73.0%). The reason for their low PL is that these soils were low in OM (0.8% for 'Urrbrae B' and 0.2% for 'Hohe Warte P'; unfortunately OM was not available for 'Hanslope 2'). In contrast, 'Highfield 4' (PL 34.0%, clay 27.0%), 'Highfield 5' (PL 34.4%, clay 23.0%), 'Boigneville B' (PL 34.5%, clay 23.6%), and 'Steppingly(4)' (PL 33.0%, clay 25.0%) soils had high PL in relation to their clay content due to high OM (OM in the range 4.8-6.5%). Effects of OM on plastic limits arc investigated in the section below. Another soil with high PL in relation to its clay content was the 'Waco' soil (PL 54.0%, clay 74.0%) from the Darling Downs in Queensland, Australia. This soil bad an OM of 1.9%, and therefore, the high PL cannot be explained with OM. This is a self-mulching soil that exhibits 'normal' swelling and shrinkage (i.e. the volume change of a soil sample is approximately equal to the volume of water that is added or lost). We suggest that this apparently high value of PL is a consequence of this swelling behaviour.
The LL was highly positively correlated with clay content (Fig. 1b). The linear relationship in Fig. 1b is the regression equation:
LL = 17.52 (2.35) + 0.860 (0.053) clay; [R.sup.2] = 0.797, P < 0.0001 (6)
The soils in Fig. 1b that have a high LL in relation to their clay content are 'Steppingly(4)' (LL 54.0%, clay 25.0%), 'Boigneville B' (LL 53.9%, clay 23.6%), and 'Highfield 4' and 'Highfield 5' (LL 53.5%, clay 27%; and LL 56.5%, clay 23.0%) (Table 1), which were described in the previous paragraph and which all are high in OM. The soil in Fig. 1b that had a low LL in relation to its clay content is 'Condoblin' (LL 23.7%, clay 23.0%; Table 1), which we associate with sub-plastic properties as described above.
The relationship with clay content is stronger for LL than for PL (cf. Eqns 5 and 6, and Fig. 1a and b). This is consistent with the findings of, for example, Mbagwu and Bazzoffi (1987) and de Jong et al. (1990). The differences in predictability of PL and LL from clay content cannot be explained with differences in test accuracy, as the reproducibility of PL was similar to that of LL (cf. previous section). We may speculate that the mechanical soil behaviour in the liquid range is more strongly controlled by SSA than the mechanical behaviour in the semi-solid range, and therefore LL is more strongly related to clay content than PL. However, we do not have a physical explanation of this.
Interestingly, for our data PL is virtually unaffected by clay content (r = 0.109, P > 0.15) for soils with clay contents below ~35% (Fig. 1a). Differences in mechanical behaviour of soil at clay contents above and below ~30% have also been observed in compaction studies (Larson et al. 1980; Imhoff et al. 2004; Keller and Hakansson 2010). We conjecture that this is a consequence of the packing arrangement of the soil particles, as was found for illite-kaolinite mixtures by Sills et al. (1973). They found that when there was an insufficient amount of smaller particles (i.e. the illite) to fill the spaces between the larger (kaolinite) particles, then the physical properties of the mixture were controlled by both of the components. However, when there was more than enough illite to fill the spaces between kaolinite particles, then the properties were controlled by the illite. With their clays, they found that the transition in behaviour occurred at ~40% of the smaller (illite) particles, which is not very different from the 35% of the smaller (clay) particles found here.
As a consequence of the high correlations of both PL and LL with clay content, PL and LL were highly correlated with each other (r = 0.806, P< 0.0001). The strong correlation between PL and LL is in accordance with earlier studies (e.g. Stanchi et al. 2009). Furthermore, PI (Eqn 1) was strongly affected by clay content too (Fig. 1c):
PI = 1.20 (2.04) + 0.551 (0.046) clay; [R.sup.2] = 0.682, P < 0.0001 (7)
The slope of the PI v. clay content relationship is defined as the activity, A (Eqn 2). The activity is related to the mineralogy of the soil, and increases with increasing SSA of the clay minerals. The activities of montmorillonite, illite and kaolinite are 7.2, 0.9, and 0.38, respectively (Lambe and Whitman 1969). From Fig. 1c, almost all of our soils lie between the activity lines of kaolinite and illite, with a mean value of A = 0.55 (Eqn 7). One soil, 'Polder 8' (classified as loam; Table 1), had a considerably lower activity (A=0.13; PI 1.9%, clay 15.1%). This polder soil is in South Flevoland (the Netherlands), originally located in the IJsselmeer (the former Zuiderzee). The 'Polder 8' soil was also found to have an unusually high porosity and a low value of tensile strength.
However, activity is not only affected by clay mineralogy, but also by OM. This is observed from Fig. 1c, where the lighter grey symbols represent the 'Highfield 1-5' soils (Table 1). This soil is from five plots of a Rothamsted long-term field experiment covering a wide range of OM. For this soil, activity, A, increased from 0.59 to 0.96 when OM increased from 1.9 (treatment with the lowest OM) to 5.4% (treatment with the highest OM). The reason is that OM increases the SSA of the soil.
Following Dexter et al. (2008a), we divided the clay content into two parts: the complexed clay, CC; and the non-complexed clay, NCC. We then obtained:
PL - 15.0 (1.35) + 0.465 (0.066) CC + 0.242 (0.034)NCC; [R.sup.2] = 0.559, P < 0.0001 (8)
which has an [R.sup.2] value similar to Eqn 5. However, it shows that complexed clay has about twice the effect of non-complexed clay on PL. For LL we obtained:
LL = 15.7 (2.59) + 1.00 (0.106)CC + 0.866 (0.055) NCC; [R.sup.2] = 0.858, P < 0.0001 (9)
which has an [R.sup.2] value similar to Eqn 6. However, this shows approximately equal effects of complexed and non-complexed clay on LL.
Division of the clay content into complexed and non-complexed did not improve the prediction of the plastic limits. This may be because for most of the soils in our dataset, all OM was complexed, which reflects the fact that the soils originate from arable land. Another reason may be that the plastic limits are properties of remoulded soil, while clay complexes play an important role especially in structured soils. However, micro-aggregates (as in sub-plastic soils) are not destroyed during moulding (see above), and textural pore space that may be stabilised by complexed OM is not modified during compaction (Dexter et al. 2008a, 2008b), and hence, clay-OM complexes could be expected to affect especially PL. This is perhaps why we see that complexed clay has about twice the effect ofnon-complexed clay on PL as mentioned above (Eqn 8). We suggest that the effects of non-complexed OM could be better evaluated in future studies by including soils with larger contents of OM, such a grassland soils.
Effects of organic matter
The relationships between OM and PL, LL, and PI, respectively, are presented in Fig. 2a-c. The outlier in Fig. 2a is the 'Waco' soil that is described in the previous section. There were weak positive correlations between OM and PL (r = 0.458, P < 0.0001) and OM and LL (r = 0.310, P<0.05), while there was no significant effect of OM on PI (r = 0.173, P>0.15). Thc scatters of data points in Fig. 2a-c are large, especially those in Fig. 2b and c. This is because the plastic limits are affected not only by OM, but also by soil texture as shown in the previous section. Therefore, we looked at soils from experimental sites with a range of OM contents, namely the 'Highfield' and the 'Boigneville' soils (Table 1). These two soils have very similar texture (Table 1), and are therefore ideal to study the isolated effect of OM. Regression between PL and OM (Fig. 2d) yielded:
PL = 10.55 (1.64) + 4.63 (0.421) OM; [R.sup.2] = 0.960, P < 0.0001 (10)
Equation 10 is very similar to that obtained by Archer (1969) for a medium silt loam in Lincolnshire, England: PL= 14.1 (4.3) + 4.3 (0.3) OM ([R.sup.2] = 0.993, P = 0.0038). Unfortunately, we do not know the clay content of that soil. We obtained the following regression equation for LL (Fig. 2d):
LL = 21.08 (2.60) + 6.50 (0.669) OM; [R.sup.2] = 0.950, P < 0.0001 (11)
The regression between PI and OM yielded:
PI = 10.53 (1.78) + 1.87 (0.457) OM; [R.sup.2] = 0.770, P < 0.0001 (12)
Equations 10-12 demonstrate a strong effect of OM on the plastic limits. However, the results reported in the literature on effects of OM on soil plasticity are inconsistent (see Campbell 1991). de Jong et al. (1990) found significant positive effects of OM on LL and PI, but no significant effect on PL. Stanchi et al. (2009), who analysed soils from the Italian Alps, found strong positive correlations between OM and PL and LL, but no effect of OM on PI. Seybold et al. (2008) reported no significant effects of OM on either LL or PI. McBride and Bober (1989) stated that OM increases PL but that LL is unaffected by OM, resulting in a decrease of PI with OM. Blanco-Canqui et al. (2006) reported that soil management significantly affected soil consistency and found significant positive correlations between OM and PL, LL, and PI for agricultural soils, which is in agreement with our results presented in Fig. 2d. Blanco-Canqui et al. (2006) found no further increase with OM in PL, LL, or PI at >4.8% organic C, but hypothesised that this was due to differences in the quality of OM between the agricultural and forest soils.
Clearly, it seems difficult to study the effect of OM on soil consistency with data from 'randomly' selected soils that were not sampled for the particular purpose of investigating effects of OM on soil plasticity. Our Eqns 10-12 and Fig. 2d suggest an important positive effect of OM on PL and LL, as well as on PI. However, this effect may be largely masked by soil texture effects when analysing a dataset consisting of soils with varying texture and OM (Fig. 2a-c).
Prediction equations for plastic limits
As shown in the previous sections, the plastic limits of soil depend strongly on both clay content and OM. Consequently, both clay and OM are needed for prediction of the plastic limits. We found the following regression model for estimation of PL:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (13)
In another multiple regression analysis, we included all texture classes (i.e. clay, silt, and sand content) and OM. However, this did not lead to inclusion of any other variables in the regression model other than those of Eqn 13, i.e. neither silt nor sand content was a statistically significant variable (P > 0.15):
For LL, we obtained:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (14)
Multiple regression including all texture classes (i.e. clay, silt and sand content) and OM additionally included the silt content (P = 0.015; partial [R.sup.2] = 0.02) in the regression model. That model (not shown) was slightly better ([R.sup.2] = 0.942; RMSE = 3.83) than Eqn 14, although it included one more variable (silt). However, the model was based on fewer data, because the silt content was not available for all soils (Table 1).
In principle, it should be possible to estimate PI from Eqns 13 and 14. However, this involves calculation of the difference of two estimated values, which may result in significant errors. Therefore, we performed multiple regression analysis for estimation of P! from texture and OM. First, we included only clay and OM in the analysis (as in Eqns 13-14). However, OM was not significant (cf. previous section), and the model produced was equal to Eqn 7. By including all textural classes, we found:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (15)
Note that the prediction equations (Eqns 13-15) are strictly only valid within the range of our data (cf. Table 1 ), i.e. for soils with a clay content in the range 11-74% and OM content in the range 0.2-6.9%. The signs of the coefficients included in Eqns 13-15 are sound, i.e. positive relationships between either clay or OM and the plastic limits, and a negative relationship between sand and the consistency limits. Further, the interaction term clay x OM is negative to compensate for the positive correlation between clay and OM (not shown).
The performances of Eqns 13-15 are presented in Fig. 3. The 95% prediction intervals span about [+ or -]8% (PL, PI; Fig. 3a and c) to [+ or -]10% (LL; Fig. 3b). In view of the natural variability (cf. Fig. 1 and Fig. 2a-c), the predictive power is satisfactory. Unfortunately, we could not compare our results with other prediction equations for plastic limits from the literature, because the prediction intervals of those were not given. However, we note that the [R.sup.2] value of our equations is higher and the RMSE lower than those of the equations given by de Jong et al. (1990) and Seybold et al. (2008). The plastic limit was slightly overestimated by Eqn 13 at low values of PL (Fig. 3a). The reason is that there was no effect (P>0.15) of clay content on PL at clay contents below ~35% (cf. Fig. 1a). The soil with the high PL that was underestimated by Eqn 13 is the 'Waco' soil, described in previous sections; however, excluding that soil from the multiple regression analysis yielded an equation almost identical to Eqn 13. Equation 13 overestimated PL for the 'Hohe Warte P' soil, which had a high clay content of 73% (Table 1), the 'Swed LA' soil, which had a high OM in relation to its clay content (Table 1), and the 'Clophill(5)' soil. It may be that these 'outliers' have different clay minerals present that have larger particles giving a lower activity of the clay per unit mass. In the case of the 'Clophill(5)' soil, we note that the soil was formed on Lower Greensand and that this area has been used in previous times for extraction of Fuller's Earth, which is a semi-plastic or non-plastic clay used in various industries. The prediction of LL was very good (Fig. 3b), as is also manifested in the high Re value of 0.93 of Eqn 14. The LL was underestimated for the 'Bindlach Al' soil, which had a high silt content (Table 1), while it was overestimated for the 'Clophill(5)' soil, described above, and the 'Urrbrae B' soil, which was low in OM in relation to its clay content (Table 1). The prediction of PI using Eqn 15 is satisfactory (Fig. 3c), with no obvious outliers.
Which soils are plastic?
As stated elsewhere, plasticity is related to the SSA of the soil particles, and hence to soil particle size distribution and especially clay content. This is reflected in the high correlations between the plastic limits and clay content that were also found in this study (Eqns 5-7; Fig. 1). It has been suggested that soil particles are surrounded by a water film that allows sliding over one another when soil exhibits plastic properties (see Campbell 1991). Although this explanation seems correct intuitively, evidence to support it is sparse. For example, a water film with a mean thickness of approximately six molecular layers still exists at a pore water suction of 15000 hPa, and this water film will cover all sizes of particles present (Dexter and Richard 2009). Nevertheless, it has been reported that sandy soils are not plastic and do not have plastic limits (e.g. Dexter and Bird 2001). We conjecture that it may not be the absolute thickness of the water film that is critical but the ratio of the film thickness to the particle size. This needs to be tested experimentally.
We made an attempt to define a limit, in terms of clay content, below which soils are not plastic and do not, therefore, have plastic limits. We can assume that a soil without a plastic range would have a value for PI of zero. However, Eqn 7 predicts negative values of clay content at PI = 0 (Fig. 1c). Although this is an extrapolation, which should be interpreted carefully, a plausible explanation for this behaviour may be that all of our soils have OM greater than zero, and OM has been shown to increase PI (Fig. 2d).
Therefore, we subtracted the effect of OM on PI in order to obtain an 'OM-neutralised' PI, [PI.sup.*], which is dependent only on soil texture. [PI.sup.*] was calculated as:
[PI.sup.*] = PI - 1.87 OM (16)
where the factor for OM (1.87) was obtained from Eqn 12. Regression between [PI.sup.*] and clay content yielded (Fig. 4):
[PI.sup.*] = -5.71 (1.96) + 0.582 (0.047) clay; [R.sup.2] = 0.759, P < 0.0001 (17)
Equation 17 yields a value for clay content of 9.8% for [PI.sup.*] = 0. Hence, we conclude that of soils with no OM, only those with clay content >10% can behave plastically. Accordingly, our data do not include any soil with clay content <11% (Table 1), because the plastic limit could not be determined for such soils.
Another approach is to intersect Eqn 5 with Eqn 6, which results in a value for clay content of 4.5%. This seems rather low, as sandy soils are referred to as non-plastic (as mentioned above). However, it is in accordance with McBride and Bober (1989), who estimated the lower clay content limit for plastic soils with an OM of 0.1% at 4.8%. They argued that it may not be possible to measure the PL of low clay content soils with the standard tests for PL, although such soils actually could behave plastically. Furthermore, data of PL found in the literature include soils with clay content <10% (Campbell 1976; McBride 1989; de Jong et al. 1990; Schjonning 1991; Seybold et al. 2008). However, in most of these studies, the drop-cone penetrometer test (Campbell 1976) was used for determination of PL, in contrast to the Casagrande plastic limit test (British Standard 1377 1975) used in our study.
Consequently, we could not unambiguously determine a lower clay content limit for PL and, hence, soil plasticity. We conclude that this limit may be dependent on factors such as the clay mineralogy and the test procedure.
The data, collected from many soil types in many countries, are remarkably consistent. Only a few soils can be considered as 'outliers', and in most of these cases the behaviour can be explained.
The plastic limits were highly related with the clay content, in accordance with previous studies. The LL was more strongly correlated with the clay content than the PL, but the reasons are unclear. An interesting observation was made that PL was not significantly affected by clay content for soils with clay contents below ~35%.
Organic matter has a strong effect on the plastic limits. This effect was especially evident when analysing soils of similar texture with a range of OM contents, e.g. soils from different treatments of long-term field experiments in which different soil management practices are compared.
We present equations (pedotransfer functions) for estimation of PL, LL, and PI from soil texture and OM. Whereas the consistency limits (PL and LL) could be predicted from clay and OM, satisfactory prediction of PI required inclusion of an additional soil textural class (sand) in the regression model. These pedotransfer functions could be valuable in the development of models of agricultural systems that include tillage as affected by climate and soil type.
We predict that the clay content must be at least 10% for soils without OM to be plastic; however, soils with <10% clay can be plastic if OM is present. However, more research is needed to better quantify the effects of OM on soil consistency.
The authors thank the INRA Soil Science Unit in Orleans, France, and its former Director, Dr Guy Richard, for permission to present the results for the French soils. These were measured by the second author during a period of research in France sponsored by the LeStudium Institute of Advanced Studies, Orleans. Dr Johannes Forkman (Swedish University of Agricultural Sciences, Uppsala) is thanked for help with the statistical analysis.
Received 20 July 2011, accepted 16 January 2012, published online 20 February 2012
Archer JR (1969) "The application of Atterberg Limits to agricultural soils.' (ADAS: Shardlow, UK)
Atterberg A (1911) Uber die physikalische Bodenuntersuchung und uber die Plastizitat der Tone. Internationale Mitteilungen.fur Bodenkunde 1, 10-43.
Barzegar AR, Hashemi AM, Herbert S J, Asoodar MA (2004) Interactive effects of tillage system and soil water content on aggregate size distribution for seedbed preparation in Fluvisols in southwest Iran. Soil & Tillage Research 78, 45-52. doi: 10.1016/j.still.2004.01.002
Blackmore AV (1976) Subplasticity in Australian soils. IV. Plasticity and structure related to clay cementation. Australian Journal of Soil Research 14, 261-272. doi:10.1071/SR9760261
Blanco-Canqui H, Lal R, Post WM, lzaurralde RC, Shipitalo MJ (2006) Organic carbon influences on soil particle density and rheological properties. Soil Science Society of America Journal 70, 1407-1414. doi: 10.2136/sssaj2005.0355
Braunack MV, Hewitt JS, Dexter AR (1979) Brittle fracture of soil aggregates and the compaction of aggregate beds. Journal of Soil Science 30, 653-667. doi: 10.1111/j.1365-2389.1979.tb01015.x
British Standard 1377 (1975) 'Methods of testing soils for civil engineering purpose.' (British Standards Institute: London)
Butler BE (1955) A system for the description of soil structure and consistence in the field. Journal of the Australian Institute of Agricultural Sciences 24, 239-249.
Butler BE (1956) Parna--an aeolian clay. Australian Journal of Science 18, 145-151.
Campbell DJ (1976) Plastic limit determination using a drop-cone penetrometer. Journal of Soil Science 27, 295-300. doi:10.1111/j. 1365-2389.1976.tb01999.x
Campbell DJ (1991) Liquid and plastic limits. In 'Soil analysis physical methods'. (lids KA Smith, CE Mullins) pp. 367-398. (Dekker Inc.: New York)
Cerato AB (2001) Influence of specific surface area on geotechnical characteristics of fine-graded soils. MSc Thesis, Department of Civil and Environmental Engineering, University of Massachusetts, Amherst, MA, USA.
Croney D, Coleman JD (1954) Soil structure in relation to soil suction (pF). Journal of Soil Science 5, 75-84. doi:10.1111/j.1365-2389.1954.tb02177.x
de Jong E, Acton DF, Stonehouse HB (1990) Estimating the Atterberg limits of southern Saskatchewan soils from texture and carbon contents. Canadian Journal of Soil Science 70, 543-554.
Dexter AR (1972) 'Penetration of spheres into soil: Part 1. Measurement techniques in the field and experimental results.' Departmental Note DN/ ER/198/1162. (National Institute of Agricultural Engineering: Silsoe, UK)
Dexter AR (1977) Effects of rainfall on the surface micro-relief of tilled soil. Journal of Terramechanics 14, 11-22. doi: 10.1016/0022-4898(77)90023-4
Dexter AR (1990) Changes in the matric potential of soil water with time after disturbance by moulding. Soil & Tillage Research 16, 35-50. doi: 10.1016/0167-1987(90)90020-E
Dexter AR, Bird NRA (2001) Methods for predicting the optimum and the range of water contents for tillage based on the water retention curve. Soil & Tillage Research 57, 203 212. doi: 10.1016/S0167-1987(00)00154-9
Dexter AR, Richard G (2009) Water potentials produced by oven-drying of soil samples. Soil Science Society of America Journal 73, 1646-1651. doi:l0.2136/sssaj2008.0294N
Dexter AR, Tanner DW (1973) The response of unsaturated soils to isotropic stress. Journal of Soil Science 24, 491-502. doi:10.1111/j.1365-2389.1973.tb02316.x
Dexter AR, Kroesbergen B, Kuipers H (1984a) Some mechanical properties of top soils from the IJsselmeer polders. I. Undisturbed soil aggregates. Netherlands' Journal of Agricultural Science 32, 205-214.
Dexter AR, Kroesbergen B, Kuipers H (1984b) Some mechanical properties of top soils from the IJsselmeer polders. 2. Remoulded soil aggregates and the effects of wetting and drying cycles. Netherlands Journal of Agricultural Science 32, 215-227.
Dexter AR, Horn R, Kemper WD (1988) Two mechanisms for agehardening of soil. Journal of Soil Science 39, 163-175. doi:10.1111/j.1365-2389.1988.tb01203.x
Dexter AR, Arvidsson J, Czyz EA, Trautner A, Stenberg B (1999) Respiration rates of soil aggregates in relation to tillage and straw-management practices in the field. Acta Agriculturae Scandinavica, Section B 49, 193 200. doi:10.1080/713782029
Dexter AR, Richard G, Arrouays D, Czyz EA, Jolivet C, Duval O (2008a) Complexed organic matter controls soil physical properties. Geoderma 144, 620 627. doi:10.1016/j.geoderma.2008.01.022
Dexter AR, Czyz EA, Richard G, Reszkowska A (2008b) A user-friendly water retention function that takes account of the textural and structural pore spaces in soil. Geoderma 143, 243 253. doi:10.1016/j.geoderma.2007.11.010
Grant CD, Watts CW, Dexter AR, Frahn BS (1995) An analysis of the fragmentation of remoulded soils with regard to self-mulching behaviour. Australian Journal of Soil Research 33, 569 583. doi:10.1071/SR9950569
Imhoff S, da Silva AP, Fallow D (2004) Susceptibility to compaction, load support capacity and soil compressibility of Hapludox. Soil Science Society of America Journal 68, 170-24. doi: 10.2136/sssaj2004.0017
ISO 10694 (1995) 'Soil quality. Determination of organic and total carbon after dry combustion (elementary analysis)." (International Standard Organization: Geneva)
Keller T, Hakansson I (2010) Estimation of reference bulk density from soil particle size distribution and soil organic matter content. Geoderma 154, 398-406. doi:10.1016/j.geoderma.2009.11.013
Keller T, Arvidsson J, Dexter AR (2007) Soil structures produced by tillage as affected by soil water content and the physical quality of soil. Soil & Tillage Research 92, 45-52. doi:10.1016/j.still.2006.01.001
Lambe TW, Whitman RV (1969) 'Soil mechanics.' (John Wiley and Sons, Inc.: New York)
Lang H-J, Huder J, Amann P (1996) 'Bodenmechanik und Grundbau. Das Verhalten von Brden und Fels und die wichtigsten grundbaulichen Konzepte.' 320 pp. (Springer-Verlag: Berlin)
Larson WE, Gupta SC, Useche RA (1980) Compression of agricultural soils from eight soil orders. Soil Science Society of America Journal 44, 450-457. doi:10.2136/sssaj1980.03615995004400030002x
Ljung G (1987) 'Mekanisk analys: beskrivning aven rationelt metod for jordardsbestamning (in Swedish) (Mechanical analysis: description of a rational method for soil textural analysis).' Avdelningsmeddelande 87 : 2. (Institutionen for Markvetenskap: Uppsala, Sweden)
Mbagwu JSC, Bazzoffi P (1987) Estimating the consistency limits moisture retention and availability of some Italian soils from physical properties. Annali Istituto Sperimentule Studio e Difesa Suolo, Firenze 18, 111-120.
McBride RA (1989) A re-examination of alternative test procedures for soil consistency limit determination: I. A simulated desorption procedure. Soil Science Society of America Journal 53, 184-191. doi:10.2136/ sssaj1989.03615995005300010034x
McBride RA (2008) Soil consistency: upper and lower plastic limits. In 'Soil sampling and methods of analysis'. 2nd edn, Ch. 58. (Eds MR Carter, EG Gregorich) pp. 761-769. (CRC Press: Boca Raton, FL)
McBride RA, Bober ML (1989) A re-examination of alternative test procedures for soil consistency limit determination: I. A compression-based procedure. Soil Science Society of America Journal 53, 178-183. doi:10.2136/sssaj1989.03615995005300010033x
McIntyre DS (1976) Subplasticity in Australian soils, I Description, occurrence, and some properties. Australian Journal of Soil Research 14, 227-236. doi:10.1071/SR9760227
McKenzie BM, Dexter AR (1985) Mellowing and anisotropy induced by wetting of moulded soil samples. Australian Journal of Soil Research 23, 37-47. doi:10.1071/SR9850037
Muller L, Sebindler U, Fausey NR, Lal R (2003) Comparison of methods for estimating maximum soil water content for optimum workability. Soil & Tillage Research 72, 9-20. doi:10.1016/S0167-1987(03)00046-1
Ojeniyi SO, Dexter AR (1979a) Soil factors affecting the macrostructures produced by tillage. Transactions of the American Society of Agricultural Engineers 22, 339-343.
Ojeniyi SO, Dexter AR (1979b) Soil structural changes during multiple pass tillage. Transactions of the American Society of Agricultural Engineers 22, 1068-1072.
Schjonning P (1991) Soil mechanical properties of seven Danish soils. Danish Institute of Agricultural Sciences, Internal Report No. S2176, Research Centre Foulum, Tjele, Denmark, 33 pp.
Seybold CA, Elrashidi MA, Engel RJ (2008) Linear regression models to estimate soil liquid limit and plasticity index from basic soil properties. Soil Science 173, 25-34. doi: 10.1097/ss.0b013e318159a5e1
Sills D, Aylmore LAG, Quirk JP (1973) An analysis of pore size distribution in illite-kaolinite mixtures. Journal of Soil Science 24, 480-490. doi:10.1111/j.1365-2389.1973.tb02315.x
Skempton AW (1953) The colloidal 'activity' of clays. In 'Proceedings of the 3rd International Conference on Soil Mechanics and Foundation Engineering'. Switzerland, Vol. 1, pp. 57-60.
Stanchi S, Oberto E, Freppaz M, Zanini E (2009) Linear regression models for liquid and plastic limit estimation in Alpine soils. Agrochimica 53, 322-338.
Utomo WH, Dexter AR (1981a) Soil friability. Journal of Soil Science 32, 203-213. doi:10.1111/j.1365-2389.1981.tb01700.x
Utomo WH, Dexter AR (1981b) Tilth mellowing. Journal of Soil Science 32, 187-201. doi:10.1111/j.1365-2389.1981.tb01699.x
Watts CW, Dexter AR (1997) The influence of organic matter in reducing the destabilization of soil by simulated tillage. Soil & Tillage Research 42, 253-275. doi:10.1016/S0167-1987(97)00009-3
Watts CW, Dexter AR (1998) Soil friability: theory, measurement and the effects of management and organic carbon content. European Journal of Soil Science 49, 73-84. doi: 10.1046/j.1365-2389.1998.00129.x
Watts CW, Dexter AR, Chamen WCT, Longstaff DJ (1995) Pre-inundation characterisation of soil physical and chemical parameters on the Tollesbury experimental set-back site. Silsoe Research Institute Contract Report CR/643/95/1326.
Wroth CP, Wood DM (1978) The correlation of index properties with some basic engineering properties of soils. Canadian Geotechnical Journal 15, 137-145.
Thomas Keller (A,B,D) and Anthony R. Dexter (C)
(A) Agroscope Reckenholz-Tanikon Research Station ART, Department of Natural Resources and Agriculture, Reckenholzstrasse 191, CH-8046 Zurich, Switzerland.
(B) Swedish University of Agricultural Sciences, Department of Soil & Environment, Box 7014, SE-75007 Uppsala, Sweden.
(C) Institute of Soil Science and Plant Cultivation (IUNG-PIB), ul. Czartoryskich 8, 24-100 Pulawy, Poland. Email: email@example.com
(D) Corresponding author. Emails: firstname.lastname@example.org; email@example.com
Table 1. Details of experimental soils: location, lower plastic limit (PL) and liquid limit (LL), and contents of sand (50-20001.[micro],m), silt (2-50 [micro]m), clay (<2 [micro],m) and organic matter (OM) References: a, Dexter (1972); b, Dexter and Tanner (1973); c, Dexter (1977); d, Ojeniyi and Dexter (F 979b); e, Braunack et al. (1979); f, Utomo and Dexter (1981 b); g, Dexter et al. (I 984a); h, Dexter et al. (F 9846); i, McKenzie and Dexter (1985); j, Dexter et al. (1988); k, Dexter (1990); l, Grant et al. (1995); m, Watts and Dexter (1997); n, Dexter et al. (1999); o, Watts et al. (1995); p, Dexter et al. (1988); q, AR Dexter (unpubl. data); r, AR Dexter (unpubl. data); s, Keller et al. (2007); t, T Keller and AR Dexter (unpubl. data); u, AR Dexter (unpubl. data) Soil no. Soil name Country Location PL Topsoils 1 Cageside England Silsoe, Bedfordshire 27.0 2 A6 England Silsoe, Bedfordshire 41.0 3 Steppingly(4) England Bedfordshire 33.0 4 Clophill(5) England Bedfordshire 17.0 5 Urrbrae Australia South Australia 19.5 6 Mortlock(I) Australia South Australia 22.5 7 Waco Australia Queensland 54.0 8 Warwick Australia Queensland 43.0 9 Mywybilla Australia Queensland 33.0 10 Allora Australia Queensland 40.0 11 Strathalbyn Australia South Australia 17.9 12 Polder I Netherlands NE Polder 27.6 13 Polder 2 Netherlands NE Polder 23.1 14 Polder 3 Netherlands E Flevopolder 26.2 15 Polder 4 Netherlands E Flevopolder 29.2 16 Polder 5 Netherlands E Flevopolder 32.6 17 Polder 6 Netherlands S Flevopolder 31.9 18 Polder 7 Netherlands S Flevopolder 33.9 19 Polder 8 Netherlands S Flevopolder 26.7 20 Mortlock 2 Australia South Australia 19.5 21 Wiesenboden Australia South Australia 24.5 22 Condoblin Australia New South Wales 13.7 23 Georgetown Australia South Australia 31.4 24 Bowenville Australia Queensland 36.7 25 Walla Walla USA Pendleton, Oregon 26.4 26 Billings USA Grand Jn., Colorado 21.6 27 Portneuf USA Kimberly, Idaho 23.7 28 Wiesenboden 2 Australia South Australia 38.3 29 Mintaro Australia South Australia 21.5 30 Portneuf 2 USA Kimberly, Idaho 21.6 31 Lawford I England Shuttleworth College 27.0 32 Lawford 2 England Shuttleworth College 27.0 33 Evesham 1 England Silsoe, Bedfordshire 48.0 34 Evesham 2 England Silsoe, Bedfordshire 38.0 35 Wickham I England Silsoe, Bedfordshire 23.0 36 Wickham 2 England Silsoe, Bedfordshire 20.0 37 Hanslope 1 England Newport Pagnell 25.0 38 Hanslope 2 England Newport Pagnell 23.0 39 Denchworth England Faringdon, Oxn. 31.0 40 Brimstone England Faringdon, Oxn. 27.0 41 Fellingsbro Sweden Orebro lan 45.0 42 Highfield 1 England Rothamsted 18.2 43 Highfield 2 England Rothamsted 23.0 44 Highfield 3 England Rothamsted 25.4 45 Highfield 4 England Rothamsted 34.0 46 Highfield 5 England Rothamsted 34.4 47 Lanna Sweden Vdstra Gotaland 28.0 48 Tollesbury 1a England Essex coast 35.3 49 Tollesbury 1b England Essex coast 33.4 50 Tollesbury 2a England Essex coast 30.9 51 Tollesbury 2b England Essex coast 35.1 52 Tollesbury 3a England Essex coast 31.9 53 Tollesbury 3b England Essex coast 31.5 54 Tollesbury 4a England Essex coast 33.0 55 Tollesbury 4b England Essex coast 29.0 56 Tollesbury 5a England Essex coast 26.5 57 Tollesbury 5b England Essex coast 26.1 58 Bindlach AI Germany Bavaria 19.0 59 Hohe Warte P Germany Bavaria 30.0 60 Avdat Israel Negev desert 18.0 61 Vreta Kloster Sweden Ostergotland 26.5 62 Fjdrdingsl6v Sweden Malmohus 16.7 63 Bjertorp Sweden Skaraborg 23.3 64 Kungsdngen Sweden Uppland 32.4 65 Landskrona Sweden Malmohus 16.8 66 Villamblain France North-central France 27.6 67 Boigneville A France North-central France 22.8 68 Boigneville B France North-central France 34.5 69 Faux Perche France North-central France 24.4 70 Ultuna Lerhdl Sweden SLU, Uppsala 27.6 71 UltunaR2-411 I Sweden SLU, Uppsala 33.4 72 Sdby I Sweden SLU, Uppsala 26.9 73 Saby II Sweden SLU, Uppsala 32.9 74 Ultuna An S Sweden SLU, Uppsala 24.4 75 Ultuna An N Sweden SLU, Uppsala 19.4 76 Saby III Sweden SLU, Uppsala 26.9 77 Sdby IV Sweden SLU, Uppsala 26.6 78 Kepa Poland IUNG, Pulawy 26.9 Max 54.0 Min 13.7 79 Urrbrae B Australia South Australia 19.7 80 Tollesbury lc England Essex coast 31.4 81 Tollesbury 1 d England Essex coast 29.3 82 Tollesbury 2c England Essex coast 32.3 83 Tollesbury 2d England Essex coast 32.3 84 Tollesbury 3c England Essex coast 33.0 85 Tollesbury 3d England Essex coast 34.1 86 Tollesbury 4c England Essex coast 29.0 87 Tollesbury 4d England Essex coast 25.7 88 Tollesbury 5c England Essex coast 20.5 89 Tollesbury 5d England Essex coast 17.2 Max 34.1 Min 17.2 Soil no. Soil name LL Sand Silt Clay OM Ref. 1 Cageside 37.0 11.0 3.0 a 2 A6 78.0 64.0 4.0 a 3 Steppingly(4) 54.0 25.0 6.5 b 4 Clophill(5) 29.0 23.0 3.7 6 5 Urrbrae 26.5 51.0 32.0 17.0 2.0 c 6 Mortlock(I) 55.0 30.0 15.0 d 7 Waco 86.0 74.0 e 8 Warwick 85.0 61.0 e 9 Mywybilla 54.0 48.0 e 10 Allora 77.0 58.0 e 11 Strathalbyn 30.0 12.0 2.8 f 12 Polder I 39.7 14.3 59.1 26.6 3.2 g,h 13 Polder 2 28.7 33.2 52.4 14.4 2.4 g,h 14 Polder 3 33.9 14.8 66.3 18.9 3.1 g,h 15 Polder 4 40.6 13.0 61.7 25.2 3.3 g,h 16 Polder 5 51.3 10.7 48.5 40.8 4.4 g,h 17 Polder 6 49.0 11.4 49.0 39.6 3.3 g,h 18 Polder 7 53.6 8.7 48.5 42.8 4.4 g,h 19 Polder 8 28.6 40.8 44.1 15.1 2.9 g,h 20 Mortlock 2 36.5 18.0 i 21 Wiesenboden 41.5 22.0 i 22 Condoblin 23.7 33.0 i 23 Georgetown 50.8 56.0 i 24 Bowenville 59.9 67.0 1 25 Walla Walla 21.8 55.9 22.3 2.0 j 26 Billings 45.9 30.4 23.7 1.0 j 27 Portneuf 18.0 58.9 23.1 0.7 j 28 Wiesenboden 2 81.0 17.4 15.8 66.8 2.4 k 29 Mintaro 30.1 26.0 54.0 20.0 2.8 k 30 Portneuf 2 31.5 19.2 52.5 28.3 1.7 k 31 Lawford I 55.0 22.8 31.2 46.0 I 32 Lawford 2 56.0 24.0 34.0 42.1 I 33 Evesham 1 70.0 11.3 17.8 70.9 I 34 Evesham 2 92.0 5.9 20.6 73.5 I 35 Wickham I 41.0 38.6 31.6 28.9 I 36 Wickham 2 42.0 38.6 32.1 29.2 I 37 Hanslope 1 40.0 30.7 30.9 38.2 I 38 Hanslope 2 49.0 29.2 19.7 51.4 I 39 Denchworth 71.0 10.5 34.3 55.3 I 40 Brimstone 74.0 10.0 32.5 56.0 I 41 Fellingsbro 65.0 2.2 28.7 68.9 I 42 Highfield 1 32.9 9.0 67.0 25.0 1.9 m 43 Highfield 2 40.0 13.0 63.0 24.0 2.6 m 44 Highfield 3 40.0 11.0 64.0 25.0 3.6 m 45 Highfield 4 53.5 11.0 63.0 27.0 4.8 m 46 Highfield 5 56.5 11.0 67.0 23.0 5.4 m 47 Lanna 43.0 3.1 n 48 Tollesbury l a 68.6 6.0 39.0 55.0 6.9 0 49 Tollesbury lb 69.0 4.0 41.0 55.0 5.3 0 50 Tollesbury 2a 60.9 1.0 42.0 57.0 6.5 0 51 Tollesbury 2b 61.1 1.0 42.0 57.0 6.3 0 52 Tollesbury 3a 61.9 6.0 42.0 52.0 4.5 0 53 Tollesbury 3b 64.0 8.0 41.0 51.0 5.1 0 54 Tollesbury 4a 61.3 7.0 42.0 51.0 4.4 0 55 Tollesbury 4b 63.4 8.0 41.0 51.0 4.8 0 56 Tollesbury 5a 53.0 9.0 49.0 44.0 3.9 0 57 Tollesbury 5b 54.9 10.0 46.0 44.0 4.7 0 58 Bindlach AI 30.0 19.0 64.0 17.0 0.2 p 59 Hohe Warte P 87.0 1.0 26.0 73.0 0.2 p 60 Avdat 24.0 25.0 56.0 19.0 0.7 p 61 Vreta Kloster 8.9 45.6 45.5 3.1 q 62 Fjdrdingsl6v 18.0 2.0 q 63 Bjertorp 14.4 55.2 30.4 3.3 q 64 Kungsdngen 4.1 39.9 56.0 1.7 q 65 Landskrona 15.7 3.6 q 66 Villamblain 43.4 1.7 65.2 33.1 2.2 r 67 Boigneville A 37.1 8.0 66.0 26.0 2.3 r 68 Boigneville B 53.9 7.8 68.6 23.6 5.0 r 69 Faux Perche 33.2 5.3 82.9 11.8 2.1 r 70 Ultuna Lerhdl 25.3 29.5 45.2 2.2 s 71 UltunaR2-411 I 10.9 31.6 57.5 3.4 s 72 Sdby I 34.6 42.4 23.0 4.2 s 73 Saby II 10.9 49.8 39.3 4.4 s 74 Ultuna An S 31.9 31.7 36.4 2.0 t 75 Ultuna An N 48.0 26.8 25.2 1.6 t 76 Saby III 38.2 40.8 21.0 4.1 t 77 Sdby IV 30.8 44.5 24.7 3.4 t 78 Kepa 11.0 65.0 24.0 2.1 u Max 92.0 55.0 82.9 74.0 6.9 Min 23.7 1.0 15.8 11.0 0.2 79 Urrbrae B 42.8 18.5 35.5 46.0 0.8 k 80 Tollesbury lc 70.0 2.0 41.0 57.0 3.1 0 81 Tollesbury 1 d 71.2 2.0 41.0 57.0 2.0 0 82 Tollesbury 2c 61.1 1.0 41.0 58.0 4.6 0 83 Tollesbury 2d 63.9 2.0 46.0 52.0 3.6 0 84 Tollesbury 3c 66.3 8.0 41.0 51.0 4.5 0 85 Tollesbury 3d 72.6 2.0 40.0 58.0 1.5 0 86 Tollesbury 4c 67.9 6.0 41.0 53.0 2.6 0 87 Tollesbury 4d 55.8 11.0 45.0 44.0 1.9 0 88 Tollesbury 5c 43.2 11.0 57.0 32.0 2.2 0 89 Tollesbury 5d 32.0 10.0 68.0 22.0 1.7 0 Max 72.6 18.5 68.0 58.0 4.6 Min 32.0 1.0 35.5 22.0 0.8
|Printer friendly Cite/link Email Feedback|
|Author:||Keller, Thomas; Dexter, Anthony R.|
|Date:||Feb 1, 2012|
|Previous Article:||Prediction of the CEC to clay ratio using mid-infrared spectroscopy.|
|Next Article:||Random sampling of stony and non-stony soils for testing a national soil carbon monitoring system.|