Colour variation in standard soil-colour charts.
Colour is a distinctive feature of soil. Early efforts to set soil-colour standards and terms for field use followed the tradition of naturalists, who since the 17th Century used colour charts to describe plants and animals (Paclt 1983). The first sets of standard soil-colour charts were published in the 1940s using cards arrayed in a loose-leaf notebook (Simonson 1993; Oyama and Takehara 2001). Like the current soil-colour charts, those charts consisted of artificially coloured papers (chips) mounted on constant hue cards, showing chroma and lightness variations in horizontal and vertical directions, respectively, as in the Munsell system (Wyszecki and Stiles 1982). Soil colours are determined by visual comparison, seeking the closest match between the soil sample and one of the standard chips in the soil-colour charts. Thus, the Munsell designation of this chip (hue, value, and chroma) is assigned to the soil sample under study.
Currently, soil colour can be measured objectively by colourimeters or spectrometers, which follow numerical specifications in colour-representation systems such as CIE x,y,Y, CIELAB, or OSA (Melville and Atkinson 1985; Torrent and Barron 1993; Sanchez-Maranon et al. 2004). However, in normal practice, the most widely used method for soil-colour specification is the use of standard soil-colour charts following the Munsell notation (Landa 2004). The colour measured in this way has far-reaching implications for field examination and description of soils (Soil Survey Staff 1993). Moreover, this method is used by all soil-classification systems as an influential diagnostic criterion for identifying the taxonomic class of a soil (FAO 1998; Soil Survey Staff 1999) as well as in studying soil formation and evaluation (Dahlgren et al. 1997; Brouwer and Fitzpatrick 2002; Brown et al. 2004).
Several problems have been described previously in relation to identifying the colour of soil specimens using standard soil-colour charts. Thus, colour determined by this technique is affected by the spectral properties and amount of light illuminating the samples, viewing angle, status of the soil sample (e.g. size, moisture content, and texture), subjectivity of the human observer, and the limited number of available chips to achieve a perfect colour match with the soil sample (Shields et al. 1966; Baumgardner et al. 1985; Sanchez-Maranon et al. 1995). On the other hand, the cylindrical structure of the Munsell system, which has a circular scale for hue and two different linear scales for value and chroma, makes Munsell variables difficult to use in certain numerical analyses. Appropriate recommendations have been outlined in order to minimise some of these problems inherent to soil-colour measurement and its application (Coventry and Robinson 1981; Melville and Atkinson 1985; Soil Survey Staff 1993).
These problems are further complicated by the existence of 2 sets of colour chips currently in use: the Munsell Soil Colour Charts, published by the Munsell Colour Co. in USA, and the Revised Standard Soil Colour Charts, manufactured by Fujihira Industry Co. in Japan. Given that the Japanese charts adapted the earlier Munsell charts to match local soils (Oyama and Takehara 2001), not only Japanese researchers have adopted them, but also scientists from many other countries. The reprinting of American and Japanese soil-colour charts has continued since the middle of the 20th Century, replacing and adding colour chips as well as manufacturing them with improved pigments. Today, several editions of American and Japanese charts are being indifferently used by the soil-science community as though they were identical. In fact, noting the subtle differences between the 2 systems, we have found an apparent improvement in the accuracy of visual determinations performed with the Japanese charts when the colour of the soil samples is close to the limits of the gamut provided by the charts, or in the reddish region from 7.5YR to 2.5YR (Sanchez-Maranon et al. 1995). Blackburn (1973) also pointed out that some chips with identical designation in the Fujihira and Munsell charts do not entirely match. Significant differences among editions of other well known colour charts have also been reported in recent literature, as is the case of the Ishihara plates, widely used to detect colour-vision deficiencies (Lee and Honson 2003).
The present study stems from visual results that suggested a potential influence of both the manufacturer and the edition of soil-colour charts on the colour specification of soil samples. Here, we seek to evaluate the accuracy and reliability of our soil-colour charts from different USA and JAPAN editions. To do so, we analyse: (1) the conformity between the colour measured on the chips and the Munsell notation shown below each chip; (2) the visual uniformity in colour space for chip sets provided by different soil-colour charts; and (3) the colour differences between chips with identical Munsell notation from different soil-colour charts.
Materials and methods
We used 6 standard soil-colour charts. These included 2 American editions of the Munsell Soil Colour Charts (Munsell Colour Company 1975, 2000), with 2 different books for the 1975 edition, and 3 Japanese editions of the Revised Standard Soil Colour Charts (Fujihira Industry Company 1970, 1987, 2001). The charts contained pages for hues of 10R, 2.5YR, 5YR, 7.5YR, 10YR, 2.5Y, and 5Y plus the 7.5R page for the Japanese editions. These charts and editions were selected on the basis of availability in our laboratories, representing examples of standard soil-colour charts published over an extensive time period (31 years). These allowed us to test the discrepancies between American and Japanese charts, as well as variations between different editions from the same manufacturer. Henceforth we shall use the following designations for the 6 soil-colour charts studied: USA1975a, USA1975b, USA2000, JAPAN1970, JAPAN1987, and JAPAN2001.
The most recent collections of soil-colour charts (USA2000 and JAPAN2001) were completely new. Colour measurements of their chips were performed immediately after opening the sealed packages. Because the other soil-colour charts (JAPAN1970, USA1975a, USA1975b, and JAPAN1987) were previously used for soil-colour assessment in the field and laboratory, their chips were cleaned by wiping off the dirt lightly with a clean wet cloth before colour measurements. All these chips appeared to be in good condition (i.e. no noticeable damage).
First, a set of visual colour measurements was made for 10 selected soil samples located in very different positions in the colour space using the chips from JAPAN1970, USA1975a, USA1975b, and JAPAN1987. Colour determinations were performed following the recommendations provided by the charts themselves and by the Soil Survey Staff (1993). Ten experienced observers, men and women with non-defective colour vision according to conventional tests (Farnsworth 1943; Ishihara 1979), performed the experiment in a VeriVide colour-assessment cabinet with a D65 (daylight) standard light source. To avoid learning effects during visual assessments, each observer made determinations for all the soil samples using the soil-colour charts one at a time at 1-week intervals.
Second, we performed spectral-reflectance measurements from 360 to 740 nm at 10-nm intervals, for each of the chips on our 6 soil-colour charts (except those with chroma notation/0 in the USA 1975 edition) using a Minolta CM-2600d spectrophotometer. This instrument used the illuminating/viewing geometry d/8 (diffused illumination, 8 degrees viewing angle), making measurements of the light reflected by the specimen surface with the specular component excluded. Because some of the chips in our soil-colour charts were made of glossy or semi-glossy materials, we excluded the first-surface reflection in instrumental measurements in order to achieve closer correlation with visual appearance (Berns 2000). Each chip was measured at 3 different positions (top left corner, middle-fight, and bottom left corner), their averages being used for calculating colour parameters, and their standard deviation for assessment of spatial repeatability. In each position, the instrument automatically made three consecutive measurements on a surface area of 21.3 [mm.sup.2], which were averaged to provide spectral-reflectance values.
We used the CIE standard illuminants D65 and C, and the CIE 1931 and CIE 1964 Standard Observers (CIE 1986), to calculate different colourimetric parameters within the Munsell system and CIELAB colour space. Munsell data were tabulated only with respect to illuminant C and CIE 1931 Standard Observer, while the CIELAB colour space was used mainly for industrial applications involving D65 illuminant and CIE 1964 Standard Observer. Within the CIELAB space, we recorded both rectangular coordinates [a.sup.*][b.sup.*][L.sup.*] and cylindrical polar coordinates [h.sub.ab][L.sup.*][C.sup.*.sub.ab]. From measured reflectance values, the Spectramagic software (available with our spectrophotometer) provided different colour parameters, including Munsell hue, value, and chroma.
We also tested the short-term repeatability of our spectrophotometer by making 10 successive measurements on the centre of a few selected chips (10R 4/8, 5YR 5/3, 10YR 8/8, 2.5Y 8/1, and 5Y 4/2) of the USA2000 charts. From these measurements the variation (standard deviation) for any CIELAB coordinate was <0.022 units, except for [h.sub.ab], which was <0.048 degrees. The middle-term repeatability was tested by measuring the central point of these chips 5 times at 10-min intervals. Now the variation (standard deviation) for any CIELAB coordinate was found to be <0.054. Finally, we tested the long-term repeatability by measuring the central point of the chips 3 times at 1-day intervals. In this case the variation (standard deviation) of any CIELAB coordinate was <0.053 units, except for [h.sub.ab], which was <0.083 degrees. The short-term repeatability represents the variation of colour measurements without moving the instrument, middle-term repeatability indicates the changes owing to lifting of the apparatus and going back on the measurement point, and long-term repeatability accounts for the effects of calibration at the beginning of each measurement session.
The conformity between measured colour and Munsell hue, value, and chroma notation shown below each chip was first analysed. All chips were arranged by groups with identical notation of hue, value or chroma in order to analyse measured colour against the Munsell notation tabbed in the charts. We calculated the mean Munsell hue of a given chart from the average Munsell hues measured in its chips, using the appropriate partition of Munsell hue (10R = 10, 2.5YR = 12.5, ..., 5Y = 25). Similarly, we compared the average measured and tabbed values for different Munsell value and chroma.
For our different soil-colour charts, the Munsell loci of constant hue and chroma were plotted in the [a.sup.*][b.sup.*] plane of CIELAB (CIE 1986) in order to determine their visual uniformity. Each point on the [a.sup.*][b.sup.*] plane was averaged from the measured chromatic coordinates of the colour chips with the same Munsell hue and chroma notation (that is, the locus refers to a central Munsell value). We calculated CIELAB hue differences ([DELTA][H.sup.*.sub.ab]) between chips of adjacent charts with identical Munsell chroma and value notation, as well as CIELAB chroma differences ([DELTA][C.sup.*.sub.ab]), and lightness differences ([Delta][L.sup.*]) between neighbouring chips at constant Munsell value and hue, and constant Munsell chroma and hue, respectively. In summary, the average [DELTA][H.sup.*.sub.ab], [DELTA][C.sup.*.sub.ab], and [DELTA][L.sup.*] is the average distance between neighbouring Munsell hues, chromas, and values, respectively.
Colour differences for pairs of chips with identical Munsell notation but from different soil-colour charts were also analysed. To provide a single value that indicates the total colour difference for pairs identically designated, we used the Euclidean CIELAB distance ([DELTA][E.sup.*.sub.ab]) between colours situated in the [a.sup.*][b.sup.*][L.sup.*] space (CIE 1986). This [DELTA][E.sup.*.sub.ab] is given by the equation:
(1) [DELTA][E.sup.*.sub.ab] = [[[([DELTA][L.sup.*]).sup.2] + [([DELTA][a.sup.*]).sup.2] + [([DELTA][b.sup.*]).sup.2]].sup.1/2] = [[[([DELTA][L.sup.*]).sup.2] + [([DELTA][C.sup.*.sub.ab]).sup.2] + [([DELTA][H.sup.*.sub.ab]).sup.2]].sup.1/2]
We used Seve's equation (Seve 1991) to calculate [DELTA][H.sup.*.sub.ab]. The [DELTA][H.sup.*.sub.ab] between 2 points having CIELAB hue-angles [h.sub.ab,1], [h.sub.ab,2] and CIELAB chromas [C.sup.*.sub.ab,1], [C.sup.*.sub.ab,2] is given by:
(2) [DELTA][H.sup.*.sub.ab] = 2 [square root of [C.sup.*.sub.ab,1][C.sup.*.sub.ab,2]] sin ([h.sub.ab,2] - [h.sub.ab,1]/2)
Because colour differences found in our research reached 5.0 CIELAB units, the use of [DELTA][E.sup.*.sub.ab] is correct. However, when smaller colour differences appear and we look for an improved correlation with visually perceived differences, we must follow the latest CIE recommendation using the new CIEDE2000 colour-difference formula (CIE 2001) given by the equation:
(3) [DELTA][E.sub.00] = [[[([DELTA]L'/[k.sub.L][S.sub.L]).sup.2] + [([DELTA]C'/[k.sub.C][S.sub.C]).sup.2] + [([DELTA]H'/[k.sub.H][S.sub.H]).sup.2] + [R.sub.T] ([DELTA]C'/[k.sub.C][S.sub.C]) ([DELTA]H'/[k.sub.H][S.sub.H])].sup.0.5]
where [DELTA]L', [DELTA]C', and [DELTA]H' are parameters transformed from CIELAB; [S.sub.L], [S.sub.C], and [S.sub.H] are weighting functions; [k.sub.L], [k.sub.C], and [k.sub.H] are parametric factors dependent on experimental conditions; and [R.sub.T] is a rotation function. A complete development of the CIEDE2000 formula can be seen in Luo et al. (2001).
The statistical techniques used to analyse colour variables were the analysis of variance to compare the mean values of colour parameters for the different soil-colour charts, the multiple-range test of Fisher's least significant differences to determine whether two means are significantly different, and the Kolmogoroff-Smirnof's, Cochran's, and Bartlett's tests to examine the conditions of normality and homoscedasticity (Schabenberger and Pierce 2002).
Results and discussion
Visual determinations using different soil-colour charts
As is well known by soil scientists, in most cases an exact visual match between a soil sample and one of the chips provided by a colour chart is not possible. In the present experiment, almost half of the visual judgements involved subjective interpolation between available chips. For soil samples with the lowest chroma, some of our observers indicated that they did not match any of the chips, and therefore those samples could not be classified.
Figure 1 shows the mean Munsell colour which our group of observers assigned to each soil sample. Munsell chroma and hue are plotted as polar coordinates, and a higher Munsell value is indicated by the greater size of the triangles (JAPAN 1970 and JAPAN 1987) or diamonds (USA 1975a and USA1975b). In general, for the same soil sample, variations in assigned colour notation using the different soil-colour charts are apparent, perhaps because our observers perceived visual colour differences among charts. The inter-observer variability remained similar regardless of the soil-colour charts employed. Specifically, for the 10 soil samples the average standard deviation in our visual judgements was between 0.52 and 0.64 for Munsell value and chroma, and between 1.28 and 1.39 for Munsell hue.
[FIGURE 1 OMITTED]
The high variation in Munsell notation for nearly achromatic samples (Fig. 1, samples 1 and 10) could be interpreted from the difficulties of our observers in matching these soil colours. In fact, the observers who provided judgements on these two soil samples made 'forced' choices, and thus, the agreement in colour notation among different charts might be a coincidence. For the remaining soil samples the variation in Munsell notation as a result of using different soil-colour charts showed the following trends: (i) A reddening in hue using the Japanese charts; (ii) an increase in Munsell chroma and also a slight increase in Munsell value using JAPAN1970 charts; (iii) a better agreement between Munsell notations found with the 2 American charts than that found with the 2 Japanese charts.
Because the same soil sample is not consistently classified by observers using different soil-colour charts, the corresponding variations of Munsell notation deserve a more detailed study. According to Lee and Honson (2003), 2 factors may contribute to these variations: the soil-colour charts might not have been produced identically, or the colour of the chips might have faded differently over time. These factors are considered separately below by analysing the relationship between Munsell notation and measured colour in American and Japanese charts, the colour spacing in the new and old charts, and the colour differences among the same chips in different charts.
Conformity between measured and tabulated colours
Figure 2 shows the average and range of Munsell hue from the spectrophotometric measurements we performed on the chips of our 6 colour charts, against the tabulated Munsell hue notation provided by the charts. For the newest soil-colour charts (USA2000 and JAPAN2001), the average measured Munsell hue was similar for the charts 5Y, 2.5Y, 10YR, 7.5YR, and 5YR, but not for the charts 2.5YR and 10R. Using CIELAB space, it was found (Table 1) that the charts 2.5YR and 10R in JAPAN2001 were redder ([h.sub.ab] =42.5[degrees] and 33.9[degrees]) than in USA2000 ([h.sub.ab] =44.7[degrees] and 38.1[degrees]). In addition, the measured and tabulated Munsell hue was more similar in USA2000 than in JAPAN2001. After converting the circular scale of hue to rectangular coordinates, as described by Fernandez and Schulze (1987), we found a better accuracy in hue for USA2000 ([+ or -] 0.20 units of chroma) than for JAPAN2001 ([+ or -] 0.35 units of chroma).
[FIGURE 2 OMITTED]
It is also noticeable in Fig. 2 that for JAPAN2001 the Munsell hue ranges overlap in some adjacent charts, such as 2.5YR and 10R or 10R and 7.5R. This means that 2 chips from different JAPAN2001 hue charts might have the same measured Munsell hue. This overlap was caused mainly by the chips with Munsell value 2/or 1.7/(almost non-existent in the USA2000 charts). Of course, Munsell hue overlap is a flaw in the JAPAN2001 charts, because an observer may be puzzled matching the colour of a soil sample with chips having different Munsell hue designations.
The old soil-colour charts had a worse conformity between measured and tabulated Munsell hue than did the new ones (Fig. 2). Wide ranges of measured Munsell hue might also be noticed, for example, in the 10YR, 7.5YR, 5YR, 2.5YR, 10R, and 7.5R hue charts of JAPAN1970. Except for the 10R chart of USA1975a and 1975b, the mean Munsell hue measured in all charts from 7.5R to 10YR proved more yellow than the tabulated Munsell notation, whereas the mean Munsell hue measured in the charts 2.5Y and 5Y was slightly redder. The yellowish hue of old colour charts was statistically more significant for Japanese than for American editions (Table 1), in agreement with the fact that the chips chosen by our observers to match soil-colour samples using the Japanese charts had redder Munsell hue notations (Fig. 1). We assume that Japanese soil-colour charts had yellowed with age to a greater extent than did American ones because of a more intensive use, although there might even have been differences in the original printing. For instance, in JAPAN1987, there were 110 differences between measured and tabulated Munsell hue which were >1.0 (63% in chips with chroma notation /1 or/8, or value notation 8/, 2/ or 1.7/), whereas in USA1975a and 1975b (which had a smaller number of chips at the ends of the gamut) only 21 differences were > 1.0.
Figures 3 and 4 show a comparison between measured and tabulated data for Munsell value and chroma in our soil-colour charts. Again colour variations between charts can be attributed to use and/or to printing differences. With respect to the newest charts (USA2000 and JAPAN2001), the most conspicuous printing difference is the appearance, in JAPAN2001, of some chips tabulated with Munsell value 1.7/ and 2/, as well as the glossy finish of some chips with chroma /4, /6, and /8. These chips in JAPAN2001 showed worse conformity between measured and tabulated Munsell value and chroma. This is consistent with the results of Fernandez and Schulze (1987), who also found a greater difference between measured and tabulated colour in unused chips of USA1975 when they have low value or high chroma.
[FIGURES 3-4 OMITTED]
In the old charts, the measured Munsell value was undervalued in the chips 5/, 6/, 7/, and 8/and overvalued in the chips noted with Munsell value 1.7/, 2/, 2.5, and 3/. As shown also by Table 2, these changes, measured from CIELAB lightness ([L.sup.*]), were more pronounced in JAPAN1970 and JAPAN1987, where the mean [L.sup.*] significantly differed from USA1975a and USA1975b (except for V=4). On the other hand, Fig. 4 also shows a greater loss of Munsell chroma in the used Japanese charts, although only the differences for chroma /6 and /8 proved statistically significant. These colourimetric results are consistent with the visual measurements reported in the previous section (Fig. 1); the mean Munsell notations given by our observers using the JAPAN1970 charts had greater Munsell value and chroma because of darkening and chroma fading in those chips.
Concerning the agreement between measured and tabulated values, we found that although each chip had a single designation, experimental measurements at different points of the chip are not completely identical (that is, there is a certain lack of spatial repeatability). Table 3 shows that using CIELAB the average colour variability within a chip was about 7-fold greater in old charts than in the new charts. The relatively large standard deviations shown in Table 3 for the chips of oldest charts might also indicate that colour fading over time cannot be uniform within a chip.
Uniformity steps in the soil-colour charts
Our experimental colour measurements showed irregularity of the spread in the colour space for the chips provided by the charts. Given that the guiding principle of the Munsell colour system was the equality of visual spacing (steps) for the 3 colour attributes (Wyszecki and Stiles 1982), soil-colour charts should provide systematically arranged chips at equal visual steps (Soil Survey Staff 1993).
Figure 5 illustrates the visual uniformity of our charts on the [a.sup.*][b.sup.*] plane of CIELAB. Beside efforts to achieve more uniform colour spaces from the perceptual standpoint (Cui et al. 2002), currently the Commission Internationale de l'Eclairage assumes CIELAB as an approximately uniform colour space, which can be used for our present purposes. Figure 5a shows the Munsell loci of constant hue and chroma for the chips available in soil-colour charts, theoretically computed in the [a.sup.*][b.sup.*] plane of CIELAB using a Munsell-CIE commercial conversion program (Kollmorgen 1991). As can be seen, the distribution of colours is quite regular, the lines of constant Munsell hue and contours of constant Munsell chroma defining a regularly spaced spider web (Melville and Atkinson 1985). The loci of constant Munsell hue and chroma measured in the new editions of soil-colour charts (Fig. 5b) approached this condition more accurately than did the old editions (Fig. 5c, d).
[FIGURE 5 OMITTED]
For USA2000 and JAPAN2001 the loci of constant Munsell hue and chroma approached evenly spaced straight lines and concentric circles, respectively. There were minor differences between USA2000 and JAPAN2001 in the spacing regularity of the colour chips, although in general the latter had a smaller [C.sup.*.sub.ab] (Fig. 5b). In contrast, the dots representing constant Munsell hue in the old charts (Fig. 5c, d) appear unaligned, and the dots of constant Munsell chroma have lost their circular regularity. This suggests that the colour chips within a soil-colour chart could fade over time in a non-uniform way, producing a deformed plot in the CIELAB colour space. In fact, all colour chips are not used with the same frequency, or are not exposed to dampness and direct sunlight for the same amount of time. Thus, the irregularity of colour spacing may be a consequence of the use of each soil-colour chart. In our case, a very similar plot resulted for the charts USA1975a and USA1975b (Fig. 5c), with the variations between JAPAN1970 and JAPAN1987 being more evident (Fig. 5d). This agrees with the results found in our visual experiment (Fig. 1), where a greater similarity of visual assessments was found employing the old American than the old Japanese charts.
For a given Munsell chroma, the distance between chips with adjacent hues ([DELTA][H.sup.*.sub.ab] in Fig. 5a) should be very similar, assuming hue differences as perceptually uniform. These [DELTA][H.sup.*.sub.ab] have been computed from chips of the different soil-colour charts and are listed in Table 4. Numerical data confirmed the resemblance between USA2000 and JAPAN2001 both in the size of visual hue spacing and in the uniformity of this spacing. The charts USA1975a and USA1975b had the greatest hue spacing, and a significant deformation at chroma /2 and /8. For these 2 Munsell chroma the standard deviation of [DELTA][H.sup.*.sub.ab] in USA1975a was about twice that at USA2000. In general, the smallest [DELTA][H.sup.*.sub.ab] along with a high standard deviation were found in JAPAN1970, according to the compact and irregular distribution shown in the [a.sup.*] v. [b.sup.*] diagram (Fig. 5d).
The chroma [DELTA][C.sup.*.sub.ab] and lightness [DELTA][L.sup.*] differences between adjacent chips also showed variations for the soil-colour charts (Table 5)--that is, the visual steps between adjacent colour chips were not constant. The shortening of chroma and lightness differences in JAPAN1970 and JAPAN1987 with respect to JAPAN2001, as well as the similar values of [DELTA][C.sup.*.sub.ab] and [DELTA][H.sup.*] in the 2 charts from the same manufacturer and age (USA 1975a, 1975b) could be effects of field use. On the other hand, the different values of [DELTA][C.sup.*.sub.ab] and [DELTA][L.sup.*] in the newest American and Japanese charts, which are on the average 0.5 and 1.0 CIELAB units greater in USA2000 than in JAPAN2001 (Table 5), along with the opposite behaviour of [DELTA][H.sup.*.sub.ab] in the used American and Japanese charts (increasing in the former and decreasing in the latter with respect the new charts, Table 4), can be explained only by original printing differences. Therefore, despite that most colour variation between charts could be the result of fading from use; it appears that a variation from printing differences is also notable.
Colour differences for chips with identical designation in different soil-colour charts
In this section, we quantify colour differences between chips identically designated in 2 different soil-colour charts. For these pairs of chips, we shall compute CIELAB and CIEDE2000 colour differences (Eqns 1 and 3), analysing whether these colour differences are perceptible to human observers with non-defective colour sight under optimal visual conditions. Table 6 lists the average colour differences for each pair of charts. On the average, we found colour differences ranging from 1.38 (USA1975a-USA1975b) to 5.46 CIELAB units (JAPAN1970-USA1975b). The main contributions to the whole colour difference come from chroma and lightness. Threshold (just noticeable) colour differences are ~0.5 CIELAB units, and suprathreshold colour differences reported for surface colours have an average value of about 1.75 CIELAB units (CIE 1995; Melgosa et al. 1997). Thus, the same Munsell notation might correspond to rather different colours, depending on which colour charts are employed.
If we consider charts manufactured by the same company and with a similar degree of use, for example, USA1975a and USA1975b, or JAPAN1970 and JAPAN1987, we find the smallest colour difference, in particular for the American charts. These colour differences have average values of 0.94 and 1.15 CIEDE2000 units, respectively (Table 6), which should correspond to 2 similar but visually discernible colour samples. In particular, the results shown in Table 6 for [DELTA][C.sup.*.sub.ab] in the charts JAPAN 1970 and JAPAN 1987 are consistent with the results found in our visual experiment, where observers were able to detect chroma differences using these 2 Japanese charts (Fig. 1).
The colour of a soil might also have different Munsell notations using old and new charts from the same manufacturer. Figure 6a shows that 86% of the chips with identical Munsell designation in JAPAN1970 and JAPAN2001 had colour differences [greater than or equal to] 3.0 CIELAB units. In this case, [DELTA][L.sup.*] contributed 56%, [DELTA][C.sup.*.sub.ab] 33%, and [DELTA][H.sup.*.sub.ab] 11% to the whole colour difference, indicating a major lightness and chroma deterioration in the chips from the oldest chart. However, colour differences between old and new American charts (USA1975 and USA2000) is approximately half the value found between Japanese charts (Table 6). Thus, it could be concluded that the charts USA1975 had greater durability than JAPAN1970, or the latter have been exposed to more extreme environmental conditions (cold, hot, sunlight, moisture, etc.).
[FIGURE 6 OMITTED]
Another possible source of variation in the Munsell colour notation of a soil is that chips from USA and JAPAN have not been identically produced. For the entirely new charts USA2000 and JAPAN2001, we measured an average colour difference of 2.12 CIELAB units and 1.52 CIEDE2000 units (Table 6). However, 70% of the pairs had colour differences <2.0 CIELAB units (Fig. 6b). Clearly perceptible colour differences (that is, colour differences >2.0 CIELAB units) appeared in 63 pairs, especially for: 10R 4/8, 3/4 and 3/6, 2.5YR 4/8, 3/3, 3/4 and 3/6, 5YR 4/6, 3/3 and 3/4, 7.5YR 3/3 and 3/4, 10YR 5/8, 4/6, 3/2, 3/3, 3/4, 2/1 and 2/2, 2.5Y 3/2 and 3/3, and 5Y 4/4 and 3/2. Colour differences in all these pairs (the most chromatic ones within the Munsell values 5/, 4/, 3/, and 2/) were [greater than or equal to] 4.0 CIELAB units, 46% of the total difference being due to [DELTA][C.sup.*.sub.ab], 35% to [DELTA][L.sup.*], and 19% to [DELTA][H.sup.*.sub.ab]. As an example, the measured spectral reflectance of the chip 10YR 3/4 is shown in Fig. 7, illustrating the similar shape of all the curves from Japanese charts (although separated by the degree of use) and American charts, while differences between Japanese and American manufacturers are also apparent.
[FIGURE 7 OMITTED]
Clearly, charts from different manufacturers and used in the field, because of production differences as well as different colour fading, can significantly affect soil-colour determination. In the worst case (JAPAN1970 v. USA1975b), the discrepancies between one pair of chips with the same Munsell designation reached an average value of 3.72 CIEDE2000 units (Table 6), which can be considered a large colour difference clearly perceptible to the human eye.
We have shown that colours from Revised Standard Soil Colour Charts and Munsell Soil Colour Charts from different editions and degrees of use have significant variations that cause discrepancies among visual determinations of soil colours. The chips from some charts yellowed and lost lightness and chroma owing to the original printing characteristics and colour fading, resulting in Munsell colour notations of redder hue, lighter value and more intense chroma. For accurate results, these soil-colour charts should be new and produced by the same manufacturer. On average, the chips from our American and Japanese charts registered a large visual colour difference, 1.52 CIEDE2000 units when they are new and 3.72 CIEDE2000 units when old. Therefore, it would be advisable to perform periodical colourimetric analysis for quality control of soil-colour charts, both for testing of the conformity between colours and Munsell notation, and the uniformity of visual spacing among colour chips.
Ministerio de Educacion y Ciencia (Spain), research projects CGL2004-02282/BTE and FIS2004-05537.
Baumgardner MF, Silva LF, Biehl LL, Stoner ER (1985) Reflectance properties of soils. Advances in Agronomy 38, 1-43.
Berns RS (2000) 'Billmeyer and Saltzman's principles of color technology.' 3rd edn (John Wiley & Sons: New York)
Blackburn G (1973) Colour charts for use in agricultural research. Journal of the Australian Institute of Agricultural Science 39, 141.
Brouwer J, Fitzpatrick RW (2002) Restricting layers, flow paths, and correlation between duration of soil saturation and soil morphological features along a hillslope with an altered soil water regime in western Victoria. Australian Journal of Soil Research 40, 927-946. doi: 10.1071/SR02009
Brown DJ, Clayton MK, McSweeney K (2004) Potential terrain controls on soil color, texture contrast and grain-size deposition for the original catena landscape in Uganda. Geoderma 122, 51-72. doi: 10.1016/j.geoderma.2003.12.004
CIE Commision Internationale de l'Eclairage (1986) 'Colorimetry.' Publication 15.2, 2nd edn (CIE Central Bureau: Vienna)
CIE Commision Internationale de l'Eclairage (1995) 'Industrial colour-difference evaluation.' Publication 116. (CIE Central Bureau: Vienna)
CIE Commision Internationale de l'Eclairage (2001) 'Improvement to industrial colour-difference evaluation.' CIE Technical Report, Publication 142. (CIE Central Bureau: Vienna)
Coventry RJ, Robinson PJ (1981) Colour groups derived from numerical classification of the Munsell soil colour chips. Australian Journal of Soil Research 19, 103-112. doi: 10.1071/SR9810103
Cui G, Luo MR, Rigg B, Roesler G, Witt K (2002) Uniform colour spaces based on the DIN99 colour-difference formula. Color Research and Application 27, 282-290. doi: 10.1002/col.10066
Dahlgren RA, Boettinger JL, Huntington GL, Amundson RG (1997) Soil development along an elevational transect in the western Sierra-Nevada, California. Geoderma 78, 207-236. doi: 10.1016/S00167061(97)00034-7
FAO (1998) 'World reference base for soil resources.' World Soil Resources Reports, No. 84. (FAO: Rome)
Farnsworth D (1943) 'Fransworth-Munsell 100-Hue Test.' (Color Co. Inc.: Baltimore, MD)
Fernandez RN, Schulze DG (1987) Calculation of soil color from reflectance spectra. Soil Science Society of America Journal 51, 1277-1282.
Fujihira Industry Company (1970) 'Revised standard soil color charts.' (Fujihira Industry Co.: Tokyo)
Fujihira Industry Company (1987) 'Revised standard soil color charts.' (Fujihira Industry Co.: Tokyo)
Fujihira Industry Company (2001) 'Revised standard soil color charts.' (Fujihira Industry Co.: Tokyo)
Ishihara S (1979) 'Test for color blindness.' 38th edn (Kanehara Shuppan Co.: Tokyo)
Kollmorgen (1991) 'CIE [left and right arrow] Munsell conversion program v. 17.20.' (Macbeth Division of Kollmorgen Instruments Co.: New Windsor, NY)
Landa ER (2004) Albert H. Munsell: A sense of color at the interface of art and science. Soil Science 169, 83-89. doi: 10.1097/01.ss.0000117789.98510.30
Lee DY, Honson M (2003) Chromatic variation of Ishihara diagnostic plates. Color Research and Application 28, 267-276. doi: 10.1002/col.10161
Luo MR, Cui G, Rigg B (2001) The development of the CIE 2000 colour-difference formula: CIEDE2000. Color Research and Application 26, 340-350. doi: 10.1002/col.1049
Melgosa M, Hita E, Poza A, Alman DH, Berns RS (1997) Suprathreshold color-difference ellipsoids for surface colors. Color Research and Application 22, 148-155. doi: 10.1002/(SICI)1520-6378(199706)22:3 <148::AID-COL3>3.0.CO;2-R
Melville MD, Atkinson G (1985) Soil colour: Its measurement and its designation in models of uniform colour space. Journal of Soil Science 36, 495-512.
Munsell Colour Company (1975) 'Munsell soil color charts.' (Munsell Color Co.: Baltimore, MD)
Munsell Colour Company (2000) 'Munsell soil color charts.' (Munsell Color Co.: Baltimore, MD)
Oyama M, Takehara H (2001) Explanation. In 'Revised standard soil color charts', pp. 3-14. (Fujihira Industry Co.:Tokyo)
Paclt J (1983) A chronology of color charts and color terminology for naturalists. Taxon 32, 393-405.
Sanchez-Maranon M, Delgado G, Delgado R, Perez MM, Melgosa M (1995) Spectroradiometric and visual color measurements of disturbed and undisturbed soil samples. Soil Science 160, 291-303.
Sanchez-Maranon M, Soriano M, Melgosa M, Delgado G, Delgado R (2004) Quantifying the effects of aggregation, particle size and components on the colour of Mediterranean soils. European Journal of Soil Science 55, 551-565. doi: 10.1111/j.1365-2389.2004.00624.x
Schabenberger O, Pierce FJ (2002) 'Contemporary statistical models for plant and soil sciences.' (CRC Press: Washington, DC)
Seve R (1991) New formula for the computation of CIE 1976 Hue Difference. Color Research and Application 16, 217-218.
Shields JA, Arnaud RJS, Paul EA, Clayton JS (1966) Measurement of soil color. Canadian Journal of Soil Science 46, 83-90.
Simonson RW (1993) Soil color standards and terms for field use--history of their development. In 'Soil color'. Special Publication No. 31. (Eds JM Bigham, EJ Ciolkosz) pp. 1-20. (Soil Science Society of America: Madison, WI)
Soil Survey Staff (1993) 'Soil survey manual.' Agricultural Handbook 18. (U.S. Government Printing Office: Washington, DC)
Soil Survey Staff (1999) 'Soil Taxonomy. A basic system of soil classification for making and interpreting soil surveys.' Agricultural Handbook 436, 2nd edn (U.S. Government Printing Office: Washington, DC)
Torrent J, Barron V (1993) Laboratory measurement of soil color: Theory and practice. In 'Soil color'. Special Publication No. 31. (Eds JM Bigham, EJ Ciolkosz) pp. 21-33. (Soil Science Society of America: Madison, WI)
Wyszecki G, Stiles WS (1982) 'Color science: Concepts and methods, quantitative data and formulae.' 2nd edn (John Wiley & Sons: New York)
Manuscript received 22 November 2004, accepted 3 May 2005
M. Sanchez-Maranon (A,C), R. Huertas (B), and M. Melgosa (B)
(A) Departamento de Edafologia y Quimica Agricola, Facultad de Ciencias, Universidad de Granada, 18071 Granada, Spain.
(B) Departamento de Optica, Facultad de Ciencias, Universidad de Granada, 18071 Granada, Spain.
(C) Corresponding author. Email: email@example.com
Table 1. Analysis of variance for CIELAB hue-angle ([h.sub.ab]) of the colour chips with a constant Munsell-hue notation by different soil-colour charts Within columns, means followed by the same letters are not significantly different (P > 0.05); for example, in the column 7.5YR the mean value 61.9 does not differ from 63.0 or 60.7, but there is a statistical difference with 59.3. The number of chips in each group is indicated by n Charts 7.5R 10R 2.5YR 5YR JAPAN1970 [h.sub.ab] 35.2a 39.6a 48.1a 57.2a n 21 28 34 38 JAPAN1987 [h.sub.ab] 33.1a 39.3a 46.4a 56.2a n 21 28 34 38 JAPAN2001 [h.sub.ab] 25.0b 33.9b 42.5b 52.1b n 21 28 34 38 USA1975a [h.sub.ab] -- 34.7b 44.7b 51.5b n 25 17 33 USA1975b [h.sub.ab] -- 34.6b 44.7b 52.6b n 25 17 33 USA2000 [h.sub.ab] -- 38.1a 44.7b 52.3b n 35 37 33 F-ratio 22.8 7.5 5.3 8.7 Charts 7.5YR 10YR 2.5Y 5Y JAPAN1970 [h.sub.ab] 63.0a 71.2a 77.8a 85.1a n 37 37 32 31 JAPAN1987 [h.sub.ab] 61.9ab 71.2a 78.0a 85.0a n 37 37 32 31 JAPAN2001 [h.sub.ab] 60.7b 69.6b 78.9a 86.9b n 37 37 32 31 USA1975a [h.sub.ab] 59.3c 67.2c 77.0b 83.4a n 20 36 18 31 USA1975b [h.sub.ab] 60.6b 69.3b 77.9a 84.1a n 20 36 18 31 USA2000 [h.sub.ab] 59.4c 68.1c 76.5b 84.5a n 35 36 31 31 F-ratio 3.3 7.0 3.3 3.8 Table 2. Analysis of variance for CIELAB lightness (L*) of the colour chips with a constant Munsell-value notation (V) by different soil-colour charts Within columns, means followed by the same letters are not significantly different (P > 0.05); for example, in the column V = 8 the mean value 73.9 is statistically different from any other. The number of chips in each group is indicated by n Charts V=1.7 V=2 V=2.5 V=3 V=4 JAPAN1970 L* 27.5a 30.2a -- 34.9a 41.6a n 6 23 33 43 JAPAN1987 L* 26.5b 29.2b -- 34.4a 41.2a n 6 23 33 43 JAPAN2001 L* 24.3c 26.8c -- 33.2b 41.9a n 6 23 33 43 USA1975a L* -- 21.2d 28.7a 33.0bc 41.6a n 2 8 22 29 USA1975b L* -- 18.7d 28.1a 32.1c 41.2a n 2 8 22 29 USA2000 L* -- 20.3d 25.7b 30.2d 40.2b n 2 14 28 35 F-ratio 50.8 62.5 13.1 23.8 6.7 V=5 V=6 V=7 V=8 JAPAN1970 L* 49.6a 57.5a 65.8a 73.9a n 43 44 38 28 JAPAN1987 L* 49.4a 57.7a 66.2b 75.4b n 43 44 38 28 JAPAN2001 L* 51.5b 61.3b 70.8c 80.3c n 43 44 38 28 USA1975a L* 51.lbc 60.6c 70.3d 79.7d n 34 36 26 23 USA1975b L* 50.8c 60.7c 70.4d 79.9cd n 34 36 26 23 USA2000 L* 50.9c 60.7c 70.6cd 80.2c n 40 42 42 35 F-ratio 60.5 275.4 350.7 249.2 Table 3. Colour variability within chips of different soil-colour charts, estimated by the average standard deviation (s.d.) of CIELAB parameters measured at 3 different places on each chip Charts s.d. L* s.d. a* s.d. b* JAPAN1970 (258 chips) 0.208 0.114 0.274 JAPAN1987 (258 chips) 0.196 0.120 0.207 JAPAN2001 (258 chips) 0.028 0.016 0.028 USA1975a (180 chips) 0.132 0.076 0.156 USA1975b (180 chips) 0.241 0.118 0.287 USA2000 (238 chips) 0.030 0.017 0.030 Charts s.d. [C*.sub.ab] s.d. [h.sub.ab] JAPAN1970 (258 chips) 0.291 0.385 JAPAN1987 (258 chips) 0.267 0.301 JAPAN2001 (258 chips) 0.028 0.067 USA1975a (180 chips) 0.171 0.153 USA1975b (180 chips) 0.304 0.282 USA2000 (238 chips) 0.031 0.061 Table 4. Mean CIELAB hue-difference [DELTA][H.sub.ab] * (standard deviation in parentheses) for chips on adjacent Munsell-hue charts, for a given chroma and different soil-colour charts, measured in n pairs of chips with the same value and chroma Charts Chroma 1 Chroma 2 Chroma 3 JAPAN1970 [DELTA][H.sub.ab]* 0.64 (0.43) 1.23 (0.5l) 1.80 (0.7l) n 51 42 40 JAPAN1987 [DELTA][H.sub.ab]* 0.72 (0.48) 1.38 (0.63) 2.01 (0.80) n 51 42 40 JAPAN2001 [DELTA][H.sub.ab]* 0.88 (0.27) 1.68 (0.33) 2.40 (0.48) n 51 42 40 USA1975a [DELTA][H.sub.ab]* 0.96 (0.26) 1.78 (0.82) 2.54 (0.48) n 17 34 15 USA1975b [DELTA][H.sub.ab]* 0.95 (0.28) 1.82 (0.68) 2.56 (0.46) n 17 34 15 USA2000 [DELTA][H.sub.ab]* 0.93 (0.33) 1.68 (0.4l) 2.44 (0.56) n 40 39 35 Charts Chroma 4 Chroma 6 Chroma 8 JAPAN1970 [DELTA][H.sub.ab]* 2.62 (0.71) 3.71 (1.00) 5.07 (1.43) n 34 29 21 JAPAN1987 [DELTA][H.sub.ab]* 2.89 (0.87) 3.97 (1.25) 5.56 (1.56) n 34 29 21 JAPAN2001 [DELTA][H.sub.ab]* 3.15 (0.67) 4.51 (0.62) 5.72 (0.77) n 34 29 21 USA1975a [DELTA][H.sub.ab]* 3.40 (0.80) 4.71 (0.75) 5.74 (1.45) n 30 24 17 USA1975b [DELTA][H.sub.ab]* 3.35 (1.06) 4.71 (0.84) 5.75 (1.19) n 30 24 17 USA2000 [DELTA][H.sub.ab]* 3.05 (0.78) 4.42 (0.78) 5.72 (0.79) n 34 26 19 Table 5. Mean CIELAB lightness ([DELTA]L*) and chroma ([DELTA][C*.sub.ab]) differences (standard deviation in parentheses) for one step of Munsell value and chroma, respectively, measured in n pairs of neighbouring chips on different Munsell-hue pages and soil-colour charts Charts 7.5R 10R JAPAN1970 [DELTA]L* 7.45 (2.30) 7.19 (1.55) n 15 22 JAPAN1970 [DELTA][C*.sub.ab] 4.28 (0.83) 4.58 (1.30) n 14 21 JAPAN1987 [DELTA]L* 7.87 (1.83) 7.46 (1.57) n 15 22 JAPAN1987 [DELTA][C*.sub.ab] 4.81 (1.15) 4.82 (1.23) n 14 21 JAPAN2001 [DELTA]L* 8.64 (2.79) 8.95 (1.68) n 15 22 JAPAN2001 [DELTA][C*.sub.ab] 4.60 (1.06) 4.78 (0.97) n 14 21 USA1975a [DELTA]L* -- 9.51 (1.10) n 19 USA1975a [DELTA][C*.sub.ab] -- 4.89 (1.29) n 20 USA1975b [DELTA]L* -- 9.70 (1.37) n 19 USA1975b [DELTA][C*.sub.ab] -- 4.81 (1.18) n 20 USA2000 [DELTA]L* -- 9.93 (0.93) n 29 USA2000 [DELTA][C*.sub.ab] -- 5.27 (1.10) n 28 Charts 2.5YR 5YR JAPAN1970 [DELTA]L* 7.07 (2.30) 7.02 (2.06) n 28 32 JAPAN1970 [DELTA][C*.sub.ab] 4.30 (1.09) 4.11 (1.47) n 27 30 JAPAN1987 [DELTA]L* 7.18 (2.11) 7.32 (1.86) n 28 32 JAPAN1987 [DELTA][C*.sub.ab] 4.67 (1.16) 4.58 (1.51) n 27 30 JAPAN2001 [DELTA]L* 8.82 (1.77) 8.90 (1.76) n 28 32 JAPAN2001 [DELTA][C*.sub.ab] 4.90 (1.17) 5.09 (1.12) n 27 30 USA1975a [DELTA]L* 8.35 (1.93) 9.13 (1.37) n 13 27 USA1975a [DELTA][C*.sub.ab] 4.90 (1.21) 5.43 n 12 26 USA1975b [DELTA]L* 8.59 (2.55) 9.21 (1.66) n 13 27 USA1975b [DELTA][C*.sub.ab] 4.63 (1.63) 5.37 (1.07) n 12 26 USA2000 [DELTA]L* 10.15 (1.48) 9.83 (0.94) n 31 27 USA2000 [DELTA][C*.sub.ab] 5.46 (1.03) 5.80 (1.05) n 30 26 Charts 7.5YR 10YR JAPAN1970 [DELTA]L* 7.62 (2.49) 7.41 (1.78) n 31 32 JAPAN1970 [DELTA][C*.sub.ab] 4.23 (1.32) 4.94 (1.65) n 29 29 JAPAN1987 [DELTA]L* 8.04 (2.80) 7.88 (1.89) n 31 32 JAPAN1987 [DELTA][C*.sub.ab] 5.04 (1.63) 5.19 (1.59) n 29 29 JAPAN2001 [DELTA]L* 9.20 (1.46) 9.06 (1.35) n 31 32 JAPAN2001 [DELTA][C*.sub.ab] 5.49 (1.25) 5.63 (1.38) n 29 29 USA1975a [DELTA]L* 8.98 (1.50) 9.45 (1.38) n 16 30 USA1975a [DELTA][C*.sub.ab] 5.31 (1.29) 5.80 (1.46) n 14 29 USA1975b [DELTA]L* 9.35 (1.17) 9.28 (1.40) n 16 30 USA1975b [DELTA][C*.sub.ab] 5.80 (1.79) 6.02 (2.56) n 14 29 USA2000 [DELTA]L* 9.88 (0.94) 10.01 (0.76) n 29 30 USA2000 [DELTA][C*.sub.ab] 5.85 (1.03) 6.33 (1.20) n 28 29 Charts 2.5Y 5Y JAPAN1970 [DELTA]L* 7.57 (1.35) 7.73 (1.50) n 26 25 JAPAN1970 [DELTA][C*.sub.ab] 5.10 (1.42) 5.66 (1.51) n 25 24 JAPAN1987 [DELTA]L* 7.97 (1.52) 7.93 (1.53) n 26 25 JAPAN1987 [DELTA][C*.sub.ab] 5.85 (1.71) 6.13 (1.55) n 25 24 JAPAN2001 [DELTA]L* 9.12 (1.57) 9.14 (1.05) n 26 25 JAPAN2001 [DELTA][C*.sub.ab] 6.09 (1.30) 6.33 (1.37) n 25 24 USA1975a [DELTA]L* 9.37 (0.92) 9.41 (1.21) n 14 25 USA1975a [DELTA][C*.sub.ab] 6.3l (0.27) 6.14 (1.02) n 12 24 USA1975b [DELTA]L* 9.45 (0.99) 9.36 (1.11) n 14 25 USA1975b [DELTA][C*.sub.ab] 6.22 (0.98) 6.18 (1.14) n 12 24 USA2000 [DELTA]L* 9.95 (0.67) 9.55 (1.22) n 25 25 USA2000 [DELTA][C*.sub.ab] 5.95 (1.83) 6.33 (1.26) n 24 24 Table 6. Average CIELAB ([DELTA][E*.sub.ab]) and CIEDE2000 ([DELTA][E.sub.00]) colour differences between all pairs of chips with identical Munsell notation from 2 different soil-colour charts Components of CIELAB colour differences ([DELTA]L*, [DELTA][C*.sub.ab], and [DELTA][H*.sub.ab]) are also shown (absolute value) Charts Pairs [DELTA]L* [DELTA][C*.sub.ab] JAPAN1970-JAPAN1987 258 0.80 1.30 JAPAN1970-JAPAN2001 258 3.13 2.40 JAPAN1970-USA1975a 171 2.90 3.04 JAPAN1970-USA1975b 171 3.29 3.33 JAPAN1970-USA2000 210 3.21 2.95 JAPAN1987-JAPAN2001 258 3.01 1.63 JAPAN1987-USA1975a 171 2.52 1.97 JAPAN1987-USA1975b 171 2.85 2.32 JAPAN1987-USA2000 210 2.84 1.97 JAPAN2001-USA1975a 171 0.84 1.17 JAPAN2001-USA1975b 171 1.22 1.53 JAPAN2001-USA2000 210 1.07 1.33 USA1975a-USA1975b 180 0.62 0.95 USA1975a-USA2000 180 1.18 1.56 USA1975b-USA2000 180 1.01 1.70 [DELTA] [DELTA] [DELTA] Charts [H*.sub.ab] [E*.sub.ab] [E.sub.00] JAPAN1970-JAPAN1987 0.35 1.74 1.15 JAPAN1970-JAPAN2001 0.97 4.56 3.34 JAPAN1970-USA1975a 1.06 4.99 3.47 JAPAN1970-USA1975b 1.01 5.46 3.72 JAPAN1970-USA2000 0.97 5.10 3.58 JAPAN1987-JAPAN2001 0.81 3.89 2.97 JAPAN1987-USA1975a 1.01 3.83 2.83 JAPAN1987-USA1975b 0.92 4.31 3.08 JAPAN1987-USA2000 0.85 4.06 2.97 JAPAN2001-USA1975a 0.99 2.00 1.54 JAPAN2001-USA1975b 0.78 2.40 1.73 JAPAN2001-USA2000 0.73 2.12 1.52 USA1975a-USA1975b 0.46 1.38 0.94 USA1975a-USA2000 0.63 2.33 1.60 USA1975b-USA2000 0.68 2.31 1.53
|Printer friendly Cite/link Email Feedback|
|Author:||Sanchez-Maranon, M.; Huertas, R.; Melgosa, M.|
|Publication:||Australian Journal of Soil Research|
|Date:||Dec 1, 2005|
|Previous Article:||Agricultural soils irrigated with acidic mine water: acidity, heavy metals, and crop contamination.|
|Next Article:||Lateritic soils of Kerala, India: their mineralogy, genesis, and taxonomy.|