Page 1
Plant and Soil 191: 77–87, 1997.
77
c
1997 Kluwer Academic Publishers. Printed in the Netherlands.
The capacity of soils to preserve organic C and N by their association with
clay and silt particles
Jan Hassink
DLO Research Institute for Agrobiology and Soil Fertility (AB-DLO), P.O. Box 129, 9750 AC Haren, The
Netherlands
Received 16 August 1996. Accepted in revised form 27 February 1997
Key words: organic C and N, protection capacity, soil texture
Abstract
Although it has been recognized that the adsorption of organics to clay and silt particles is an important determinant
of the stability of organic matter in soils, no attempts have been made to quantify the amounts of C and N that can
be preserved in this way in different soils. Our hypothesis is that the amounts of C and N that can be associated
with clay and silt particles is limited. This study quantifies the relationships between soil texture and the maximum
amounts of C and N that can be preserved in the soil by their association with clay and silt particles. To estimate
the maximum amounts of C and N that can be associated with clay and silt particles we compared the amounts
of clay- and silt-associated C and N in Dutch grassland soils with corresponding Dutch arable soils. Secondly, we
compared the amounts of clay- and silt-associated C and N in the Dutch soils with clay and silt-associated C and
N in uncultivated soils of temperate and tropical regions.
We observed that although the Dutch arable soils contained less C and N than the corresponding grassland soils,
the amounts of C and N associated with clay and silt particles was the same indicating that the amounts of C and N
that can become associated with this fraction had reached a maximum. We also observed close positive relationships
between the proportion of primary particles < 20
m in a soil and the amounts of C and N that were associated with
this fraction in the top 10 cm of soils from both temperate and tropical regions. The observed relationships were
assumed to estimate the capacity of a soil to preserve C and N by their association with clay and silt particles. The
observed relationships did not seem to be affected by the dominant type of clay mineral. The only exception were
Australian soils, which had on average more than two times lower amounts of C and N associated with clay and
silt particles than other soils. This was probably due to the combination of low precipitation and high temperature
leading to low inputs of organic C and N.
The amount of C and N in the fraction > 20
m was not correlated with soil texture. Cultivation decreased the
amount of C and N in the fraction > 20
m to a greater extent than in the fraction < 20
m, indicating that C and N
associated with the fraction < 20
m is better protected against decomposition.
The finding of a given soil having a maximum capacity to preserve organic C and N will improve our estimations
of the amounts of C and N that can become stabilized in soils. It has important consequences for the contribution
of different soils to serve as a sink or source for C and N in the long term.
Introduction
Fine-textured soils have higher organic C and N con-
tents than coarse-textured soils when supplied with
similar input of organic material (Amato and Ladd,
1992; Hassink, 1994; Hassink et al., 1997; Jenkin-
FAX No: + 31505337291. E-mail: J. Hassink@ab.dlo.nl
son, 1988) The difference is assumed to result from
the greater physical protection of soil organic mat-
ter in fine-textured soils (Jenkinson, 1988; Van Veen
and Kuikman, 1990). Considerable published evidence
indicates that one of the principal factors responsible
for physical protection of organic matter in soils is its
ability to associate with clay and silt particles (Theng,
PIPS NO.:136711 (M) BIO2KAP
*136711*
plso6537.tex; 17/06/1997; 18:52; v.7; p.1
Page 2
78
1979). Several studies showed that the amount of C
and N associated with clay and silt particles is mainly
affected by soil texture and not by the input of organics
to the soil, while organic matter in larger size fractions
is mainly affected by the input organics and not by
soil texture (Christensen, 1992; Garwood et al., 1972;
Hassink, 1995; Quiroga et al., 1996). Although it has
been recognized that physical protection mechanisms
are important determinants of the stability of organic
matter in soil (Van Veen and Kuikman, 1990) it has not
been determined whether the capacity of soils to phys-
ically protect organic matter is limited and whether
the capacity of soils to preserve organic matter can be
quantified. Not only soil texture but also clay type may
affect the capacity of soils to physically protect organic
C and N. Specific surface areas of clays vary from 6–
39 m
2
g
1
for kaolinite (Dixon, 1977) to 800 m
2
g
1
for smedite and vermiculite (Robert and Chenu, 1992).
Soils dominated by clays with a high specific surface
area are expected to adsorb more humic substances
than soils dominated by clays with a low specific sur-
face area (Tate and Theng, 1980). As the amounts of
soil organic C and N are higher in old grassland and
uncultivated arable soils (Bonde et al., 1992; Feller et
al., 1991; Jenkinson, 1988; Lugo and Brown, 1993;
Stevenson, 1982; Zhang et al., 1988), we assume that
the relationship between soil texture and clay- and silt-
associated C and N in old grassland and uncultivated
soils represents the opacity of soils to preserve organ-
ic C and N. The first objective of this paper is to test
whether there is really a maximum in the amount of
C and N that can become associated with clay and silt
particles. Therefore we compared the amounts of clay
and silt associated C and N and C and N in coarser
size fractions in two grassland with two corresponding
arable soils. The second objective of this paper is to
test whether for soils from both temperate and tropi-
cal regions a general relationship between soil texture,
the dominant type of clay mineral and the capacity to
preserve organic C and N can be found. To test this we
compared clay and silt associated C and N in Dutch
grassland soils with published results of correspond-
ing measurements in uncultivated soils from different
continents.
Materials and methods
We sampled the top 10 cm of grassland soils which had
been under grass for at least 30 years. The grasslands
were grazed by dairy cattle and received 400–500 kg
fertilizer N per ha per year. To reach the first objective
we sampled adjacent arable fields at two locations. In
Tynaarlo the arable field had been under a 4 year rota-
tion of winter wheat, sugarbeet, barley and ware pota-
toes for at least 25 years. In Cranendonck we compared
the distribution of C and N in a 30 years old grassland
field with a field that had been under maize for at least
25 years. We also compared the distribution of C and
N in the top 10 cm with deeper soil layers (30–40 cm
and 60–80 cm). Some characteristics of the grassland
and arable soils are given in Table 1. Information about
the dominant clay minerals in the different soils was
taken from the literature (Breeuwsma, 1990; Favejee,
1949; Marel, 1949). The average precipitation in the
Netherlands is 700–800 mm and the average tempera-
ture 8
C.
Three mixed samples, each consisting of 20 bulked
cores, were taken from every location. The samples
were sieved through a 0.008 m mesh screen; roots
and stubble were removed and the samples were ana-
lyzed separately. Dry soil (50 g) was suspended in
250 mL water for 24 hours. The samples were treated
ultrasonically for 15 minutes with a probe-type ultra-
sound generating unit (Soniprep 150). Probe output
was calibrated by measuring the temperature produced
upon ultrasonifying a known mass of water for a spec-
ified time (North, 1976). Output power a was 30 W.
Heat buildup in the soil suspension was controlled by
a water cooling jacket. The dispersed soil suspension
was transferred to a 1 L glass cylinder. The cylinder
was shaken end over end until the soil was suspended.
A table showing the settling-times for 20
m particles
at temperatures between 15 and 25
C was construct-
ed by applying Stokes’ law and a particle density of
2.675 g cm
3
. After the correct settling time, particles
< 20
m were isolated by siphoning the suspension
at the appropriate depth. The fractions were dried for
24 hours at 105
C. They were ground and analyzed
for total C and N using a Carlo Erba NA 1500 ana-
lyzer. Particle size distribution was determined after
oxidation of organic matter with H
2
O
2
and removal of
CaCO
3
with HCl. The amounts of C and N associated
with the particle size fraction < 20
m were calculated
by using the percentage of particles < 20
m obtained
after removal of organic matter and CaCO
3
and the C
and N contents of the fraction < 20
m obtained after
sonification.
Total C in soil was defined as dichromate-
oxidizable C according to Kurmies (Mebius, 1960).
Total N was determined according to Deys (1961) by
digestion with sulfuric acid and salycylic acid. To com-
plso6537.tex; 17/06/1997; 18:52; v.7; p.2
Page 3
79
Table 1. Some characteristics of the top 10 cm of the Dutch soils that were sampled
Location
C
Cn
pH
Granular composition
C associated
Dominant
(%)
(KCl)
% particles <
with particles
types of
< 20
m
clay
2
m
20
m
50
m
(g kg
1
)
minerals
a
Grasslands
Jubbenga
2.87
21.6
4.6
1.0
1.9
3.7
4.1
Q, K
b
Cranendonck
1.50
19.5
5.4
1.0
3.5
8.5
3.6
Q, K
Heino 1
1.98
18.0
5.0
1.9
2.8
8.7
0.9
Q, K
Holten
2.27
17.5
5.1
2.5
3.6
12.0
1.0
Q, K
Tynaarlo
4.38
17.8
4.4
2.4
4.4
23.5
7.5
Q, K
Achterberg
3.09
17.5
4.9
3.0
5.8
9.0
10.2
Q, K
Markelo
3.91
17.2
5.2
3.4
5.4
12.4
3.0
Q, K
Heino 2
4.68
15.6
5.2
5.6
8.9
29.7
3.6
Q, K
Finsterwolde
5.37
10.9
5.0
8.4
13.3
23.2
11.1
I, S
c
Lelystad
3.07
11.1
7.1
21.6
35.6
56.4
18.6
I, S
Burum
5.37
9.8
4.8
24.1
36.5
71.7
27.5
I, S
Aduard
3.55
10.1
5.6
29.8
45.5
69.4
13.5
I, S
Zaltbommel 1
3.99
9.7
5.8
25.8
42.6
67.2
19.9
I, S
Zaltbommel 2
6.07
9.2
5.4
51.1
76.0
86.4
31.0
I, S
Arable land
Cranendonck
0.93
19.4
5.4
1.1
3.5
8.6
3.6
O, K
Tynaarlo
2.38
17.5
4.6
3.9
6.2
27.8
7.4
Q, K
a
Favejee (1949), Marel (1949), Breeuwsma (1990).
b
Q – quartz, K – kaolinite.
c
I – illte, S – smectite.
pare these C and N analysis methods with the Carlo
Erba analyzer (used to determine C and N in the <
20
m fraction), we also determined total C and N in
the grassland and arable soils from Tynaarlo and Cra-
nendonck using the Carlo Erba NA 1500 analyzer. We
found that differences in total C and N between the
methods were less than 10%.
The amounts of C and N not associated with clay
and silt particles was defined as total soil C and N minus
C and N in the fraction < 20
m. For the grassland and
arable soils from Tynaarlo and Cranendonck we also
determined the amounts of C and N in the fraction >
20
m, by washing the sonicated suspension on a 20
m sieve; the material accumulating on this sieve was
considered to be the fraction > 20
m. C and N in this
fraction was determined using a Carlo Erba NA 1500
analyzer. For the Tynaarlo and Cranendonck soils, the
difference between the amounts of C and N in the >
20
m fraction determined by the Carlo Erba analyzer
and the amounts of C and N in this fraction defined as
the difference between total soil C and N and C and N
in the < 20
m fraction was less than 20% (Table 3).
To reach the second objective we compared total
soil C and N and the distribution of C and N over
the fractions < 20
m and > 20
m of the grassland
soils and arable soils with published results of corre-
sponding measurements in uncultivated soils and in
corresponding soils after different periods of conver-
sion to cultivated arable land. We chose 20
m as the
upper size limit for the clay and silt fraction (instead of
the previously used 50
m; Hassink, 1994), because
in most studies 20
m was taken as the upper size lim-
it. We found data of the top layer (approximately 10
cm) of the following uncultivated soils and cultivated
grasslands: a grassland soil in Germany (Leinweber
and Reuter, 1992), prairie soils in Canada (Elustondo
et al., 1990) and North America (Tiessen and Stew-
art, 1983; Balesdent et al., 1988; Zhang et al., 1988);
savanna and forest soils in Africa (Nigeria, Senegal,
Togo and Ivory Coast; Bates, 1960; Feller et al., 1991),
forest soils in Middle and South America (Guadeloupe
and Brazil; Feller et al., 1991; Bonde et al., 1992) and
virgin soils in Australia (Turchenek and Oades, 1979;
Dalal and Mayer, 1986, 1987). In most of these studies,
plso6537.tex; 17/06/1997; 18:52; v.7; p.3
Page 4
80
Table 2. Percentage of soil particles < 20
m, actual, calculated maximum (max.) amount of C associated
with the particles < 20
m (C in fraction < 20
m = 4.09 + 0.37
% particles < 20
m; Figure 3) and
their difference (maximum – actual amount), the dominant types of clay minerals in the uncultivated soils of
temperate and tropical regions referred to in this study, and the mean annual precipitation and temperature of
the locations
Particles
Amount of C associated
Dominant types
Mean annual
Mean annual
< 20
m
with particles < 20
m
of clay minerals
a
precip.
temp.
(%)
Actual
Max
Difference
(mm)
(
C)
(g kg
1
)
Australia: Dalal and Mayer (1986 a, b)
86.8
13.4
36.2
–22.8
S, K
670
18.5
66.2
15.4
28.6
–13.2
Q, R
630
19.5
55.8
12.6
24.7
–12.1
R, I
670
18.5
52.1
11.1
23.4
–12.3
Q, R
610
19.9
82.3
6.0
34.5
–28.5
K, I
480
20.5
27.4
9.5
14.2
–4.7
K, Q
580
20.3
Australia: Turchenek and Oades (1979)
50.7
19.7
22.8
–3.1
K, I
530
b
11.7
b
Senegal: Feller et al. (1991 a, b)
18
6.0
10.8
–4.8
K, Q
800
29
Togo: Feller et al. (1991 a, b)
18
5.5
10.8
–5.3
NA
c
1040
27
Ivory Coast: Feller et al. (1991 a, b)
35
12.7
7.0
–4.3
K, Q
1360
28
Nigeria: Bates, 1960
13.4
7.0
9.0
–2.0
K
1230
27
Guadeloupe: Feller et al. (1991 a, b)
70
30.0
3.0
0.0
K, H
3000
25
Brazil: Feller et al. (1991 a, b)
70
26.1
30.0
–3.9
K, Go
1200
21
Brazil: Bonde et al. (1992)
62.2
31.8
27.1
+4.7
K
1000–2000
b
16–20
b
Missouri, US: Balesdent et al. (1988)
78
d
30.5
33.0
–2.5
M
800–1400
b
NA
Iowa, US: Zhang et al. (1988)
55
28.8
24.4
+4.4
S
833
8
58
31.5
25.6
+6.0
S
833
8
Saskatchewan, Canada: Tiessen and Stewart (1983)
41
d
26.3
19.3
+7.0
S
d
420
f
1
f
29
d
16.7
14.8
+1.9
S
420
1
74
d
23.4
31.5
–8.1
S
420
1
Quebec, Canada: Elustondo et al. (1990)
61.2
23.6
26.7
–3.1
1 + Cl, Q
g
800–1000
b
4
h
43.7
20.9
20.3
+0.6
1 + Cl, Q
800–1000
4
15.4
15.3
9.8
+5.5
Mi
e
800–1000
4
plso6537.tex; 17/06/1997; 18:52; v.7; p.4
Page 5
81
Table 2. Continued
65.8
37.2
28.4
+8.7
Mi
800–1000
4
63.5
24.2
27.6
+3.4
I + Cl, Q
800–1000
4
16.7
15.7
10.3
+5.4
Mi
800–1000
4
34.8
18.6
17.0
+1.6
Mi
800–1000
4
Germany: Leinweber and Reuter (1992)
37.6
17.9
18.0
–0.1
C, I
500–600
b
8
b
a
S – smectite, K – kaolinite, Q – quartz, R – randomly interstratified, I –illite, H – hematite,
Go – goethite, M –montmorillonite, Mi – mica, CI – chlorite, C – calcite.
b
Estimated from The World Atlas of Agriculture. 1970. Committee for the World Atlas of
Agricuiture, Instituto Geografico de Agostini, Novara, Italy.
c
Not available.
d
Interpolated from the data.
e
Kodama (1979).
f
Campbell et al. (1993).
g
De Kimpe et al. (1979).
h
De Kimpe et al. (1974).
Table 3. Amounts of C and N in the size fractions < 20
m and > 20
m and total
soil C and N in the top cm of the grassland and arable soil in Tynaarlo and the top 10
cm and the soil layers at 30–40 and 60–80 cm depth in the grassland and maize field
in Cranendonck analyzed by a Carlo Erba analyzer (g kg
1
)
Location
C in fraction
Total C
N in fraction
Total N
Treatment
< 20
m > 20
m
< 20
m > 20
m
Tynaarlo
Grassland
7.5
34.2
43.8
0.74
1.64
2.46
Arable
7.4
14.4
23.8
0.72
0.56
1.36
Cranendonck
0–10 cm grass
3.6
12.2
15.0
0.32
0.49
0.77
30–40 cm grass
3.4
7.5
12.1
0.31
0.31
0.62
60–80 cm grass
3.4
6.4
9.7
0.31
0.26
0.55
0–10 cm maize
3.6
5.9
9.3
0.33
0.17
0.48
30–40 cm maize
3.5
3.3
7.1
0.31
0.09
0.36
60–80 cm maize
3.1
2.6
7.2
0.29
0.07
0.37
total soil C and N and the amounts associated with the
particle size fraction < 20
m were determined both
in grasslands and uncultivated soils and in adjacent
soils with similar characteristics that had been culti-
vated and used as arable land for different periods of
time (5–120 years). In the soils from Central and South
America, only C was determined. The dominant clay
type and the climatic conditions were given in most
studies. When these data were not presented in the
paper, they were derived from other sources (Table 2).
Statistical analysis
The relationships between the percentage of soil par-
ticles < 20
m and total soil C and N and C and N
associated with the particle size fractions < 20
m and
> 20
m were analyzed with correlation and regression
techniques (Genstat, 1987).
plso6537.tex; 17/06/1997; 18:52; v.7; p.5
Page 6
82
Results
C end N in size fractions in soils with similar textures
but differing in organic matter input (objective 1)
The first objective was to test whether the amounts of C
and N that can become associated with clay and silt par-
ticles are limited. Despite the fact that the arable field
from Tynaarlo, the maize field from Cranendonck and
the deeper layers of the grassland arid maize field from
Cranendonck had much lower total amounts of soil C
and N than the corresponding top layers of the grass-
land fields, the amounts of C and N associated with
the clay and silt fraction were not less (Table 3). The
amounts of C and N in the fraction > 20
m were less in
the arable field from Tynaarlo arid the maize field from
Cranendonck than in the corresponding grassland sites,
and decreased with increasing depth at Cranendonck
(Table 3).
Relationship between soil texture and total amounts
of soil C and N in uncultivated and grassland soils
from different continents
For uncultivated and grassland soils the total amounts
of C and N in the top 10 cm, ranged from 8 to 60 and
0.5 to 6.5 gram kg
1
soil, respectively. There was no
clear relationship between the C and N content of a soil
and its clay and silt content (Figures 1 and 2). C and N
contents varied considerably between soils with simi-
lar clay and silt content and were generally higher in
the Dutch grassland soils than in the uncultivated soils
of North, Middle and South America and Africa. Aus-
tralian soils generally had the lowest C and N contents
(Figures 1 and 2).
Relationship between soil texture and the amounts of
soil C and N associated with clay and silt particles in
uncultivated and grassland soils from different
continents (objective 2)
In contrast to total C and N, there were highly sig-
nificant positive correlations between the clay and silt
content of a soil and the amounts of C and N associ-
ated with this fraction (Figures 3 and 4; C (g kg
1
)
= 4.09 + 0.37
% particles < 20
m; N (g kg
1
) =
0.40 + 0.037
% particles 20
m; excluding the soils
from Australia). The relationship between the clay and
silt content of a native soil and the amount of C and
N associated with this fraction held for soils from the
Netherlands, North, Middle and South America and
Africa. The amount of C associated with the clay and
soil fraction was less than 10 g kg
1
soil in coarse-
textured soils and up to 37 g kg
1
soil in fine-textured
soils. For N, the corresponding amounts were 10 times
lower. The dominant type of clay mineral present in
the soil did not affect the relationship between the clay
and silt content of the soils and the amounts of C and
N associated with this fraction (Table 1, Figures 3,
4). The Australian soils showed different results; low-
er amounts of C and N were associated with the clay
and silt fraction than in other uncultivated soils and no
correlation between their clay and silt content and the
associated amounts of C and N was found (Figures 3
and 4). The Australian soils were not different from the
other soils with respect to dominant clay type. Com-
pared to other locations, the Australian sites had a low
precipitation in combination with a high temperature
(Table 2).
Relationship between soil texture and the amounts of
soil C and N in the fraction > 20
m in uncultivated
and grassland soils from different continents
The amounts of C and N in the fraction > 20
m
did not correlate with the clay and sift fraction and
varied considerably between soils with similar clay
and silt content (Figures 5 and 6). The amounts of C
and N in the fraction >
m were generally higher in
the grassland soils of the Netherlands (20–45 and 1–
3.7 g kg
1
soil for C and N, respectively) than in the
other soils (3–20 and 0.2–2.0 g kg
1
soil for C and
N, respectively) and lowest in the soils from Australia
(Figures 5 and 6).
Effect of converting uncultivated and grassland soils
to cultivated arable land on the amounts of C and N
in the fractions < 20
m and > 20
m in soils from
different continents
The amounts of soil organic C and N were compared
in pairs of soils with similar texture that had uncul-
tivated or under grassland, or converted to cultivated
arable land for different periods of time (5–120 years)
in different experiments. Conversion to arable land
generally led to a decrease in the amounts of soil C
and N and the relative decreases in soil C and N were
comparable.
In more than 90% of the soils, the relative decrease
in the amount of C or N associated with the fraction
< 20
m was smaller than the relative decrease in the
amount of C or N in the fraction > 20
m (Figure 7).
plso6537.tex; 17/06/1997; 18:52; v.7; p.6
Page 7
83
Figure 1. Relationship between total soil C (g kg
1
soil) and the percentage of soil particles < 20
m in uncultivated and grassland soils of
temperate and tropical regions.
Figure 2. Relationship between total soil N (g kg
1
soil) and the percentage of soil particles < 20
m in uncultivated and grassland soils of
temperate and tropical regions.
Figure 3. Relationship between C in the particle size fraction < 20
m (clay and silt in g kg
1
soil) and the percentage of soil particles < 20
m in uncultivated and grassland soils of temperate and tropical regions. C in fraction < 20
m (g kg
1
soil, excluding Australian soils) = 4.09
(1,59) + 0.37 (0.04)
% particles < 20
m. (r = 0.89; 0 = SE).
plso6537.tex; 17/06/1997; 18:52; v.7; p.7
Page 8
84
Figure 4. Relationship between N in the particle size fraction < 20
m (clay and silt in g kg
1
soil) and the percentage of soil particles < 20
j
mum in uncultivated and grassland soils of temperate and tropical regions. N in fraction < 20
m (g kg
1
soil, excluding Australian soils) =
0.40 (0.15) + 0.037 (0.004)
% particles < 20
m. (r = 0.90; 0 = SE).
Figure 5. Relationship between C in the particle size fraction > 20
m (g kg
1
soil) and the percentage of particles < 20
m in uncultivated
and grassland soils of temperate and tropical regions.
Figure 6. Relationship between N in the particle size fraction > 20
m (g kg
1
soil) and the percentage of soil particles < 20
m in uncultivated
and grassland soils of temperate and tropical regions.
plso6537.tex; 17/06/1997; 18:52; v.7; p.8
Page 9
85
Figure 7. Relationship between the percentage of soil particles < 20
m and the percentage of soil C or N in the particle size fractions < 20
m
and > 20
m after conversion of grasslands or uncultivated soils to cultivated arable land.
Cultivated arable soils generally contained less than
50% of the amount of C or N in the fraction > 20
m
of uncultivated and grassland soils; for the fraction <
20
m the proportion of C or N left after conversion to
arable land was generally more than 60%.
Discussion
The first objective was to test whether the amount of C
that can become associated with clay and sift particles
is limited. We found that although the top layers of the
sandy grassland soils in Tynaarlo and Cranendonck
contained more C than the corresponding arable sites
and the deeper soil layers, the amounts of C associated
with the clay and silt fraction were not different. The
increases in soil C were only observed in the larger
size fractions. This suggests that the amount of C that
can become associated with the clay and silt fraction
had reached a maximum in these sandy soils.
The second objective was to quantify a general rela-
tionship for soils from different continents between
soil texture and the capacity of a soil for C and N to
be associated with clay and silt particles. Therefore,
we determined clay- and silt-associated C and N in
old grassland soils and compared that with clay- and
sift-associated C and N in uncultivated soils. Grass-
lands and uncultivated soils generally have higher C
and N contents than cultivated arable soils, due to the
higher incorporation of organic C, the absence of soil
tillage and reduced exposure to erosion (Christensen,
1992; Elliott et al., 1993; Jenkinson, 1988; Lugo and
Brown, 1993). We assumed that in all old grassland
and uncultivated soils the amounts of C and N were
at equilibrium and had reached a maximum under the
given conditions.
We found close correlations between the amounts
of C and N associated with the clay and silt fraction
and the percentage of soil particles in this size fraction.
Although Dutch grassland soils contained relatively
high amounts of total soil C and N and C and N in the
faction > 20
m, C and N associated with the frac-
tion < 20
m was similar as for the uncultivated soils.
This also suggests that the amounts of C and N that
can become associated with the clay and silt fraction
reached a maximum in all soils. The observation that
the C content of clay particles was similar for grassland
and arable soils, in spite of their differences in total C
content (Hassink et al., 1997) is also in line with this
assumption. Apparently the input of organic C into the
soil was high enough to saturate the clay particles under
arable farming. The extra input of C under grassland
management could not be bound to clay particles and
accumulated in fractions with a greater particle size.
The observations are in agreement with the results
of laboratory studies with pure clays, where it has also
been found that the amount of organics that can be
bound to clay particles is limited (Harter and Stotzky,
1971; Marshman and Marshall, 1981; Pinck et al.,
1954). In those studies, the amount of organics that
could be bound per amount of dispersed, pure clay was
generally higher than the amount of organics bound per
amount of clay and silt in our study. This is logical, as
the clustering of clay packets into stable microaggre-
gates in the soil leads to a reduction in the specific
surface area of the clays (Stotzky, 1986).
It may be expected that the amount of C and N
that can be adsorbed by clay minerals is also affected
plso6537.tex; 17/06/1997; 18:52; v.7; p.9
Page 10
86
by clay type. Specific surface areas vary from 2–4 m
2
g
1
for quartz (Wilding et al., 1977), 10–70 m
2
g
1
for kaolinite, chlorite and mica (Huang, 1990), 50–
100 m
2
g
1
for illite to 800 m
2
g
1
for smectite and
vermiculite (Robert and Chenu, 1992) and differences
in specific surface could affect the capacity of clays
to adsorb humic substances (Tate and Theng, 1980).
Most clays with a low specific surface area, however,
have a relatively high ratio of external to internal sur-
face a whereas for clays with a high specific surface
area the opposite is true (Robert and Chenu, 1992).
Many humic compounds do not penetrate the interlayer
space and adsorb only on the external surfaces (Robert
and Chenu, 1992; Tate and Theng, 1980). In a recent
study it was shown that the amount of fulvic adds that
adsorbed to clay minerals was not related to the sur-
face area of the clays (Varadachari et al., 1994). Also
in the present study we observed no relation between
the dominant clay type in the soil and the amount of
organic C and N that was associated with the clay and
silt fraction. This suggests that the surface area of clays
determined in the laboratory is not a good indicator of
the capacity of clays to adsorb organic C and N.
The amounts of C and N that can be associated by
clay and sift particles also was not different for soils
from temperate and tropical regions. In uncultivated
systems in the tropics, the rate of input can be so high
that the equilibrium level of organic matter is high in
spite of the fact that the decomposition rate is also high
(Greenland et al., 1992). The amounts of total soil C
and N and C and N bound to clay and silt particles were
lower for most Australian soils than for other soils.
The combination of a very low annual precipitation
and high temperatures (Table 2) lead to unfavourable
conditions for plant growth and a limited input of plant
residues into the soil. The input of organic C and N
was obviously too low to reach an equilibrium and
saturation level of the clay and silt particles similar to
those in other uncultivated soils.
In agreement with the hypothesis, there was no
correlation between the amounts of C and N in the
fraction > 20
m and soil texture. The amounts of C
and N in coarser size fractions are related to the amount
of residue that is incorporated into the soil and not to
soil characteristics (Bonde et al., 1992; Hassink, 1995;
Hassink et al., 1997). Due to the wide variation in the
amounts of C and N in the fraction > 20
m, correla-
tions between total soil C and N and the percentage of
particles < 20
m were not clear.
The observation that the decrease in clay- and silt-
associated C and N upon cultivation of soils was gen-
erally less than the decrease in C and N in the particle
size fraction > 20
m confirms that clay and sift parti-
cles protect C against microbial degradation. It is also
in line with the general observation that in arable soils
most of the soil organic matter is found in the clay and
silt size fraction, whereas in forest and grassland soils
the contribution of sand size organic matter to total soil
organic matter is greater (Adams, 1980; Christensen,
1992).
The assumption that the protective capacity of a soil
is limited is expected to have important consequences
for the long term dynamics of soil organic matter and
the role of different soil types as a source and sink term
in the global C and N cycle on the long term.
Acknowledgement
I thank L Brussaand, H G van Faassen, P J Kuikman, J
J Neeteson, P C de Ruiter and K B Zwart for critically
reading and improving the manuscript and for helpful
discussions.
References
Adams T McM 1980 Macro organic matter content of some Northern
Ireland soils. Record Agric. Res. 28, 1–11.
Amato M and Ladd J N 1992 Decomposition of
14
C-labelled glucose
and legume materials in soil: properties influencing the accumu-
lation of organic residue C and microbial biomass C. Soil Biol.
Biochem. 24, 455–464.
Balesdent J, Wagner G H and Mariotti A 1988 Soil organic matter
turnover in long-term field experiments as revealed by carbon-13
natural abundance. Soil Sci. Soc. Am. J. 52, 118–124.
Bates J A R 1960 Studies of a Nigerian forest soil. The distribution
of organic matter in the profile and in various soil fractions. J.
Soil Sci. 11, 246–257.
Bonde T A Christensen B T and Cem C C 1992 Dynamics of soil
organic matter as reflected by natural
13
C abundance in parti-
cle size fractions of forested and cultivated oxisols. Soil Biol.
Biochem. 24, 275–277.
Breeuwsma A 1990 Mineralogical composition of Dutch soils. In
Soil Science of the Netherlands. Eds. W P Locher and H de
Bakker. pp 103–107. Maimberg, Den Bosch.
Campbell C A, Moulin A P, Curtin D, Lafond G P and Townley-
Smith L 1993 Soil aggregation as influenced by cultural practices
in Saskatchewan: I. Black Chemozemic soils. Can. J. Soil Sci.
73, 579–595.
Christensen B T 1992 Physical fractionation of soil and organic
matter in primary particle size and density separates. Adv. Soil
Sci. 20, 1–89.
Dalal R C and Mayer R J 1986 Long-term trends in fertility of soils
under continuous cultivation and cereal cropping in Southern
Queensland. III. Distribution and kinetics of soil organic carbon
in particle-size fractions. Aust. J. Soil Res. 24, 281–292.
plso6537.tex; 17/06/1997; 18:52; v.7; p.10
Page 11
87
Dalal R C and Mayer R J 1987 Long-term trends in fertility of soils
under continuous cultivation and cereal cropping in Southern
Queensland. VI. Loss of total nitrogen from different particle-
size and density fractions. Aust. J. Soil Res. 25, 83–93.
De Kimpe C R, McKeague J A and Topp G C 1974 Soil properties
in relation to water regime at a site near Quebeck city. Can. J.
Soil Sci. 54, 427–446.
De Kimpe C R, Laverdiere M R and Martel Y A 1979 Surface area
and exchange capacity of clay in relation to the mineralogical
composition of gleysolic soils. Can. J. Soil Sci. 59, 341–347.
Deys W B 1961 The determination of total nitrogen in herbage
samples. Inst. Biol. Scheik. Ond. Wageningen Jaarb. Med. 89, 90
(In Dutch).
Dixon J B 1977 Kaolinite and serpentine group minerals. In Minerals
in Soil Environments. Ed. R C Dinauer. pp 357–403. SSSA,
Madison, WI.
Elliott E T, Cambandella C A and Cole C V 1993 Modification of
ecosystem processes by management and the mediation of soil
organic matter dynamics. In Soil Organic Matter Dynamics and
sustainability of tropical Agriculture. Eds. K Mulongoy and R
Merckx. pp 257–268. John Wiley and Sons, New York.
Elustundo J, Angers P A, Laverdiere M R and N’Dayegamiye A 1990
Etude comparetive de l’agregation et de la matiere organique
associee aux fractions granulometriques de sept soils sous culture
de mais ou en prairie. Can. J. Soil Sci. 70, 395–202.
Favejee J Ch L 1949 The mineralogical composition of the clay
fraction of Dutch soils. Landbouwk. Tijdschr. 61, 167–171 (In
Dutch).
Feller C, Fritsch E, Poss R and Valentin C 1991 Effet de la texture
sur le stockage et la dynamique des matieres organiques dans
quelques sols ferrugineux et ferrallitiques, Cahier Otstom Ser.
Pedologie 26, 25–36.
Garwood E A, Clement C R and Williams T E 1972 Leys and soil
organic matter III. The accumulation of macro-organic matter in
the soil under different swards. J. Agric. Sci. Camb. 78, 333–341.
Genstat Manual 1987 A General Statistical Program. Clarendon,
Oxford.
Greenland D J, Wild A and Adams D 1992 Organic matter dynamics
in soils of the tropics-from myth to complex reality. In Myths and
Science of Soils of the Tropics. Proceedings of an International
Symposium. Eds. R Lal and P A Sanchez. pp 17–33. ASA/SSSA,
Madison, WI.
Harter R D and Stotzky G 1971 Formation of clay-protein complex-
es. Soil Sci. Soc. Am. J. 35, 383–389.
Hassink J 1994 Effects of soil texture and grassland management on
soil organic C and N and rates of C and N mineralization. Soil
Biol. Biochem. 26, 1221–1231.
Hassink J 1995 Density fractions of soil macroorganic matter and
microbial biomass as predictors of C and N mineralization. Soil
Biol. Biochem. 27, 1099–1108.
Hassink J, Matus F J, Chenu C and Dalenberg J W 1997 Interactions
between soil biota, soil organic matter and soil structure. Adv.
Agroecol. (In press).
Huang P M 1990 Role of soil minerals in transformations of natural
organics and xenobiotics in soil. In Soil Biochemistry, Vol. 6.
Eds. J M Bollag and G Stotzky. pp 29–115. Marcel Dekker Inc.
New York.
Jenkinson D S 1988 Soil organic matter and its dynamics. In Russel’s
Soil Conditions and Plant Growth, 11th ed., Ch. 18A. Ed. A Wild.
pp 564–607. Longman, New York.
Kodama H 1979 Clay minerals in Canadian soils: their origin, dis-
tribution and alteration. Can. J. Soil Sci. 59, 37–58.
Leinweber P and Reuter G 1992 The influence of different fertil-
ization practices on concentrations of organic carbon and total
nitrogen in particle-size fractions during 34 years of a soil forma-
tion experiment in loamy marl. Biol. Fert. Soils 13, 119–124.
Lugo A E and Brown S 1993 Management of tropical soils as sinks
or sources of atmospheric carbon. Plant Soil 149, 27–41.
Marshman N A and Marshall K C 1981 Bacterial growth on proteins
in the presence of clay minerals. Soil Biol. Biochem. 13, 127–134.
Mebius L J 1960 A rapid method for the determination of organic
carbon in soil. Anal. Chim. Ada 22, 120–124.
North P F 1976 Towards an absolute measurement of sail structural
stability using ultrasound. J. Soil Sci. 27, 451–459.
Pinck L A, Dyal R S and Allison F E 1954 Protein-montmorillonite
complexed, their preparation and the effect of soil micmorgan-
isms on their decomposition. Soil Sci. 78, 109–118.
Quiroga A R, Buschaiazzo D E and Peinemann N 1996 Soil organic
matter particle size fractions in soils of the semiarid Argentinian
Pampas. Soil Sci. 161, 104–107.
Robert M and Chenu C 1992 Interactions between soil minerals and
microorganisms. In Soil Biochemistry, Vol. 7. Eds. G Stotzky and
J M Bollag. pp 307–404. Marcel Dekker Inc., New York.
Stevenson F J 1982 Humus Chemistry. John Wiley and Sons, New
York.
Stotzky G 1986 Influence of soil mineral colloids on metabolic pro-
cesses, growth, adhesion, and ecology of microbes and viruses. In
Interactions of Soil Minerals with natural Organics and Microbes.
Eds. P M Huang and M Schnitzer. pp 305–428. SSSA Spec. Publ.
No. 17. Madison, WI.
Tate K R and Theng B K G 1980 Organic matter and its interactions
with inorganic soil constituents. In Soil with a variable Charge.
Ed. G K G Theng. pp 225–249. New Zealand Soc. Soil Sci.,
Lower Hutt.
Theng B K G 1979 Formation and properties of clay-polymer com-
plexes. Elsevier, Amsterdam.
Tiessen H and Stewart J W B 1983 Particle-size fractions and their
use in studies of soil organic matter. II. Cultivation effects on
organic matter composition in size. Soil Sci. Soc. Am. J. 47,
509–514.
Turchenek L W and Oades J M 1979 Fractionation of organo-mineral
complexes by sedimentation and density techniques. Geoderma
21, 311–343.
Van der Marel H W 1949 Mineralogical composition of a heath
podzol profile. Soil Sci. 67, 193–207.
Van Veen J A and Kuikman P J 1990 Soil structural aspects of decom-
position of organic matter by microorganisms. Biogeochem. 11,
213–233.
Varadachari Ch, Mondal A H, Nayak D C and Ghosh K 1994 Clay-
humus complexation: effect of pH and the nature of bonding. Soil
Biol. Biochem. 26, 1145–1149.
Wilding L P, Smeck N E and Drees L R 1977 Silica in soils: quartz,
cristobalite, tridymite, and opal. In Minerals in Soil Environ-
ments. Ed. R C Dinauer. pp 471–552. SSSA, Madison, WI.
Zhang J, Thompson M L and Sandor J A 1988 Compositional dif-
ferences in organic matter among cultivated and uncultivated
argiudolls and hapludalfs derived from loess. Soil Sci. Soc. Am.
J. 52, 216–222.
Section editor: R Merckx
plso6537.tex; 17/06/1997; 18:52; v.7; p.11