JOURNAL OF THE SCIENTIFIC AGRICULTURAL SOCIETY OF FINLAND

195

Maataloustieteellinen Aikakauskirja
Voi. 48: 195—202, 1976

Factors affecting the availability to radishes
of cadmium added to soil

Antti Jaakkola1)

Department of Agricultural Chemistry, University of Helsinki
Matti J. Siren

Abstract. In a five week period, the cadmium uptake of radishes in pot experiments
made in the greenhouse was on average 7.8 % of the cadmium mixed with the soil (0.4
mg Cd per 400 ml of soil). In a material consisting of samples from 25 plough layers
the uptake varied from 1.4 to 11 % and in a material of samples from the corresponding
subsoils from 14 to 22.6 %. The increase in cadmium content (A c) of the radish tops,
which increase was regarded here as describing best the availability of added cadmium,
varied in the plough layer material from 2.2 to 27.1 mg/kg and in the subsoil material
from 4.3 to 94.9 mg/kg Cd on a dry matter basis. The increment A c was related to
the amount of organic carbon (xj) and to proportion of hydrogen (x 2)) of total

-ir b.
A *2 2exchangeable cations according to the equation A c =k • . In this equation

xi
h
i

o<bj < 1 and 0< b 2 '■ 1 In the complete material of 50 soil samples, the coefficient
of determination of the logarithmically transformed equation was R 2 = 0.73. The
corresponding coefficient of determination in a material of plough layer samples (R2

=

0.48) was imporoved when the increase in cadmium content of the radish tops grown
in corresponding subsoil samples was included as an additional independent variable
(R 2 = 0.59). This showed that some character other than any determined on the
experimental soil was acting in the plough layer material.

According to present ideas, large amounts of cadmium in the human diet
constitute a considerable risk to health (Friberg et ai. 1974). Cadmium in the
soil finds its way into the diet both directly in food plants and indirectly
through fodder crops and animal products. The most important source of
cadmium to plants is usually the soil, although considerable amounts may be
deposited locally from the air (Lagerwerff and Specht 1971). In areas fer-
tilized with sewage from settlements or industry, the soil is in particular
danger of contamination (Linnman et al. 1973). In addition, some fertilizers
contain cadmium levels which may increase the danger (Stenström and

2 ) Present address: Agricultural Research Centre, Department of Agricultural Chemistry
and Physics, 01300 Vantaa 30.

https://www.c-info.fi/en/info/?token=xK-NrIdNjh86yIaR.84OYpnFW5vt5Ln5HIOOywg._Nf5R5LBlXqkG3P2dSoMW7qlsEyriNcfVDII8rr_qjvBV3CjLuzNg5WaLSispJrq0NX_CKcVxvKZVmYjw_yLbH-B3R4BX4ai3SDRylV1FI5_4I6LkBCfwnhTDbbkC24S25emtfMrMmIzsYQm3QHs8EXbzcRUyhqfxt7kIVyb2qWw9YoYvoLI1MJ40AM


196

Vahter 1974). The aim of this study was to clarify the effects of certain soil
characteristics upon the availability of cadmium contaminating the soil.

Material and methods
The soil sample material consisted of 50 samples taken from 25 sites in

various parts of South and East Finland. At each site samples were taken
from the plough layer (topsoils) and below it to a depth of 40 cm (subsoils).
The samples were allowed to reach an air-dry condition, and were ground to
pass a 4 mm sieve.

A 400 ml sample of each soil was put into a plastic pot, and an equal
quantity weighed into another pot. The pots were fertilized with

N P KMg Mn Cu Zn B Mo
10 2 3 2 0.05 0.05 0.05 0.05 0,002 mmol per pot.

To one of each pair of pots 0.4 mg of Cd were added as the nitrate. Water
was added to the pots to give a soil water potential of pF 2.5. This tension
was measured with a pressure plate apparatus. In each pots, four radish plants
(Raphanus salivus L. cv. Non Plus Ultra) were grown from seed to 36 days
old in a greenhouse during April and May 1973. The water lost was made
up at intervals by watering the pots to their original weight.

The aerial (tops) and underground (roots) parts of the radish plants were
harvested separately. The roots were freed from the soil by washing with
deionized water. Yields were weighed air-dry. Subsamples were used for a
dry matter determination at 105° C. After a wet combustion, the cadmium
content was determined with a flameless atomic absorption spectrophotometer
(Perkin Elmer). The cadmium uptakes were calculated from yields and their
cadmium contents.

The following determinations were made on the experimental soils: pH
in 0.01 M CaCl 2 suspension, organic carbon, clay fraction), iron extract-
able in acid ammonium oxalate at pH 3.3, and also exchangeable hydrogen at
pH 7 and cation exchange capacity in soils saturated with I N ammonium
acetate. The amounts of organic carbon, clay, soluble iron and of exchangeable
hydrogen and cation exchange capacity were calculated for the soil in each
pot by multiplying the weight of soil by the corresponding content.

The statistical treatment of the results was based mainly on correlation
and regression analyses. Since the linear regression equation y= a + bjXj
+ b 2x 2 + b 3 x 3 + ....

+ bk xk embodies the assumption that the variables
x 1( x 2, x 3 etc. define additive parts of y, parts independent of one another,
and since this assumption is presumably not valid, it was decided to test another
type of model. The linear regression equation log y = log a' + b/ log x x +

v log x 2 + b3
' log x 3 + + bk

' log x k incorporating the log-
arithmically transformed variables was chosen for the test. When the
independent variables (x) influence one another, this model is more logical.

iti / v /

The equation can be expressed in the form y = a' Xj 1 x 2 2 x 3 3
xk^k ,in which it is clear that the relative effect of each variable is assumed to
be independent of the effect of other variables.



197

Table 1. Amounts and variation of some characters determined on experimental soils.

Topsoils Subsoils Whole material
n = 25 n = 25 n = 50

mean standard mean standard mean standard
deviation deviation deviation

soil, g/pot 385 60 466 47 425 67
org. C, g/pot 17.7 8.7 4.0 2.9 10.9 9.4
clay (< 2 /im), g/pot 134 85 177 117 155 104
CEC, mval/pot 129 48 113 53 121 51
exch. H, mval/pot 72 19 57 19 65 20
Fe, soluble in ammonium

oxalate, g/pot 3.62 1.99 3.73 2.30 3.67 2.13
PHCaCl 49 046 5,1 055 5 0 0,51

Table 2. Dry matter yield, its cadmium content and cadmium uptake of radish without
added cadmium.

Topsoils Subsoils Whole material
n = 25 n = 25 n = 50
mean standard mean standard mean standard

deviation deviation deviation

Tops
yield, g/pot 1.37 0.22 0.99 0.26 1.18 0.31
Cd content, /ig/g 0.83 0.51 0.50 0.32 0.66 0.45

Roots
yield, g/pot 1.74 0.45 1.48 0.55 1.61 0.51
Cd content, [ig/g 0.24 0.09 0.17 0.09 0.21 0.09

Cd uptake, /jg/pot 1.50 0.75 0.72 0.46 1.11 0.73

Results and discussion
On the basis of amounts present in the pots of soil (Table 1), the following

average contents were obtained;
topsoils subsoils whole material

org. C, % 4.6 0.9 2.6
clay, % of total weight 34.8 38.0 36.5
CEC, mval/100 g 33.5 24.2 28.5
Fe, mg/g, extractable in ammonium

oxalate, pH 3.3 9.4 8.0 8,6

The pots produced on average a dry matter yield of 2.8 g (Table 2), of
which the proportion of roots was a little over half. Both the top and root
yields of radishes grown on topsoils were on average significantly (P < 0.05)
greater than yields from subsoils. The native cadmium of the soil produced in
the material a cadmium content of the tops at least three times higher on
average than that of the roots. Deposition from the air was unlikely in this



Table 3. Response of dry matter yield, its cadmium content and cadmium uptake of radish
to cadmium application (400 /rg/pot).

Topsoils Subsoils Whole material
n = 25 n = 25 n = 50
mean standard mean standard mean standard

deviation deviation deviation

Tops
yield, g/pot 0.05 0.24 + 0.01 0.16 - 0.02 0.20
Cd content, //g/g +11.9 6.5 +35.1 21.3 +23.5 19.5

Roots
yield, g/pot - 0.04 0.45 + 0.07 0.31 + 0.02 0,38

Cd content, ug/g + 3.1 1,6 + 7.5 6.6 + 5.3 5.2

Cd uptake, //g/pot +20.2 10.3 +41.9 20.9 +31.1 19.6

experiment. Of the total cadmium uptake (Table 2) in the whole material,
an average of 70 % was present in the aerial parts. Cadmium uptake was
significantly greater from the topsoils than from the subsoils (P< 0.001).

Addition of cadmium to the soils before sowing had no influence on the
yield (Table 3). Differences in yield between the cadmium and no cadmium
treatments were probably due to random influences. The variations of the
differences were rather large and did not differ significantly from yield variations
in the no cadmium treatments (cf. Table 2). Thus, there is some evidence that
the latter variations were random, too, rather than dependent on differences
between soils.

The increase in cadmium content of the tops was on average 45 times as
large as the corresponding increase in the root content. In the whole material,
radishes took up an average of 7.8 % of the cadmium added. The range of
variability was from 1.4 to 11.0 % in the topsoils and from 1.4 to 22.6 % in
the subsoils. The increment in cadmium content of both tops and roots of
radishes grown on subsoils was significantly greater than the increment for
topsoils (P< 0.01). Owing to this, also the uptake of applied cadmium was
significantly greater from the subsoil than from the topsoil (P < 0.001).

Of the quantities determined, cadmium uptake probably provides in theory
the best measure of cadmium availability, whose assesment was aimed at.
The uptake of applied cadmium was calculated according to the equation
Ay =Yi-y« ; y* = c 1( a xi + c 2i a 2i ,(i = 1,0), where

/Jy = uptake of applied cadmium
yi = cadmium uptake
Cjj = cadmium content of tops
ajj = yield of tops
c 2j = cadmium content of roots
a 2 j = yield of roots
subscript i = 0 with no addition of cadmium

i = 1 after addition of cadmium

198



199

Table 4. Correlation coefficients between logarithms of difference in Cd content of tops and
of certain soil properties.

Topsoils Subsoils
n = 25 n = 25

Whole material
n = 50

organic C -0.58** -0.57*» -0.78*»*
clay -0.32 -0.55** -0.23
CEC -0,63** -0.58*» -o.sB***
exch. H -0.36 0.03 -0.33*
exch. H/CEC .... 0.44* 0.56** 0.31*
pH 1 ) -0.15 -0.38 -0.15
Fe -0.32 -0.62** -0.38»

*) untransformed values

It is apparent from the equation that all of the variation in yields is express-
ed in the calculated values of cadmium uptake. When this variation is large
(cf. Tables 2 and 3) and presumably random, it seems likely that the use of the
cadmiumcontent as the independent variable instead of the calculated cadmium
uptake would allow the attainment of a higher coefficient of determination and
a more correct result in the regression analysis. Since the bulk of the cadmium
in the plants is contained in the tops, variations were studied of the cadmium
content of the tops in particular. The error caused by ignoring the roots is
reduced by the fact that the cadmium contents of tops and roots showed a
close, positive correlation with each other (r = 0.956***). The procedure used
thus assumed that top yield is constant for the whole material, and that the
top yield contains a constant proportion of the total plant cadmium. The fact
that in almost all cases the cadmiumcontent was better correlated than cadmium
uptake with other soil characters tested argues in favour of the procedure. The
cadmium contents of plants have been used for assessing the availability of
soil cadmium by John et al. (1972) and Haghiri (1974).

Correlations between the logarithms of the change in cadmium contents
and of characters determined on the soils were rather poor (Table 4). In the
whole material, organic carbon was rather well correlated with the dependent
variable, but in each of the partial materials the correlation was poorer.

On the basis of pot experiments made with oats, Haghiri (1974) concluded
that the depression in cadmium content of the tops associated with an increase
in organic matter results solely from an increased cation exchange capacity.
The results of the present study do not justify such an conclusion, since the
cation exchange capacity accounted for significantly less (r 2 = 0.33) of the
variation in the cadmium content of the radish than did organic carbon (r 2 =

0.61). When organic carbon was taken into account in addition to the exchange
capacity, the coefficient of determination was raised from 33 % to 68 %

(P < 0.001) in the whole material and from 34 % to 44 % (P < 0.05) for the
subsoils. The increase from 39 % to 48 % for the topsoils was scarcely signif-
icant (0.05 < P< 0.10).

When the effects of organic carbon, clay, and of cation exchange capacity
correlating well with these (R 2 = 0.80) were considered simultaneously (Table



200

5), organic carbon was found to be the most important independent variable
in the whole material and in the subsoil material. The coefficients of determina-
tion for the alternatives I and II in Tabel 5 did not differ significantly, but
the effect of the proportion of exchangeable hydrogen (II) was more consistent
than that of the cation echange capacity (I). The elimination of clay, whose
correlation with the dependent variable is not significant, led to alternative
111, to be considered in more detail below. The coefficient of determination
for the whole material was 73 %, or greater than in the topsoil material (48 %)

or the subsoil material (59 %).

The regression equations calculated according to alternative 111 in Table
5 express the following relationships:

x 0.83

A c = 116.6 • (R 2 = 0.48 for logarithmic transformation)
xi

topsoils

x 0.72

Ac = 89.2 • (R 2 = 0.59 - *-)
y 0.61X 1

subsoils

x 2 »- 74
whole material Ac = 99.0 (R 2 = 0.73 » —), in which equations

X 1 *■

A c = the difference, mg/kg, in cadmium contents between tops from cadmium and no
cadmium treatments,

Xj = amount of organic carbon, g/pot
x 2 = proportfon oi hydrogen of exchangeable cations.

The exponentials (regression coefficients after logarithmic transformation)
of all the equations given were significant. The high hydrogen proportion
and the low amount of organic carbon thus represent a large increase in the
cadmium content of the tops. The simplest explanation is that the increase in
cadmium content is directly proportional to the proportion of exchangeable
hydrogen, and inversely proportional to the amount of organic carbon. While
this hypothesis regarding exchangeable hydrogen could not be disproved,
both in the subsoil material and in the whole material the exponential of the
amount of organic carbon was significantly smaller than unity. This indicates
that a given relative increase in organic carbon corresponds on average to a
smaller relative decrease in cadmium content response of radish tops.

John et ai. (1972) have also established that the cadmium content of radish
tops depends upon the content of soil organic matter and soil acidity expressed
as pH measured on a suspension of soil in salt solution. In their material, an
additional independent variable was provided by a separate determination of
the soil capacity for retaining cadmium from a solution. Part of the variation
accounted for by this new variable would otherwise have been explained by
organic matter and pH. In the material of John et al. (1972), the iron soluble
in dilute mineral acids raised the coefficient of determination considerably.
In the material of the present study, however, iron extracted from the soil
with acid ammonium oxalate at pH 3.3 had no significance as an independent
variable after the effects of organic carbon and exchangeable hydrogen had
been eliminated. The present material is not strictly comparable with that of
John et al. (1972), as they determined iron by a different method. The iron
content in their material averaged about 80 mg/kg, in the present material
8 600 mg/kg.



201

Table 5, Partial correlation coefficiants (r) and coefficient of determination (R2) between
increase in cadmium content in radish tops (y) and organic carbon (1), clay fraction (2), cation
exchange capacity (3) and hydrogen percentage of exchangeable cations (4) in soil (logarithmic
transformation).

Topsoils
n = 25

Subsoils Whole material
n = 25 n = 50

I. Omitting effects of hydrogen percentage
ryl 23 -0.34 -0.45* -o.69***
ry2.i 3 0.00 -0.26 -0.11
ryBl2 -0.35 -0.01 -0.20
H 2 y. 123 0.48** 0.48** o.6B***

11. Omitting effects of cation exchange capacity
ryl 24 -o.s9*** -0.60** -0.84**»
r y2.j 4 —0.05 —0.04 —0.03
r y4 12 0.33 0.46* 0.42**
R2ym 0.48** o.s9*** o.73***

111. Omitting effects of clay fraction and cation exchange capacity
r yl 4 -0.59** -0.63** -o.B3***
r y4 j 0.47* 0.63** o.s6***
R2y,14 o,4B*** o.s9*** o.73***

The higher coefficient of determination in the whole material than in either
of the component materials is due to differences in the partial correlation
coefficients incorporating organic carbon. On the basis of its level of organic
carbon, the material could be naturally divided into two components, whose
mean responses of the cadmium content of the tops to added cadmium were
different (cf. Table 3). This difference in means is not necessarily due solely
to a difference in the amount of organic carbon between the component materials,
but may also be caused by some character not determined here.

A larger proportion of the variation in increase of the cadmium content of
radish tops could be accounted for in the subsoil material than in the topsoil
material. It is possible that some character common to the plough layer and
subsoil and not investigated here was influencing the results. For these reasons,
an attempt was made to improve the coefficient of determination in the topsoil
material by taking into account as an independent variable the increase in
cadmium content of radishes grown in the corresponding subsoil sample. This
measure increased significantly the coefficient of determination, from 48 %

to 57 %. The nature of the common character was quite unresolved. When
the increase in cadmium content of radishes grown in subsoil was replaced by
the proportion of hydrogen of subsoil exchangeable cations as an independent
variable, the coefficient of determination did not fall significantly (to 55 %),

but in any case this latter variable was not significant (0.05 < P < 0. 10). It
is conceivable that the proportion of exchangeable hydrogen in the subsoil
characterizes soil acidity under constant conditions when organic carbon does
not cause interference by its variations which are nevertheless apparent in the
plough layer.

Acknowledgement. The authors wish to express their gratitude to Elintarviketuotannon
edistämissäätiö (the Foundation for Promoting Foodstuff Production) for a grant which made
it possible to carry out this study.



202

REFERENCES

Friberg, L., Piscator, M., Nordberg, G. F. & Kjellström, T. 1974. Cadmium in the
Environment. 2nd Ed. 248 p. Cleveland.

Haghiri, F. 1974. Plant uptake of cadmium as influenced by cation exchange capacity,
organic matter, zinc, and soil temperature. J. Environ. Quality 3: 180 189.

John, M. K., VanLaerhoven, C. J. & Chuah, H, H. 1972, Factors affecting plant uptake
and phytotoxicity of cadmium added to soils. Environ. Sci & Technol. 6: 1002 1009.

Lagerwerff, J. V. & Specht, A. W. 1971. Trace substances in environmental health. IV.
Occurrence of environmental cadmium and zinc, and their uptake by plants. Proceed-
ings of University of Missouri’s annual conference on trace substances in environmental
health. Colombia, Missouri 1971. p. 85 93.

Linnman, L., Andersson, A., Nilsson, 0., Lind, 8., Kjellström, T. & Friberg, L. 1973.
Cadmium uptake by wheat from sewage sludge used as a plant nutrient source. Arch,

Environ. Health 27:45 47.
Stenström, T. & Vahter, M. 1974. Cadmium and lead in Swedish commercial fertilizers.

Ambio 3: 91-92.

MC received March 17, 1976.

Selostus

Maahan lisätyn kadmiumin käyttökelpoisuus retiisille astiakokeessa

Antti Jaakkola1 )

Maanviljelyskemian laitos, Helsingin Yliopisto
Matti J. Sir£n

Tutkimuksessa pyrittiin selvittämään eräiden maan ominaisuuksien vaikutus maahan
lisätyn kadmiumin käyttökelpoisuuteen kasville. Tutkitut ominaisuudet olivat orgaaninen
hiili, saves, kationinvaihtokapasiteetti, vedyn osuus vaihtuvista kationeista ja happameen
ammoniumoksalaattiin liukeneva rauta.

Maanäyteaineiston muodostivat 25 paikasta etelä- ja itä-Suomesta otetut näytteet. Jokai-
sesta paikasta otettiin näyte kyntökerroksesta ja jankosta. Kasvihuoneessa järjestetyssä
astiakokeessa määritettiin maihin sekoitetun kadmiumnitraatin (0.4 mg Cd/400 ml maata)
vaikutus viiden viikon ikään kasvatetun retiisin kadmiumpitoisuuteen ja kadmiuminottoon.
Lisätystä kadmiumista retiisin laskettiin ottaneen yhteensä 1.4 22.6 %. Maanpäällisten osien
kadmiumpitoisuuden nousun (A c) ja koeastian sisältämän orgaanisen hiilen määrän (xj sekä

. 2vaihtuvan vedyn osuuden (x 2) välillä todettiin yhteys ZJ c= k •
—— . Lisäksi osoitettiin, että
*i i

o<bj<lja 0 < b 2 1. Koko 50 näytteen aineistossa, jossa A c vaihteli 2.2 94.9
ing/kg Cd, tämän mallin logaritmimuunnos selitti 73 % vaihtelusta. Kyntökerrosnäytteiden
aineistossa, jossa selvitysaste oli huonompi (48 %) kuin jankkoaineistossa (59 %), vaihtelusta
pystyttiin selittämään enemmän (57 %) ottamalla lisäksi selittävänä muuttujana huomioon
retiisin kadmiumpitoisuuden nousu vastaavassa jankkoaineistossa. Tämän pääteltiin osoitta-
van, että kyntökerrokseen lisätyn kadmiumin käyttökelpoisuus riippui orgaanisen hiilen
ja vaihtuvan vedyn osuuden lisäksi jostakin ominaisuudesta, jota ei ollut määritetty kyntö-
kerrosnäytteistä ja joka oli korreloitunut jankossa vaikuttavien lisätyn kadmiumin käyttö-
kelpoisuutta säätävien tekijöiden kanssa.

l ) Nykyinen osoite; Maanviljelyskemian ja -fysiikan laitos, Maatalouden tutkimuskeskus,
01300 Vantaa 30