Comparison of analytical methods in testing soil fertility

Analytical methods for testing soil fertility were compared in a material of 430 topsoil samples. The samples were analyzed for particle-size distribution, organic carbon content, pHfCaCy, exchangeableCa and Mg extracted with 1 M ammonium acetate (pH 7) and I M KCI, exchangeable K extracted with 1 M ammonium acetate (pH 7) and P extracted by the Bray 1 method. These soil properties were compared with the soil textural class and humus content class estimated visually, pH(H20) and Ca, Mg, K and P extracted with acid ammonium acetate. The estimation of soil textural class was quite successful, but the content of organic matter was frequently underestimated. pH(H 20) and pH(CaCl 2 ) were highly correlated and 95 % of the variation in pH(H zO) was explained by pH(CaCl2). Exchangeable Ca together with pH(CaCl 2) explained about 90 % of the variation in Ca extracted with acid ammonium acetate. Exchangeable Mg explained about 70 % of the variation in Mg extracted with acid ammonium acetate. Exchangeable K explained 90 % of the variation in K extracted with acid ammonium acetate. The Bray 1 P and pH(CaCl 2 ) explained 60 % of the variation in P extracted with acid ammonium acetate. pH(CaCl 2), clay and organic carbon content explained 72 —83 % of the variation in Ca. Mg, K and P were not highly dependenton pH, particle-size distribution and organic carbon content of soil.


Introduction
For the estimation of soil fertility status, various methods have been developed.In Finland, acid ammonium acetate extraction (Vuorinen and Mäkitie 1955) is extensively used in testing the soil nutrient status.In soil testing, soil textural class and humus content are estimated visually and soil pH is measured in water suspension.The purpose of this study was to compare the results of the adopted soil testing analysis with other soil test values.The reference methods were chosen among those widely used in other countries.

Material and methods
The research material consisted of 430 plough layer (0 -25 cm) samples collected for mapping down soil characteristics at the agricultural area of the Viikki Experimental Farm.The characteristics of the soil samples have previously been described by Jokinen  (1983 and 1984).
Index words: exchangeable Ca, Mg, K, extractable P, humus content class, pH, textural class, soil testing 183 JOURNAL OF AGRICULTURAL SCIENCE IN FINLAND     The samples were air-dried and ground to pass a 2-mm sieve.The particle-size distribu- tion of the inorganic matter in the soil samples was determined by the pipette method (Elonen 1971).The pH of the soil was measured in a stirred soil-0.01M CaCl 2 suspension (1 : 2.5) (Ryti 1965).The organic carbon content of the soil samples was determined using a modified (Graham 1948) Alten's wet combustion method.Exchangeable Ca and Mg were extracted from 10 g soil by four suc- cessive treatments with 50 ml of 1 M ammo- nium acetate (pH 7.0) at... with 50 ml of 1 M KCI and determined by aiomic absorption spectrophotometry.Exchangeable K was ex- tracted with 1 M ammonium acetate (pH 7.0) and determined by flame photometry.Phos- phorus was extracted with Bray 1 extractant (0.03 M NaF, 0.025 M HCI) (Bray and  Kurtz 1945), the extraction ratio being 1 : 10 w/v (Kaila 1965) and determined by a molybdenum blue method (Kaila 1955).The pH of the soil-H 2 0 suspension (1 : 2.5), and Ca, Mg, K and P extractable with acid ammonium acetate (0.5 M acetic acid, 0.5 M ammonium acetate, pH 4.65, ratio 1 : 10 v/v) (Vuorinen and Mäkitie 1955) were determined at a com- mercial soil testing laboratory (Viljavuuspalvelu Oy).The textural class and the humus content of the soils were estimated visually at the same laboratory.
The particle sizes of soils were classified as follows: Particle-size fraction Diameter, mm The textural classes aimed at were as follows: Coarser soils are classified according to the dominating coarser fraction.Gyttja clay is a soil with a clay content over 30 % and with a 3-6 % organic matter content in the subsoil.Soils with a clay content under 30 % may be defined as clayey if the clay content is considerable.Coarser soils with a considerable silt content may be defined as silty and silt soils with considerable content of coarser fractions as sandy.
The soils were classified according to the content of organic matter as follows:

Content of
Corresponding Humus organic matter content of organic content in the soil, % The content of organic matter is obtained by multiplying the organic carbon content (%) with the coefficient 1.72.

Results
Soil textural class (visual estimation) and particle-size distribution.Particle-size distri- bution is presented in Table 1.The particle- size distribution varied greatly in different fields.However, the proportion of sand frac- tions was low in the whole area and finesand fractions averaged 20 and 28 °7o of the particle-size distribution.The mean clay and silt contents were 31 and 17 % of the inorganic material.
The greatest group consisted of 169 coarser finesand samples (Table 2).In this group the mean content of coarser finesand fraction was 47 % of the inorganic material, the contents ranging from 28 % to 81 %.The group of finer finesand samples included only 15 samples in which the content of finer finesand fraction averaged 25 %.In addition, 12 sam- ples were identified as finesand soils without a more accurate classification.In these sam- ples the proportions of finer and coarser 184 T a b l e finesand fractions were almost equal (on the average 24 % and 28 %, respectively).Four clay soils were also included in the finesand soil groups in which the clay content was 34 % at its highest.
The second greatest group included 154 sandy clay soils with a mean clay content of 40 % (Table 2).The content of clay fraction ranged from 28 % to 59 %.The clay content of two samples was under 30 %, thereby these sam- ples were not clay soils.The group of sam- ples identified as silty clay soils included 40 samples in which the mean contents of clay and silt fractions were 41 and 31 %, respec- tively.However, the content of coarser frac- tions was over 20 % in all samples, and all the samples in this group should be classified as sandy clays.
Four samples were identified as heavy clays, and their content of clay fraction was over 60 %.The group of gyttja clays consisted of 28 samples, average clay content 44 %.Only one sample was identified as silt soil.In this soil the content of silt fraction was under 50 % as well as the content of coarser frac- tions.So, this soil belongs to the group of loam soils.
Fifty-four samples the clay content of which averaged 25 %, range 17-34 %, were defined as clayey (Table 2).The clay content was over 30 % in four samples of this group.These samples should be classified as clay soils.
Twenty-three samples (excl.silty clays) the silt content of which averaged 27 %, range 21-35 %, were defined as silty.Thirty-two sam- ples (excl.sandy clays) in which the finer and coarser finesand fractions ranged from 8 to 33 % and from 3 to 33 %, respectively, were defined as sandy.
Organic carbon content and humus content class.Organic carbon content averaged 5.2%, range 1.8-14.6%(Table 1).The mean organic carbon content was highest (7.0 %) in gyttja clays and was high also in the silty and sandy clays (6.6 % and 5.4 %, respectively) (Table 3).The humus content class »medium» included 327 samples in which the organic carbon content averaged 4.4 %, range 1.9-11.2% (Table 3).On the basis of organic carbon content 186 samples of this group should be classified as »rich» and 25 samples as »very rich».The humus content class »rich» included 88 samples of an average organic carbon content of 7.4 %, range 2.5 13.2 %.On the basis of organic carbon con- tent 57 samples should be classified as »very rich», one sample as mould soil and one sam- 1.8 2.5-13.2Very rich 10 9.2 2.0 6.9 -14.2 pie as »medium».The humus content was considered very rich in 10 samples, one of them should be included in mould soils.Three samples were classified as mould soils.Soil pH.The pH measured in water suspen- sion ranged from 4.2 to 7.1, mean 5.8 (Table 4).The mean pH measured in 0.01 M CaCl 2 suspension was 5.3, range 4.0-6.8.The pH values measured in water and 0.01 M CaCl 2 suspensions were highly correlated (r = o.9B***).The regression equation pH(H 2 0) = 0.94 + 0.91 pH(CaCl 2 ) explained 95 % of the variation in pH(H 2 0).
The applicability of the regression equation in predicting the mean pH(H 2 0) values in different fields was tested so that pH(H 2 0) values were calculated using the mean pH(Ca€l 2 ) values in different fields.The pre- dicted pH(H 2 0) was equal to the mean pH(H,O) in the fields 54, 84 and 89.In the fields 49, 94, 96 and 98 the predicted value was 0,1 pH unit higher, in the fields 88 and 97 0.1 and in the field 86 0.2 pH units lower than the mean pH(H 2 0).The pH measured in water or 0.01 M CaCl 2 suspensions was not highly related to particle-size distribution or organic carbon content of soil.The organic carbon content explained 42 % of the variation in pH(H 2 0) and 32 % of the variation in pH(CaCl 2 ).
When the effect of silt fraction was also con- sidered, 46 % and 36 % of the variation in pH(H 2 0) and pH(CaCl 2 ), respectively, was explained.
The pH(CaCl 2 ) and neutral ammonium acetate-or potassium chloride-extractable cal- cium together explained most of the variation in acid ammonium acetate-extractable calcium.Acid ammonium acetate-extractable calcium was dependent on pH(CaCl 2 ) and calcium extracted with 1 M ammonium acetate (pH 7) according to the equation Ca = -2039 + 506 pH + o.47Ca(Acet.pH 7).The coefficient of determination was 91.2 %.In different fields the values of acid ammonium acetate- extractable calcium calculated on the basis of the mean pH(CaCl 2 ) and calcium extracted The dependence of the exchangeable cal- cium on soil properties was studied using clay and silt content (%), organic carbon content (%) and pH(CaCl 2 ) as independent variables in the regression analysis.The pH(CaCl 2 ) of soil alone explained 54 % of the variation in calcium extracted with acid ammonium acetate.Adding the clay content to the variables increased the coefficient of determination to 74 %.Adding the organic carbon content to variables increased the coefficient of determination to 82.5 °7o, the regression equation being Ca = -4744+ 16.8 clay-% + 1037 pH + 93org.C-%.Including silt content to the variables increased the coefficient of determination only to 84 °Io.
Calcium extracted with 1 M ammonium acetate (pH 7) was not highly correlated with pH, clay nor organic carbon content alone, but pH(Ca€l 2 ) and organic carbon content together explained 63 % of the variation in calcium.Addition of clay content to the variables increased the coefficient of determination to 75 %, the regression equation being Ca(Acet.pH7) = -6104 + 22.8c1ay-% + 1205 pH + 2610rg.C-%.The silt content was an insignificant explainer.
The organic carbon content and pH(CaCl 2 ) of soil explained 55 % of the variation in calcium extracted with 1 M KCI.Addition of clay content to the variables increased the coefficient of determination to 72 %.With these variables the regression equation was Ca(KCI) = -4BBB + 24.8 clay-% + 953 pH + 2350rg.C-%.Exchangeable magnesium.The content of magnesium extracted with acid ammonium acetate averaged 163 mg/1 soil, range 25 - 1850 mg/1 soil (Table 6).The mean contents of magnesium extracted with 1 M ammonium acetate (pH 7) and 1 M KCI were 187 and 167 mg/kg soil, respectively.Magnesium extracted with 1 M ammonium acetate (pH 7) explained 72.4 % of the variation in magnesium ex- tracted with acid ammonium acetate, the re- gression equation being Mg = -23.6 + 1.0 Mg(Acet.pH 7).Considering the effect of pH(CaCl 2 ) the coefficient of determination which was in this case 73.3 % did not essentially increase.In different fields the values of acid ammonium acetate-extractable magne- sium calculated on the basis of the mean values of magnesium extracted with 1 M am- Table 6.Exchangeable magnesium extracted by acid ammonium acetate, 1 M ammonium acetate (pH 7) and 1 M KCI in the research material.

Field Number
Acid NH 4 OAc extractable 1 M NH4

2.3
Magnesium extracted with 1 M KCI ex- plained 68.6 % of the variation in acid am- monium acetate-extractable magnesium, the regression equation being Mg = -lO.O + I.OMg(KCI).Considering the effect of pH(CaCl 2 ), the coefficient of determination increased to 70.2 °7o.
Exchangeable magnesium was not highly related to the particle-size distribution, organic carbon content nor pH(CaCl 2 ) of soil.Clay and silt content together explained only 28 % of the variation of acid ammonium acetateextractable magnesium.These variables together with pH(CaCl 2 ) explained 50 %of the variation and adding the organic carbon content to the variables increased the coefficient of determination only to 52 %.Clay content and pH(Ca€l 2 ) together ex- plained 56 % of the variation in the content of magnesium extracted with 1 M ammonium acetate (pH 7) and 52 % of the variation in the content of magnesium extracted with 1 M KCI.Clay and silt content together with pH(CaCl 2 ) explained 64 °7o of the variation in acetate-extractable magnesium and 60 % of the KCI-extractable magnesium.Adding the content of organic carbon to the variables increased the coefficient of determination to 66 % (1 M acetate-extractable Mg) and 62 % (1 M KCI-extractable Mg).
Exchangeable potassium..The content of potassium extracted with acid ammonium acetate averaged 224 mg/1 soil, range 20 - 1250 mg/1 soil (Table 7).The content of potassium extracted with 1 M ammonium acetate (pH 7) averaged 305 mg/kg soil, range 53 - 1440 mg/kg soil.Potassium extracted with acid ammonium acetate was highly correlated with potassium extracted with 1 M ammonium acetate (r = o.9s***).The regression equation was K = -3.07+ o.7sK(Acet.pH7) and the coefficient of determination 90.3 °7o.Addition of pH(CaCl 2 ) to variables increased the coef- ficient of determination to 92.3 °7o.In dif- ferent fields the values calculated from the regression equation deviated from the mean content of potassium extracted with acid ammonium acetate as follows: Field Deviation, °7o (No. of Field The dependence of exchangeable potassium on soil properties was weak.The content of clay and organic carbon together with pH(CaCl 2 ) explained only 19 % of the va- riation in potassium extracted with 1 M am- monium acetate (pH 7).The content of clay, silt and organic carbon together explained 22 % of the variation in potassium extracted with acid ammonium acetate.
Extractable phosphorus.Phosphorus ex- tracted with acid ammonium acetate averaged 23 mg/1 soil and ranged from 4 to 92 mg/1 soil (Table 8).Phosphorus extracted by the Bray 1 method averaged 112 mg/kg soil, range 2-355 mg/kg soil.The correlation between phosphorus extracted by these two methods was not very close (r = o.62***).Phosphorus extracted by the Bray 1 method explained only 37.7 »ft of the variation in phosphorus ex- tracted with acid ammonium acetate.When the effect of pH(CaCl 2 ) was considered, the coefficient of determination increased to 59.5 %, the regression equation being P = -42.4+ 9.9 pH + 0.11P (Bray).Adding the content of clay and organic carbon to the variables increased the coefficient of determination only to 62.1 %.Extractable phosphorus was only weakly re- lated to soil properties.The content of silt and organic carbon together with pH(CaCl 2 ) ex- plained 31.5 % of the variation in acid am- monium acetate-extractable phosphorus.The variable which explained the variation in phosphorus extracted by the Bray 1 method was primarily the clay content.However, it explained only 26.7 % of the variation in phosphorus.When the effects of pH(CaCl 2 ) and the content of silt fraction were also con- sidered, the coefficient of determination in- creased to 30 %.

Discussion
In the visual humus content classification, the content of organic matter was frequently underestimated.Only 36 % of 327 samples classified in the humus content class »me- dium» truly belonged to this class.On the basis of organic carbon content the majority of these samples should be classified as »rich».
The humus content class »rich» included 88 samples of which 33 % truly belonged to this class, while nearly all other samples should be classified as »very rich».A factor rendering the humus content classification more difficult is the relatively high content of organic mat- ter in the present material.If soils with greater variation in the content of organic matter are not included for comparison, the possibility of faulty estimations increases.In the visual classification of soil textural class, clay soils were well distinquished from coarser soils.In the present material of 430 samples only four clay samples were included in coarser soil groups.However, classification of clay soils into groups of silty and sandy clays was not equally successful.According to the estimation, these clay soil groups included a total of 194 samples.About 20 % of them were estimated as silty clays which was not a true classification.If the criterion of silty clay is a content of coarser fractions under 20 %, all 194 samples are sandy clays.
The values of pH(H 2 0) and pH(CaCl 2 ) were highly correlated, which is in agreement with previous results of Ryti (1965), Mänty-  lähti and Yläranta (1980) and Sillanpää  (1982).In the material of Ryti (1965) the mean difference between pH(H 2 0) and pH(Ca€l 2 ) was 0.5 pH units.Also in the present study the pH(H z O) values were about 0.5 pH units higher than the pH (CaCl 2 )   values.The regression equation between pH(H 2 0) and pH(CaCl 2 ) obtained in this study agreed fairly well with the regression 191 192 equation obtained by Ryti (1965) for sand and finesand soils (pH(H 2 0) = 0.81 + 0.94 pH(Cad 2 ), n = 109, r = o.97***).The mean pH(H 20) values of different fields were well predicted with the regression equation cal- culated from the whole material.The difference between measured and calculated value was no more than 0.1 pH units exclud- ing field No. 86 where the content of organic carbon was lower than in most other fields.The values of pH(H 2 0) and pH(CaCl 2 ) were not closely related to the soil particle-size distribution or organic carbon content.This could also be expected on the basis of a previous study with nearly the same material (Jokinen 1984).
In cultivated soils, the majority of the cat- ion exchange capacity is saturated with ex- changeable calcium.According to Kaila   (1972), the degree of saturation with calcium varies from 60 % to 80 %.Under these cir- cumstances it was not surprising that pH(CaCl 2 ), clay content and organic carbon content, factors on which the cation exchange capacity is highly dependent (Heinonen 1960,   Marttila 1965, Kaila 1971), explained more than 70 % of the exchangeable calcium ex- tracted by three methods.More than 90 % of the variation in acid ammonium acetate extractable-calcium was explained by together with calcium extracted with 1 M am- monium acetate (pH 7) or 1 M KCI.The mean values of acid ammonium acetate-extractable calcium in different fields were predicted rather well on the basis of pH and calcium ex- tracted by the comparative methods.How- ever, the contents were not similar.The acid acetate extracted much less calcium from the soil than did the neutral acetate or KCI.The different extraction method explains part of the difference.According to the regression analysis, an increment of one unit of the latter values corresponded to 0.6 unit of the former.
The magnesium status fluctuated more than did calcium status in the experimental area.Exchangeable magnesium, a minor component in saturation of cation exchange capaci-ty, was not very highly related to soil charac- teristics which explained 52-66 % of the variation in exchangeable magnesium.Ac- cording to Kaila (1972), 10-30 % of the cation exchange capacity of cultivated soils is saturated with magnesium.The mean values of acid ammonium acetate-extractable magnesium in different fields were not predicted as well as calcium values on the basis of comparative methods.At its highest the deviation was about 30 %.On the average, the contents determined by different methods did not de- viate very much from each other.According to the regression analysis, the mean dif- ferences between samples were equal irre- spective of the method used.
As previously observed (Jokinen 1984), exchangeable potassium was poorly related to soil pH, particle-size distribution and organic carbon content.Potassium extracted with acid ammonium acetate was highly related to po- tassium extracted with 1 M ammonium acetate.With the exception of one field (94, 6 samples), the mean values of acid ammonium acetate-extractable potassium were rather well predicted on the basis of the comparative method.The acid acetate extracted on an average 75 % of the potassium extracted by the neutral acetate.
Phosphorus extracted with acid ammonium acetate, representing relatively well available phosphorus in soil, was not highly related to phosphorus extracted by the Bray 1 method, which indicates the capacity factor of the soil phosphorus status (Kaila 1965).More phosphorus was extracted by the Bray 1 method than with acid ammonium acetate.In the study of Aura (1978) with 30 soils and in the study of Sippola and Jaakkola (1980)   with 20 soils, a 5-fold amount of phosphorus per liter of soil on the average was extracted by the Bray 1 method as compared with phosphorus extracted with acid ammonium acetate.The relationship between the methods in this study was also of the same magnitude.An accurate comparison was not possible be- cause the former content was expressed on a volume basis, the latter on a weight basis.In the study of Aura (1978), the phosphorus uptake of oats in four growings in pots was better explained by phosphorus extracted with acid ammonium acetate (77 %) than by phosphorus extracted by the Bray 1 method SELOSTUS Maan viljavuuden määrittämiseen käytettyjen analyysimenetelmien vertailu Raina Niskanen ja Antti Jaakkola Helsingin yliopisto, Maanviljelyskemian laitos, 00710 Helsinki 71 Tutkimusaineisto koostui 430 Viikin opetus-jakoetilan peltojen muokkauskerroksesta otetusta maanäytteestä.Näytteistä määritettiin lajitekoostumus, orgaaninen hiili, pH(CaCI 2 ), vaihtuva Ca ja Mg uutettuna 1 M ammo- niumasetaatilla (pH 7) sekä 1 M KChlla, vaihtuva K uutettuna 1 M ammoniumasetaatilla (pH 7) ja P uutet- tuna Bray 1 -menetelmällä.Näitä maan ominaisuuksia verrattiin Viljavuuspalvelu Oy:ssä tehdyn viljavuusana- lyysin tuloksiin.

Table 3 .
Organic carbon content in textural and humus content classes.

Table 4 .
pH values measured in water and 0.01 M CaCl 2 suspensions.

Table 5 .
Exchangeable calcium extracted by acid ammonium acetate, 1 M ammonium acetate (pH 7) and 1 M KCI in the research material.

Table 7 .
Exchangeable potassium extracted by acid ammonium acetate and 1 M ammonium acetate (pH 7) in the +

Table 8 .
Phosphorus extracted by acid ammonium acetate and the Bray 1 method in the research material.