Soil-fertilizer reactions and plant response to coated and non-coated concentrated superphosphate


Material Information

Soil-fertilizer reactions and plant response to coated and non-coated concentrated superphosphate
Physical Description:
xi, 293 leaves : ill. ; 28 cm.
Nicholaides, J. J ( John J. ), 1944-
Publication Date:


Subjects / Keywords:
Phosphatic fertilizers   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1973.
Includes bibliographical references (leaves 281-292).
Statement of Responsibility:
by John J. Nicholaides, III.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000317480
notis - ABU4304
oclc - 08819515
System ID:

Full Text



John J. Nicholaides, III

A Dissertation Presentea to the Gracuate
Council of the University of Florida
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy




Gwynne and Elizabeth


The author wishes to express his sincere appreciation to

Dr. John G. A. Fiskell, Chairman of the Supervisory Committee, for his

valued counsel, guidance, and assistance throughout the entire course of

this study, and for his valuable suggestions and excellent assistance

in the preparation of this manuscript.

Appreciations are also extended to Drs. Lucian W. Zelazny,

David H. Hubbell, William G. Blue, and Salvadore J. Locascio for their

interest, constructive advice, participation on the Supervisory Committee,

and review of this manuscript. Recognition is made of the interest,

valued assistance, and participation on the Supervisory Committee of the

late Dr. Curtis E. Hutton.

Gratitude is expressed to Dr. Charles F. Eno, Chairman, Soil Science

Department, for awarding the NDEA Title IV Fellowship and for his words

of encouragement which facilitated this study.

Special appreciation must be extended to Dr. Frank G. Martin for

his useful statistical analyses of the field data.

Appreciations are extended to Dr. Raymond B. Diamond and Mr. George

Slappey of the Tennessee Valley Authority for furnishing the sulfur-

coated fertilizers used in this study.

To Mr. Frank Sodek and Mrs. Elma del Mundo of the Soil Science

Analytical Service Laboratory and to Mr. Lex Carver and Mr. James

Chichestar of the Extension Soil Testing Laboratory, appreciations are

extended for their labors with certain elemental analyses. Though space


does not permit mention of individual names, thanks are in order for

the aid of many others at the University of Florida and at the Agricul-

tural Research Center at Jay, Florida.

For his excellent draftmanship of the figures in the final draft,

appreciation is given to Mr. Ray Garman.

The author wishes to commend Mrs. Loan Tinsley for her excellence

in typing and proofing the first draft and the final copy of this dis-

sertation. Appreciations are also extended to Mrs. Susan Mickelberry

and Mrs. Carol O'Dell for their excellent typing and proofing of the

final draft of the appendix tables.

Above all, the author wishes to express the deepest gratitude to

his wife, Gwynne, whose understanding, unyielding encouragement, and

labor contributed immeasurably to the successful completion of this




ACKNOWLEDGEMENTS ............................................. iii

ABSTRACT .................................... .............. ix

INTRODUCTION ............................ .... ........... I

LITERATURE REVIEW ...................... ................... 3

Factors and Mechanisms Contributing to Fertilizer
Phosphate Retention by Soils ............................. 3
Mineralogical Reactions .......................... .... 3
Aluminum and Iron ................ ..................... 7
Exchangeable and crystalline aluminum and iron ..... 9
Amorphous aluminum hydroxides and iron oxides ...... 11
Organic Matter Complexes ............................. 16

Measurement of Phosphate Availability in Soils ........... 18
Indices for Phosphate Availability .................... 20
Phosphate Potential .....................**............ 24
Phosphorus Sorption Isotherms ....................... 26

Efficiency of Phosphate Fertilizers ...................... 28
Phosphate Fertilizers ............................... 28
Secondary Forms of Phosphate Fertilizers in Soil ...... 29
Plant Responses to Phosphate Fertilization ............... 30
Field Response ......................................... 30
Fertilizer placement ............... ............ 30
Residual effects ................. ............ 33
Foliar Composition ..........................**... 34

MATERIALS AND METHODS ........................................ 37

Site Selection .......................................... 37

Laboratory Experiments ................................ .. 38
Native Soil Phosphorus ............................... 38
Total phosphorus .................................. 38
Inorganic phosphorus fractionation ................ 38
Plant-available phosphorus ......................... 38
Soil Retention of Applied Phosphorus ................. 39
Phosphate potential ................................ 40
Phosphate sorption isotherms ....................... 41

Components Contributing to Soil
Phosphorus Retention ........................... .... 42
Organic matter removal ............................ 42
Citrate-dithionite-bicarbonate treatment ........... 43
Dissolution with hot sodium hydroxide .............. 44
Check treatments ...................... .......... 44
Phosphorus additions ........................... 45
Studies using 3ap ................................. 46
Fertilizer Phosphorus Dissolution .................... 47
Fertilizer dissolution without soil ................ 47
Fertilizer dissolution and retention in soil ....... 48

Field Experiments .....................................**. 48
Phosphorus Sources, Rates, and Placements, 1971 ....... 48
Phosphorus sources, rates, and placements .......... 48
Experimental design ................................. 49
Fertilization ...................................... 49
Planting and cultural practices .................... 50
Plant sampling .................................. ... 50
Soil sampling ..................... ............. 51
Harvesting .................................. ..... 52
Statistical analysis of data ....................... 52
Fertilization with Mixtures of Coated and
Non-coated Phosphorus Fertilizers, 1971 .............. 53
Design and cultural practices ..................... 53
Statistical analysis of data ....................... 54
Phosphorus Sources, Rates, and Placements, 1972 ....... 54
Fertilization ...................................... 54
Plant sampling .................................... 55
Soil sampling ..................... ..... .....*... 55
Harvesting ......................................... 56
Statistical analysis of data .........(............ 56
Various Combinations of Coated and Non-coated
Fertilizers with and without Strip Mulch, 1972 ....... 56
Fertilization, planting, and cultivation practices.. 57
Statistical analysis of data ....................... 57
Evaluation of Residual P from Previous Two Years,
1973 ...........e...***.*..**.********* ...*** *.. 58
Plant sampling ..................................... 58
Soil sampling ............ ......................... 59
Harvesting .............................*.....* .. 59
Statistical analysis of data ....................... 59

Analytical Procedures .................................... 59
Clay Mineralogy and Associated Procedures ............. 59
Pretreatments ............ ......................... 59
Clay mineralogical analyses ........................ 61
Selected clay properties .......................... 63
Clay chemical analyses ....................... .... 64
Soil Particle-Size Distribution ............. ........ 65
Soil Chemical Analyses ................................ 65
Phosphorus ......................................... 65

Other analyses .. ............................. 66
Plant Chemical Analyses .............................. 67
Phosphorus .....................................*** 67
Other elements ..................*...............** 68
Fertilizer Analysis ............... ......... 68
Phosphorus ..............................*......... 68
X-Ray diffraction ............................... 68

RESULTS AND DISCUSSION ..................................... 69

Clay Mineralogical Analysis ........................... 69
X-Ray Diffraction ................................... 69
Differential Thermal Analysis ....................... 71
Infrared Analysis .................................... 72
Non-Crystalline Components ......................... 72
Selected Clay Properties ... ... .............. ....... 72

Native Soil Phosphorus ................................ 74

Soil Retention of Applied Phosphorus ................... 75
Phosphorus Adsorption Isotherms ............. .... 75
Phosphate Potential ................ ........... .... 87
Components Contributing to Soil Phosphorus
Retention ......................................... 91
Studies using 32p ...................... ......... 92
Fertilizer Phosphorus Dissolution .................... 102
Fertilizer dissolution without soil ............... 102
Fertilizer dissolution and retention in soil ...... 104

Field Experiments ................................. ... 108
Phosphorus Sources, Rates, and Placements, 1971 ...... 108
Yield analyses .................................. 109
Soil analysis .................................... 112
Fertilizer analysis ............................... 113
Tissue analyses .............. ...... ............. 118
Correlations of yield, soil, and tissue analyses .. 120
Fertilization with Mixtures of Coated and Non-coated
Phosphorus Fertilizers, 1971 ........................ 122
Yield analyses .................................. 122
Soil and tissue analyses .......................... 123
Correlations of yield, soil, and tissue analyses... 124
Phosphorus Sources, Rates, and Placements, 1972 ...... 124
Yield analyses ................. ............... .. 124
Soil analyses ................................. 127
Tissue analyses ............................... 130
Correlations of yield, soil, and tissue analyses .. 133
Various Combinations of Coated and Non-coated
Fertilizer with and without Strip Mulch, 1972 ....... 133
Yield analyses .... .. ........................... 133
Soil analysis .................. ................. 134
Tissue analyses ................................. 136


Correlations of yield, soil, and tissue analyses .. 136
Evaluation in 1973 of Residual Phosphorus from
Previous Two Years ............................... 136
Yield, soil, and tissue analyses .................. 136
Correlations of yields, soil and tissue analyses .. 142
Correlations Between Experiments .................... 145

SUMMARY AND CONCLUSIONS .................................... 148

APPENDIX .......................... ........ .................. 153

LITERATURE CITED ........................................... 281

BIOGRAPHICAL SKETCH ......................................... 293


Abstract of Dissertation Presented to the
Graduate Council of the University of Florida in Partial Fulfillment
of the Requirements for the Degree of Doctor of Philosophy



John J. Nicholaides, III

December, 1973

Chairman: Dr. John G. A. Fiskell
Major Department: Soil Science

Nature of soil-fertilizer reactions and associated corn response

to coated and non-coated concentrated superphosphate (SCSP and CSP,

respectively) were investigated in a Paleudult of high P-retention ca-

pacity. The soil used in the study was primarily Red Bay fine sandy

loam [Rhodic Paleudult]. Included in the comparison of P sources for

field corn were SCSP, CSP, and ordinary superphosphate (OSP). This

phase of research led to an evaluation of these sources on soil-P avail-

ability and on corn responses, first to current and later to residual

rates of application.

Varied laboratory studies dealt with characterization and compo-

nents of P retention in this soil. Removal of organic matter and amor-

phous aluminosilicates had little effect on soil-P retention. However,

removal of crystalline Fe and Al reduced soil-P fixation by 89%. At

various P rates and equilibrium times, P sorption by Red Bay soil was

best described by use of a Gunary rearrangement of the parabolic equa-

tion with inclusion of a square-root term. Langmuir and Freundlich

equations and the phosphate potential were nearly equally satisfactory

for this purpose.

Phosphorus dissolution from CSP was retarded by use of S-coating;

however, observation of wide variance in P released from SCSP pellets was

noted in both laboratory and field studies. Though P dissolution from

SCSP pellets was slower than that from CSP granules, amounts of P dis-

solved from these sources were equal at the end of growing season.

After 10 weeks of equilibration, soil moderated acidity of leachate

from both sources by one-half of a pH unit. Soil sorption of P dissolv-

ing from fertilizer granules (1,340 pg P/g soil) was comparable to P-

sorption maximum described by Langmuir equation (1,560 pg P/g soil) for

P rates up to 5,000 pg P/g soil. Monocalcium phosphate (MCP) in CSP

and SCSP pellets was converted to less soluble P forms, primarily

dicalcium phosphate (DCP), in laboratory and field trials.

Corn yield response to both initially applied and residual P was

alike for CSP and SCSP sources. This indicated that P supplied to plants

either in an immediately soluble form or in small, but continual, amounts

of soluble P were equally effective for corn. However, with various

mixtures of coated and uncoated P sources, maximum tissue-P levels and

yields occurred when mixture contained 20% SCSP. Minimum tissue-P levels

and yields were noted when mixture contained 80% SCSP. A residual study

undertaken 2 years later revealed a reversal of these observations and

indicated that SCSP was more effective as a residual P source than was

CSP. No difference between use of either OSP or CSP as the uncoated

source of P in the mixtures was found. Other residual studies revealed

that for both SCSP and CSP maximum yields were obtained 1 year after

fertilization with 84 kg P/ha, while 168 kg P/ha were required to produce

maximum yields 2 years following fertilization.

Corn yield response was alike between broadcast and band placement

over a range of P rates. Corn yield response, soil-P levels, and tissue-

P levels were linearly related to rate of applied P. Both available

soil-P values and P content of corn ear-leaf tissue sampled at late-

silking stage were linearly and curvilinearly correlated with yield;

highest accounts of yield variation by the above P levels were slightly

greater than 50%.


Soluble phosphates applied to most acid soils with high sesquioxide

contents are known to be rapidly changed into slightly soluble forms.

Attempts at overcoming this poor efficiency or in maintaining sufficient

phosphate for crops can be classed either as studies.dealing with rates

and placement of P fertilizer or in testing of P compounds for evidence

of improved P availability to crops over that obtained by such conven-

tional P fertilizers as concentrated superphosphate (CSP) and ordinary

superphosphate (OSP). The fact that the latter sources contain mono-

calcium phosphate (MCP), which solubilizes upon reaction with the soil

solution, is a contributing factor to the resulting P retention and to

the response of crop yields to P applications. Soil testing procedures,

usually chemical, are used as a measure of plant-available P content

which has resulted from recent fertilization or from residual P levels

present from prior fertilization.

Customarily, P fertilization of row crops in made at planting time,

and hence much of the soluble P reacts with soil components prior to

root development. Slow dissolution of fertilizer P, to give a more even

apportionment of P over the growing season, should allow plant roots

sufficient time to develop and utilize some fertilizer P prior to the

formation of unavailable P compounds in the soil. Slower dissolution

should allow plant roots to intercept soluble P prior to subsequent P

reactions with soil components. Accordingly, to prevent rapid dissolu-

tion of fertilizers, a coating of elemental S with a wax sealant and a

microbicide was developed by the Tennessee Valley Authority for use with

pelleted urea, muriate of potash, and CSP fertilizers.

For an adequate evaluation of S-coated CSP (SCSP), an agriculturally

important soil with a high P-retention capacity was desired. Ultisols

have been noted to be among those Florida soils with the highest P-fixa-

tion capacities (Ballard and Fiskell, 1973). Red Bay fine sandy loam

[Rhodic Paleudult] is such a soil which is extensively cropped in western

Florida, southeastern Alabama, and areas of Mississippi, Georgia, South

Carolina, and North Carolina.

The objectives of the present studies were 1) to characterize P

retention and its contributing factors in Red Bay fsl, 2) to compare SCSP,

CSP, and OSP as P sources for corn in field studies, and 3) to evaluate

the effects of these P treatments on soil-P availability and on composi-

tions of corn tissue.


Factors and Mechanisms Contributing to
Fertilizer Phosphate Retention by Soils

Phosphorus fixation is defined by Kardos (1967) as "the process

whereby readily soluble phosphorus is changed to less soluble forms by

reaction with inorganic or organic components of the soil, with the

result that the phosphorus becomes restricted in its mobility in the

soil and suffers a decrease in its availability to the plant." The

amounts of fertilizer P which an acid soil immobilizes and the rate of

that immobilization have been found to vary greatly depending on the

soil's mineralogical composition, the activity of its exchangeable and

crystalline Al and Fe, its amorphous aluminum hydroxides and iron oxides,

and organic matter-metal cation complexes. Phosphorus fixation, reten-

tion, and immobilization are used interchangeably in the following


Mineralogical Reactions

By using chemically pure soil minerals, < 2i, at 26, 49, and 95C

at pH values of 3, 5, and 7 in the range 0.01 to iM P, Haseman, Brown,

and Whitt (1950) showed the decreasing order in which minerals fixed P

was gibbsite >> goethite > illite > kaolinite > montmorillonite. As

temperature and P concentration increased and pH decreased, the authors

noted that P-fixation rates increased 10-fold in montmorillonite, 8-

fold in illite, and 2-fold in kaolinite. While rates of P retention by

gibbsite and goethite increased slightly with temperature, they observed

that there was little effect on these rates by variation in the pH of

0.iM P additions. In addition to fixing much more P (133 mg P/g gibb-

site) than any of the other minerals, gibbsite also was noted to have a

faster P-retention rate (0.133 mg P/g gibbsite/hour) than for the other

minerals tested. Reaction with all the minerals was described as having

an initial rapid rate of reaction with P during the first 30 minutes

followed by a subsequently slower rate. These two reaction rates were

interpreted by the authors as being indicative of the same general type

of chemical reaction. The rapid fixation,they thought, resulted from P

reaction with readily available Al and Fe, while the slower fixation was

believed to result from P reaction with Al and Fe which were released

through decomposition of the respective minerals.

The decomposition reaction with soil minerals is similar to the P-

induced decomposition of kaolinite reported by Low and Black (1948). In

their work, kaolinite was purported to dissociate into Al and Si ions

with P precipitating the Al, thereby disturbing the equilibrium and

causing the clay to dissolve in accordance with solubility-product prin-

ciples. To support their hypothesis, Low and Black (1950) studied

kaolinite digestion at pH 4.9 for 2 and 3 weeks at 45C and used P solu-

tions ranging from 0 to 3N. They reported the digestion to result in

release of Si in proportion to the P fixed. They noted that the vari-

ance with time on the rate of Si release from kaolinite during P fixa-

tion indicated the existence of two types of reactions. The first and

more rapid reaction was thought to be the surface replacement of the Si

tetrahedra by P tetrahedra, and the second one was the P-induced decom-

position of the surface compound with subsequent precipitation of an


Al-P compound.

Volcanic ash soils containing allophane and halloysite were

treated by Wada (1959) with 1M ammonium phosphate solution at pH 4 and

7. After comparative test periods of 24 hours, it was found that at

pH 4, P was fixed at a reaction rate of 152 mmole P/day/100 g of allo-,

phane and 11 mmole/day/100 g of halloysite. However, he reported that,

after 3 weeks, a total of only 417 and 238 mmole P/100 g was retained

by allophane and halloysite, respectively. He observed that the rate

of reaction of allophane with P was initially quite rapid, while rate

of reaction of halloysite with P was maintained at the initial rate over

the 3-week period. When the pH was increased to 7 in this study, P

reaction with allophane was retarded to a rate of 98 mmole/day/100 g,

and for halloysite the reaction with P more than doubled to 24 mmole/

day/100 g. X-ray diffraction of the P-reaction products in these

samples indicated the presence of ammonium-substituted taranakites.

Also noted was a shift of the 10.1A spacing observed in halloysite at

pH 4 to 13.2A at pH 7 where both cases were reacted with ammonium phos-

phate. Since high P retention was found for halloysite at pH 7, Wada

concluded that the shift to 13.2A was produced by a monolayer of

(NH4)2HP04 being physically adsorbed between the silicate layers in the


Fox, DeDatta, and Sherman (1962) investigated P retention in seven

Hawaiian soils of diverse mineralogical composition, pH, and Al activi-

ties over a 4-day period. To each of three Oxisols, two Inceptisols,

an Ultisol, and a Vertisol were added 44 ppm of 32P. They reported the

general intensity of P fixation for the various mineralogical systems

according to the average percentage of applied P which was fixed. These

percentages were as follows: 99.5% for amorphous hydrate oxides of Fe

and Al > 98.5% for crystalline gibbsite-goethite > 98.0% for kaolinite

> 92.5% for 2:1 lattice layer clays. In a comparison of the relative

effect on P retention of active Al compared to mineralogical composition,

these workers found that even when active Al was reduced to zero by

liming, soils with different clay mineralogy differed as much as 40 times

in P retention. They concluded that the apparent overriding factor in

P retention by these Hawaiian soils was differences in their mineralogy.

Hall and Baker (1971) examined P fixation in montmorillonite and

vermiculite, each having a CEC of 116 meq/100 g. They used Al additions

from 0-90 meq/100 g, P additions in the range from 0-1600 ppm, and a pH

range from 4-7. In this experiment, they found that for all Al and P

rates at pH 4, both clay minerals fixed 67% of the added P, and that

when the pH was increased to 7, the percentage of P retained by mont-

morillonite increased to 82%, whereas that held by vermiculite dropped

to 61%. They also reported that X-ray diffraction of vermiculite at

300C showed stable 14A spacings indicating interlayered stable Al

polymers, the edges of which were postulated to be Al-P reaction sites.

Since montmorillonite, in their studies, did not exhibit these stable

interlayers, the authors contended that the Al-P reaction sites were a

separate solid phase in montmorillonite. Consequently, the authors

postulated that, for soils with clays reacting like the vermiculite of

their study, lime addition prior to P application will reduce P fixation

by reducing the specific surface of reactive Al through stable inter-

layer Al polymer formation, thereby reducing Al surfaces available for

P reactions. Similarly, with soils containing clays having properties

similar to the montmorillonite they studied, liming would increase P


Aluminum and Iron

Numerous workers have confirmed the work of Chang and Jackson (1958)

which contended that, when fertilizer P is added to an acid soil, some

form of calcium phosphate (Ca-P) is the initial reaction product, but

this form soon changes to aluminum phosphate forms (Al-P) and, with

further time, reverts to some form of iron phosphate (Fe-P). Unless

so stated in the ensuing review, the phosphate forms are by definition

those extracted by Chang and Jackson fractionation procedure (1957),

with the Fife modification (1962). By this terminology, Al-P is ob-

tained by extraction with 0.5N NH4F at pH 8.3, Fe-P is found by extrac-

tion with 0.1N NaOH for 24 hours, and Ca-P is that removed subsequently

by 0.5N H2SO4. Yuan, Robertson, and Neller (1960) noted in laboratory

studies that when 50 and 100 ppm P were applied to Red Bay fine sandy

loam [Rhodic Paleudult], 44.6 and 61.2%, respectively, of the applied P

were in the Al-P form, while 49.1 and 34.3% of the applied P, corre-

spondingly, were in the Fe-P form. Using the above soil and another

Paleudult, these workers also noted that as P rates were changed from

0-1000 ppm, the ratio of Al-P: Fe-P was also increased from 4:6 to 7:1.

By adding 200 ppm P to six different Taiwan soils having a pH

range from 5.3 to 7.5 and maintained at field-moisture levels for 3 days,

Chang and Chu (1961) found that 56% of added P was fixed as Al-P, 23%

as Fe-P, and 7% as Ca-P. After 100 days under the same conditions, they

noted Fe-P increased to 43% of the total recovered P, while Al-P and

Ca-P decreased to 44 and 5%, respectively. By adding 200 ppm to each

of these soils and equilibrating the suspensions for 3 hours, the

authors found a net increase of 32, 39, and 4% in the P recovered as

Al-P, Fe-P, and Ca-P, respectively. From these data, they suggested

that the first stage of fixation of added P by various cations occurs

on the solid phase surface with which P comes in contact, and that the

relative amounts and kinds of P formed depended on the specific surface

area of solid phases providing Al, Fe, and Ca for the reactions.

Having based P-fixing capacity of four acid Ohio soils on the

ability of a soil to retain 26 ppm of soluble P against 0.03N NH4F in

0.025N HC1 extraction, Volk and McLean (1963) defined low and high P-

fixing soils as those retaining 25% or less and 50% or more of the

applied P, respectively. These workers found that in soils of low-fixa-

tion capacity, most of the applied P was in the Al-P form, while in

soils with high-P fixation, most of the applied P was found in the Fe-P

form. When a high P-fixing soil was limed, the percentage of the ap-

plied P increased from 30 to 50% as Al-P percentage decreased from 70

to 50%

In a study on Lakeland fine sand [Typic Quartzipsamment] where 800

lbs/acre of triple superphosphate (TSP) were applied in 18 application

with and without lime over a 6-year period, Fiskell and Spencer (1964)

recovered the primary form of applied P as Al-P, with lesser amounts of

dicalcium phosphate dihydrate (DCPD), Fe-P, and octacalcium phosphate

(OCP) or hydroxyapatite (HAp). They found, even with high liming rates,

that 50% of the applied P was recovered as Al-P, while the remainder

was in the form of Fe-P and Ca-P. They reported that both tricalcium

phosphate (TCP) and DCPD which resulted from the TSP applications to

the soil were dissolved by iN NH4OAc at pH 4.8 used as the first step

in the sequential extractions. They stated that use of IN NH4Cl as

recommended in the sequence by Chang and Jackson (1957) might result in

some Ca-P forms being recovered in the Al-P fraction.

After initial applications of 343, 685, and 1,371 kg P/ha, Shelton

and Coleman (1968) analyzed samples taken from a Georgeville soil

[Typic Hapludult] over an 8-year period. This soil was high in kaoli-

nite, with lesser amounts of 2:1 to 2:2 lattice clay minerals, a small

amount of gibbsite and about 5% free-iron oxide. Its initial pH of 5.3

had been increased to 6.0 with liming. Over the 8-year period, they

found that a marked decrease of Al-P occurred, while Fe-P increased.

At the lowest P level, Al-P/Fe-P decreased from 1.4 to 0.2. Reductant

soluble-P, as defined in the method by Chang and Jackson (1957), was

detected after 3 and 5 years at the two higher rates of P fertilization.

This same pattern of Al-P eventual reversion to Fe-P was noted by

Appelt and Schalscha (1970) in volcanic ash soils of Southern Chile. In

evaluating P-retention capacity of 14 Costa Rican soils, Fassbender

(1968) found that young volcanic-ash-derived soils, Andepts or Incepti-

sols, exhibited a larger P-fixation capacity than did the Oxisols, many

of which are Ultisols.

Exchangeable and crystalline aluminum and iron

Coleman, Thorup, and Jackson (1960) investigated P sorption on 60

sub-soil samples of six Typic Hapuludults from the North Carolina

Piedmont. Aluminum exchangeable by 1N KC1 extraction was found to be

correlated (r = 0.838) with P sorption. These researchers found that

salt leaching of the exchangeable Al reduced P sorption in these soils

having appreciable exchangeable Al. However, salt leaching of clayey,

kaolinitic soils, such as Cecil, which contained little exchangeable

Al (< 1 meq/100 g), but large quantities of Fe and Al oxides and

hydroxides, resulted in little decrease in P sorption. These workers

also found P binding in soils to be at a maximum near pH 7 in salt-free

Al systems and near a maximum at pH 4 in presence of Cl salts which

displaced Al from exchange sites. The phosphate-binding reactions,

they suggested, involved exchangeable Al and led to the formation of a

crystalline substance similar to variscite.

A similarly high correlation, r = 0.925, between P sorption and

exchangeable and citric acid-soluble Al was obtained by Franklin and

Reisenauer (1960) on 14 acid soils which they classed as Latosols,

Podzols, Alluvial, Humic-gley, Sierozem, and Prairie soils. They

reported that citrate acid-extractable Al was 160 times more active in

P sorption in these soils than was citrate acid-extractable Fe. In

contrast to the work presented previously on effects of the soil miner-

alogy on P retention, these authors could find no relationship of P

sorption to either the mineralogy or content of clay in any of the

soils used.

In 11 Florida Ultisols, correlations between P retention and amounts

of Al and Fe extracted by various reagents were reported by Yuan and

Breland (1969). They stated that P retention was more highly correlated

(r = 0.96) with citrate- dithionite-bicarbonate (CDB)-extractable Al

plus Fe than with Al and Fe extractable by reagents such as IN_ NH4OAc

at pH 4.8, 0.1N HC1, 0.2M (NH4)2C204 at pH 3.0, or 0.5N NaOH. Both P

retention and extractable Al correlations, they found, were highest

(r = 0.90) for CDB-extractable Al which is purported to extract much of

the crystalline Al. In their studies, P retention and extractable Fe

were correlated (r = 0.92) for Fe extracted by CDB reagent and were

similar to the relationship of P retention and oxalate-extractable Fe

(r = 0.92), wherein the latter reagent is useful in removal of amorphous

Fe forms. They also reported that correlation coefficients between P

retention and Al, Fe, and Al plus Fe extracted by hot 0.5N_ NaOH were

0.81, 0.87, and 0.82, respectively.

Velez and Blue (1971) recorded P sorption in a Rhodic Paleudult

(Red Bay fsl) from Florida and a tropical alluvial Entisol from Costa

Rica that received three treatments: no pretreatment, oxalate extracted,

and CDB extracted. After addition of 5000 ppm P, P sorption in the

Paleudult for the above treatments was 50.1, 47.5, and 27.6% of the

applied P, respectively; the P sorption in the Entisol after the same

treatments was 96.0, 47.4, and 50.6%, respectively. From their data,

they concluded that P sorption in the Paleudult resulted primarily

from precipitation with crystalline Al and Fe, while in the Entisol P

retention was controlled by the amorphorus Al present.

Amorphous aluminum hydroxides and iron oxides

In agreement with studies by Coleman et al. (1960) and Franklin

and Reisenauer (1960), Syers et al. (1971) found P sorption to be cor-

related (r = 0.84) with exchangeable Al in the surface samples of 15

Brazilian soils classified as Ultisols and Oxisols. These soils

retained from 60 to 490 pg P/g from the addition of 500 pg P/g soil.

However, in these soils, removal of exchangeable Al produced only a

small decrease in the amount of P sorbed. This response was quite

similar to that of Typic Hapuludults reported by Coleman et al. (1960)

which also had little exchangeable Al, but large quantities of Fe and

Al present as oxides and hydroxides. Reeve and Sumner (1970) could

find no apparent relationship between exchangeable Al and P-fixation

capacity of eight Oxisols in Natal. Since in these studies these

workers did not determine the Al and Fe oxides and hydroxides, no

comparison could be made between their work and that of Coleman et al.


Using 10 Podzols from Scotland, Williams (1960) reported that acid

oxalate-extractable Al was highly correlated (r = 0.88) with P sorption.

Oxalate-extractable Fe also was correlated (r = 0.68) with P retention,

but addition of a term accounting for Fe in the equation never did

improve the estimation of P sorption accounted for by Al alone.

Bromfield (1964) used biological reduction proposed by Bromfield and

Williams (1963) and 0.051N HC1 to remove Fe and Al from three acid

Scottish soils with P-retention capacities varying from 3.8 to 7.5 mmole

P/100 g soil. In his investigations, as Al was left unchanged, P sorp-

tion was observed to decrease slightly by 1.1 mmole as the first 11

mmole Fe/100 g soil were removed; as the next 10 mmole Fe were removed,

P sorption was found to remain constant. However, as Al was removed

from the same soils by acid leachings, P sorption was decreased to as

much as 60% of the initial value, regardless of whether Fe was removed

- or 4eft intact.

For the majority of 47 acid soils from Australia, Bromfield (1965)

found Al extracted by ammonium oxalate at pH 3.2 to correlate better

(r = 0.98) with P sorption than oxalate-extractable Fe (r = 0.79).

However, he stated that the relative contributions of the amorphous Al

and Fe to P sorption could not be accurately assessed due to the posi-

tive correlations (r = 0.79) between extractable Al and Fe.

Using 12 acid soils of varying Fe oxide contents, including the Ai

and B2 horizons of Cecil [Typic Hapludult], Ramulu, Pratt, and Page

(1967) noticed P retention was significantly correlated to both CDB-

extractable Fe (r = 0.77) and oxalate-extractable Fe (r = 0.95). These

authors showed that, at the same level of CDB-extractable Fe, soils with

only kaolinite as the dominant clay mineral fixed more P than soils with

both kaolinite and vermiculite. The inference drawn from their work was

that there was a finite amount of Fe present in those soils with vermic-

ulite that was CDB-extractable, but which was not active in P retention.

They postulated that Fe-hydroxy polymers form in the interlayer spaces

of vermiculite and that these Fe polymers were not active in P retention,

but were CDB-extractable. Since they did not find that oxalate extrac-

tion removed Fe from vermiculite interlayers, it was concluded that

oxalate provided a more quantitative means of determining the relation-

ship between Fe oxides and P retention than did CDB treatment. However,

in the Cecil Ai sample which contained gibbsite and had an oxalate-Fe/

CDB-Fe ratio of 0.4, it was observed that more P (5.86 mg/g) was fixed

than in the other soils with ratios of twice this magnitude. However,

Ramulu et al. (1967) explained such anomolous behavior by stating that

perhaps some crystalline Fe of this soil entered into the P-retention


Another probable occurrence is that the gibbsite was responsible

for the additional P retention (Haseman et al., 1950). Syers et al.

(1971) found CDB also removed almost equal quantities of Al (1.1 to 19

mmole/100 g) and Fe (1.4 to 41 mmole/100 g) from the 15 Brazilian

Oxisols and Ultisols studied. Although Ramulu et al. (1967) had cor-

rected for iN KCl-extractable Al in their correlations between Fe and

P retention, they felt that in the presence of high amounts of free-iron

oxides, it was reasonable to assume that fixed P was mainly due to the

iron oxides. However, the contributions to P retention by Al which

was CDB-extractable were essentially ignored by Ramulu et al. (1967).

Syers et al. (1971) reported that oxalate removal of amorphous Fe

and Al components from the Brazilian Oxisols and Ultisols, when compared

to untreated samples, reduced P sorption by 19 to 100, while CDB

removal of Fe and Al changed P sorption in a range from 12% increase to

84% decrease. Their observation was that the amount of the added P

which was sorbed by untreated samples was better correlated with oxalate-

Al (r = 0.78) than with oxalate Fe (r = 0.47). These workers also

suggested that in P sorption by these untreated soils, amorphous Al

components were more active per unit weight than were amorphous Fe


The results of these studies by Ramulu et al. (1967) and Syers et

al. (1971) give credence to the contention of Hsu (1965) that surface-

reactive amorphous aluminum hydroxides and iron oxides, rather than

crystalline Al and Fe, dominate P retention in soils. Hsu (1964)

reasoned that since the activity of A13+ or Fe3* in solution is limited

by pH (being negligible above pH 5), and since no such limitation is

imposed on the surface reactivity of amorphous aluminum hydroxides and

iron oxides, evidence for P sorption being a surface reaction rather

than precipitation could be obtained by adding dilute P solutions with

pH adjusted to 7. Accordingly, he found that when solutions containing

0-60 ppm P at pH 7 were added to a slightly acid Elstow clay loam

(pH 6.4), two fixation reactions operated at different rates. This

author postulated that the more rapid reaction was due to surface ad-

sorption of P on amorphous aluminum hydroxides and iron oxides present

in the soil, while the much slower reaction was believed to be due to

a similar surface adsorption on secondary amorphous hydroxide and oxide

surfaces that developed during ageing for one year. In a hypothetical

consideration of P retention work, Hsu (1965, p. 403) concluded:

1. In principle, precipitation and adsorption
result from the same chemical force. Whether
precipitation or adsorption occurs is dependent
on the form of aluminum and/or iron present
at the moment of reaction. In common soils,
because of the effect of pH, surface-reactive
amorphous aluminum hydroxides and iron oxides
dominate the process of phosphate fixation rather
than A13+ and Fe3+ in solution.

2. The concentration of phosphate in solu-
tion is determined by the sum of the activities
of all forms of aluminum and iron; thus, the
study of the development of the activity of
phosphate-fixing cations, particularly amor-
phous aluminum hydroxides and iron oxides, is
a more fundamental approach than is the study
of the solubility of phosphate compounds.

The above conclusions at least partially rejected the phosphate

phase equilibria study of Lindsay and Moreno (1960) in which activity

isotherms for the variscite, A1P04*2H20, and strengite, FePO4*2H20,

among others, were represented on a solubility diagram. A function of

the phosphate activity in solution was plotted against pH on this

solubility diagram. By assuming the respective A1l and Fe+ activities

in soils to be limited by the solubilities of gibbsite and goethite,

the authors were able to assess the relative stabilities of variscite

and strengite and to predict their transformations in soils after

fertilizer or lime applications.

However, Hsu (1965) did not take into account the finding of Lindsay,

Frazier, and Stephenson (1962) that in the microenvironment around a

particle of monocalcium phosphate (MCP) derived from fertilizer, such

as concentrated superphosphate (CSP), an extremely low pH of 1.5 occurred

within an hour after application to the soil. The resulting metastable

triplepoint solution(MTS$ was defined by the authors as the solution

saturated with respect to Ca(H2P04)2*H20 and CaHPO4*2H20. They noted

that after 15 minutes, MTSP had dissolved from the contacts with

Hartsells fsl [Typic Hapludultj 40 and 138 mmole/liter of Fe and Al,

respectively. After 3 days, they noted that the pH of the filtrate had

risen from 1.8 to 2.1, and that the Fe and Al compositions of that

filtrate were 53 and 322 mmole/liter, respectively. Filtrates obtained

during the 3-day reaction period, upon standing, yielded precipitates

identified as colloidal ferric aluminum phosphate, (Fe,A1,X)PO4 nH20

of indefinite composition, where X is any cation other than Fe and Al,

and n is a variable. At the low pH present in the microenvironment of

the P fertilizer, Fe3 and Al1+ were active, according to Lindsay and

Moreno (1960), and came into solution, thereby precipitating P. This

might be contrasted with the laboratory study of Hsu (1964) in which P

added at pH of 7 to soil was retained in a situation less likely to

simulate actual field conditions which made it appear that amorphous

aluminum hydroxides and iron oxides were responsible for P retention.

Under field conditions, A3+ and Fe3+ resulting from fertilization

acidity could also enter into P sorption by soils.

Organic Matter Complexes

As previously presented, Williams (1960) found P retention in 10

Podzols from Scotland to depend primarily on oxalate-Al. In all soils

in these tests, he found P retention to be highly correlated with easily

oxidizable soil C (r = 0.77-0.89). After finding high correlations

between this C and Al (r = 0.54-0.89), and to a lower degree between C

and Fe (r = 0.15-0.59), the author concluded that humate-Al and possibly

humate-Fe complexes also entered into P retention of the soils. Weir

and Soper (1963) studied P interaction with humic acid (HA)-Fe complexes.

They found that, while ferric iron usually precipitated as ferric

hydroxide above pH 2.3, in the presence of HA considerable amounts of

Fe were active in solution. They demonstrated that this Fe completed

with HA and sorbed considerable added P. When the authors subjected

the HA-Fe-P complex to shaking with a hydroxide-exchange resin, the

complex was able to retain 38% of the P. From 90 to 100% of this P was

exchangeable with 32p.

In 24 acid New Brunswick soils, Saini and MacLean (1965) found that

organic matter (OM), as determined by the method of Walkley-Black

(1934), was related to P retention by a correlation coefficient of 0.60.

Also, Al extracted at 50C with sodium hydrosulphite in citrate buffer

at pH 4.75 had been found to be highly correlated (r = 0.91) with P re-

tention and with OM content (r = 0.79) in these soils. The authors

suggested that part, but not all, of the active Al in these soils was

associated with OM in the form of P-retaining complexes.

On a wide variety of 50 New Zealand topsoils ranging in their clas-

sification from Brown-Gray Earths to Podzols, Saunders (1965) found

soil OM was not directly concerned with P retention; however, a close

correlation (r = 0.70) existed between OM and oxalate-extractable Al

and Fe. The author postulated that the OM-Al-Fe complex was a gel

complex which could sorb P. It was probable, he stated, that the amor-

phous Al and Fe extracted by the oxalate were derived from this gel and

that this gel was the major site for P retention in the soils he inves-


Fox and Kamprath (1971) found added P to be readily leached from a

muck soil possessing only 0.02 meq/100 g of exchangeable Al. At the

low P rate of 25 kg/ha, all of the applied P was removed from this muck.

At a rate of 100 kg P/ha, the small amount of P not recovered by leachings

was considered to be held by.metal cations completed by the OM.

Measurement of Phosphate Availability in Soils

Although applied P is retained by various Al and Fe compounds in

acid soils, some of this fixed P has proven to be available to plants.

Plant availability of soil P has generally been dependent on the form

in which P is held in the soil. The availability of P from synthetic

amorphous and crystalline Al-P compounds was investigated by Juo and

Ellis (1968) using suidangrass as the indicator crop in quartz sand

cultures. Their study indicated that relative availability of P from

the compounds studied was strengite < variscite <<< colloidal Fe-P Z

colloidal Al-P. The authors found surface area of the above compounds

in the order listed to be 1.98, 24.9, 27.5, and 10.5 m2/g, respectively.

However, it was noted that crystallization of the colloidal Fe-P to

strengite was essentially complete after 12 hours of digestion at 105C,

while crystallization of amorphous Al-P to variscite required 40 days

of digestion at this temperature. At 35C, amorphous Fe-P was observed

to crystallize within 9 months, while Al-P was found to remain amorphous

over this period. These results indicated to the authors that degree

of crystallinity seemed to be more important than surface area in deter-

mining relative availability of Al-P and Fe-P compounds.

The above results were in agreement with previous work by Taylor,

Gurney, and Lindsay (1960), Lindsay and Dement (1961), and Taylor et al.

(1963). These three greenhouse experiments compared the availability

to corn of various Al-P and Fe-P compounds on a Hartsells fsl [Typic

Hapludult] using monocalcium phosphate (MCP) as a reference P source.


Total P uptake by three successive corn crops from colloidal Al-P was

found by Taylor et al. (1960) to be much greater than that from colloi-

dal Fe-P at a soil of pH 4.8. These authors observed during the three

successive crops that P uptake from colloidal Al-P was essentially

unchanged, while that from colloidal Fe-P decreased drastically after

the first crop. These findings were interpreted by Juo and Ellis (1968)

to be due to Al-P remaining colloidal during the cropping period,while

the colloidal Fe-P crystallized rapidly. Although strengite was found

to be completely unavailable in Hartsells fsl at pH 4.9, colloidal Fe-P

was 78% as available as was MCP (Lindsay and Dement, 1961). With liming

to pH 6.5, colloidal Fe-P was noted to be as available as MCP. In

addition to colloidal Al-P and Fe-P compounds, calcium ferric phosphate

H4CaFe2(P04)4 5H20), potassium taranakite (H6K3A15(P04)8 18H20), and
ammonium taranakite, at rates of 200 and 1000 mg P/pot, were found to

be relatively available P sources for corn when applied to the above

acid Hapludult (Taylor et al., 1960; Taylor et al., 1963).

On acid soils, the Al-P fraction, as determined by 0.5N NH4F, at

pH 8.3, has been reported by many workers to be more closely correlated

with plant-available P than was the Fe-P fraction removed by 0.1N NaOH

extraction. Smith (1965) observed that Al-P was the main source of P

available to wheat grown in a slightly acid Red-Brown Earth. However,

in the absence of wheat, this Al-P which was taken up by wheat in other

pots was found to be converted to Fe-P compounds. From this he concluded

that the plant probably was taking up some soluble P from the Al-P

fraction which in the absence of plant roots would be precipitated as

Fe-P. Payne and Hanna (1965) found that P content of millet tops grown

on three acid Podzols of low, medium, and high P-retention capacity was

chiefly correlated with the Al-P fraction (r = 0.56). Using 12 soils

varying from 28 to 72% fixation of soluble P against 0. 03N NH4F in

0.1i HC1, Lavery and McLean (1961) found the lower P-fixing soil tied

up 2.5 times as much P in the Al-P fraction as in the Fe-P fraction. In

their work, higher P-fixing soils were observed to have twice as much P

in the Fe-P fraction as in the Al-P fraction. These workers stated

that these results suggested P from the Al-P fraction would be more

plant available than that from Fe-P fractions.

Indices for Phosphate Availability

Phosphorus availability and retention in soils have been measured

by a variety of methods. Foremost among these is the use of chemical

extractants designed to indicate levels of P available to plants. Since

most work on acid soils indicates Al-P to be the fraction most closely

correlated with plant growth, a selective extractant for primarily Al-P

would seem to be desirable. Extractants which have been used to esti-

mate plant-available P are given in Table 1.

For purpose of brevity, the methods listed in Table 1 will be

referred to as reagents 1-12 and, unless otherwise stated, the first

soil to solution ratio and shaking time are understood to have been

used. Comparing P extracted from 74 U.S. soils by reagents 1, 2, 5,

and 9, Fitts (1956) reported reagent 1 to be best for determining

available P (r = 0.70) and least affected by soil pH or cation exchange

capacity to anion-exchange capacity ratio. When the soils were limited

to those with pH less than 6.1, the correlation coefficient for reagent

1 increased to 0.80. Similar results were found by C. F. Griffin (1971)

after 13 acid Connecticut soils were limed. In this work, extractable

P was determined by reagents 1, 5 (15 minutes' shaking time) and 9 for




C )



9O .













,-4 -




o H

C) 4
'- 0

E 0

o 0
o C:
9 ^

0 0

0 00


dcO 0

O$ 0


4, -^ to
-o U %

- -o 0

C, 0 U h
0d 0\ 0

x A n mf rz .0
o fo T- ( o > o ir\



0 0 0 0 \ 0 0 0
S-1 0\ OS'





o U

0 ,r'

0 0 0 0 0 0
CN V V) -4Ul s

N. 0 CC 0 C CI 0_ ( v 0


0 N0

w-c 0 + 0'
ou ) Ot r
00 0 0 0 F0
C vC\ (I) 0
w N 0 0 i
0. 0 w I

-. 1 N 1 01
Scud Zic;

NC 3 \0 C- OD ON 0o V-4 NC
V-4 V-4 1-

-4 V1 -i v\



each soil at five liming levels, giving a pH range from 5.3 to 7.7. It

was found that changes in extractable P by reagent 1 were less affected

by soil differences than P extracted by the other methods. On virgin

Red Bay fsl [Rhodic Paleudult] Robertson and Hutton (1953) found that

reagent 3 (1:40 soil to solution ratio) removed an amount of P represen-

tative of plant-available P. However, they report that when residual P

exceeded 20 and 80 kg P/ha, reagents 1 and 2 gave a better measure of

available P than other methods tested. In a detailed comparison of

reagents 1-7 at all the soil to solution ratios and shaking times listed

in Table 1, Sherrell (1970) deemed reagents 1 and 2 at a 1:10 soil to

solution ratio to be the most rapid, practical, and accurate for availa-

ble P analysis of 16 varied New Zealand soils. Minor alterations in

time of shaking were shown in this study to have little effect on P

extracted by reagents 1 and 2. Phosphorus extracted by reagent 1 was

found by Sherrell (1970) and Payne and Hanna (1965) to be significantly

correlated for both extraction times with the Al-P fractions in the

soils, while P extracted by reagent 2 was correlated with Ca-P. However,

on five Minnesota soils with pH from 5.9 to 7.5 and textures ranging

from loamy sand to clay, reagents 1 and 2 were reported by Chai and

Caldwell (1959) to be very closely correlated with Al-P (r = 0.96 and

0.98) and with Fe-P (r = 0.93 and 0.89), but not with Ca-P fractions as

determined by the method of Chang and Jackson (1957). Phosphorus

extracted by reagent 2 was noted by Sherrell (1970) to be correlated

with white clover yield after P fertilization (r = 0.89), with yield

without P additions (r = 0.87) and with P uptake (r = 0.79). In an

evaluation of P extracted by reagents 1, 2 (both 1:7 soil to solution

ratio), 3, 4 (1:25 ratio), 5 (15 minutes), 6, and 8 from acid, volcanic-

derived soils of the Windward Islands, Walmsley, Twyford, and Cornforth

(1971) reported that results by reagent 2 correctly predicted banana-

yield response to P fertilization for 85% of the tests.

However, the reagents containing NH4F (reagents 1, 2, and 12) have

been shown to dissolve compounds other than Al-P. Bromfield (1967)

reported that another researcher found DCP and TCP to be "extensively

dissolved" by reagent 12, although no data were given. Extraction with

reagent 12 was reported by Bromfield (1967) to release primarily Al-P

and variable amounts of Fe-P and Ca-P from 14 New Zealand acid surface

soils having wide P-sorption capacities and Fe and Al contents. Of the

P in the various compounds, he noted that reagent 12 dissolved 96% of

DCP, 88% of TCP, and 23% of the P residue from water-leached SP. Since

appreciable amounts of DCP and anhydrous DCP were shown to be formed in

the vicinity of a SP granule (Lindsay et al., 1962), Bromfield (1967)

suggested that reagent 12 can clearly overestimate the Al-P fraction in

a fertilized soil. Blue (1970) used reagent 12 at pH 8.2 and found 14

and 53% of P in TCP were dissolved after extraction times of 1 and 96

hours, respectively. In the above study, reagent 11 was noted to dis-

solve 34% P from TCP after 1 hour.

On 29 acid to neutral surface soils from California, Pratt and

Garber (1964) found that P extracted by reagents 2 and 11 were correlated

(r = 0.82), as were those of reagents 2 and 11 (r = 0.65). Using four

Mollisols and two Alfisols which differed in pH and P fertilization,

Tandon, Motto, and Kurtz (1967) compared elements extracted by reagent

1 and those extracted by each of its components. In the acid soils of

their study, they found that P extracted by reagent 1 exceeded the sum

of P removed separately by each component. Neither P extracted by

reagent 1 nor by 0.03N NH4F was correlated with soil pH, Al, Fe, Ca,

Mg, Mn, or Si levels. Reagent 1 was found by these workers to extract

an average of 21 ppm P, 850 ppm Al, 575 ppm Ca, 200 ppm Mg, and 65 ppm

Mn from these soils. However, when these soils were first treated with

reagent 12, the average amounts of P, Al, Ca, Mg, and Mn subsequently

extracted by reagent 1 were 7, 14, 25, 20, and 15 ppm, respectively.

Phosphate Potential

Another index of soil-P availability used in the literature is

phosphate potential, defined as pCa + pH2PO4 by Aslyng (1954) and

Schofield (1955). Phosphate potential, reported in some literature as

Potsp and termed herein as PP, is a measure of the ease of P removal

from soil. Quantity (Q) and intensity (I) were defined by Williams (1967)

as the amounts of P sorbed by a soil and the P concentration remaining

in solution, respectively. In his study, the rate at which P moved

from sorbed (Q) to soluble (I) form was denoted capacity (C). Beckett

and White (1964) termed the Q/I relation as 6Q/6I and called it poten-

tial buffering capacity, PCB. Such definitions are consistent with those

employed herein.

Ramamoorthy and Subramanian (1960) compared correlations of crop

response with P removed by reagents 1 and 4, P sorption, desorption and

anion-exchange capacities, PP, and their own equilibrium PP (EPP),

using a wide variety of paddy Indian soils which varied in pH from 5.2

to 7.4. Equilibrium PP was obtained by equilibrating soil samples with

small increments of MCP solution and determining at which level of MCP

was PP found to remain constant. A higher correlation for crop responses

to P fertilization on the 10 soils was found with EPP (r = 0.77) than

with any other method tested. Values for PP by the Schofield method

were found to range from 6.87 to 8.42, while those for EPP ranged from

7.43 to 8.15. The relationship between PP values and P content of

cotton plants from 13 fields of a red clay loam in Uganda was reported

by Le Mare (1960) to be very close (r = 0.88). The range of values for

pCa + pH2PO4 in this samples was 6.59 to 8.15. Using 26 Rhodesian and

4 Zambian soils which were derived from acid and basic igneous and

sedimentary rocks, Salmon (1966) found Q/I in these acid soils to be

better correlated with percent P taken up by grass (r = 0.93) than was

either Q (r = 0.84) or I (r = 0.89). The Q/I relationship was reported

to have accounted for 87% of the variation in P content of grass in this


Taylor and Gurney (1965c) added CaCO3, sufficient to provide from

0 to 200% of the CEC, to 5 acid Ultisols and examined the resulting PP

and resin-extractable P values. They reported that, while PP showed

little change on addition of excess CaCO3, amounts of resin-extractable

P increased with pH. The authors inferred from the results that a

greater quantity of P was adsorbed in a more readily removable form in

limed soil than in unlimed soil. White and Haydock (1967) evaluated P

availability as measured by PP against that by reagents 3, 4, and 5 on

15 soils covering a range of P availabilities which they classified as

deficient to adequate. By adding CaCO3 where required, the pH of each

soil was raised to a range of 5.5-6.0. Using Phaseolus lathyroides L.

as an indicator crop on acid soils of variable Ca status and with high

Al, these authors reported no difference between any of the methods

tested. All methods were reported to be highly correlated with PP on

the acid soils.

Phosphorus Sorption Isotherms

The equilibrium in soils between sorbed and solution P also has

been described through sorption isotherms. Such isotherms have been

reported by Fox and Kamprath (1970) to include both the I and C values

as defined by Williams (1967) and described in an earlier section.

Employed in soil-P studies have been both the Freundlich and the

Langmuir sorption isotherms.

The Freundlich equation is

x/m = kcn [1]

where x is ng of P sorbed, m is adsorbent in g, c is the solution, P

concentration in jg/ml, and k and n are constants. Sorption of P by

silt loams and silty clay loams from Illinois with pH range from 4.8 to

6.0 was shown by Kurtz, DeTurk, and Bray (1946) to follow a Freundlich-

type isotherm. Russell and Low (1954) found that P sorption on kaoli-

nite also obeyed the Freundlich equation. Olsen and Watanabe (1957)

reported that the Freundlich equation generally applies to a wider range

of equilibrium P concentrations and for larger amounts of sorbed P than

does the Langmuir described below.

The primary advantage of the Langmuir equation over the Freundlich

is that an adsorption maximum can be calculated from the Langmuir equa-

tion. The linear form of the Langmuir equation is
S = 1/kb/b+ [2]

where c, x, and m are as defined in equation [1], b is the sorption

maximum, and k is a constant. However, Olsen and Watanabe (1957)

reported that in a 24-hour equilibration period with various acid Ulti-

sols, the P concentration could not exceed 9 x 10 ii for the linear

relationship to exist.

Compared with Freundlich and Langmuir equations, better estimations

of P retention by 24 acid soils was found by use of a rearrangement of

the least squares parabolic equation (Gunary, 1970). By inclusion of

a square-root term, the equation is

S= A + D + Be [3]

where c, x, and m are as described in equation [1] and A, D, and B are


Recently, two, and possibly three, kinds of adsorption sites with

differing affinities for P have been postulated by Syers et al. (1973).

These workers used the Lanimuir equation in studies of sorption of added

P by Brazilian Oxisols and Ultisols of low, moderate, and high ability

to sorb P. Plotting the data according to the conventional Langmuir

equation, they found two linear relationships. These were interpreted

as P sorption by sites differing widely in P affinities. By rearranging

the Langmuir equation to the form

x/m = b (x/m)/(kc) [4]

where the terms are as defined in equations [1] and [2], they found that

the P-sorption curve could not be satisfactorily resolved into two

straight-line components. From this result, they contended the ex-

istence of a third type of site for P sorption was possible. However,

previous work by Kurtz and Quirk (1965) showed that small or no reduc-

tions in P-sorption maxima resulted after large P additions to acid Red-

Brown Earths and Lateritic Podzols. They suggested that P did not

permanently occupy the initial adsorption sites, but with time formed

discrete P minerals.

In Typic and Rhodic Paleudults, Langmuir-defined P-sorption maxima

and surface areas were closely related (r = 0.96) in studies by Olsen

and Watanable (1957). They suggested that surface adsorption of P is

likely to occur in the initial adsorption stages. In acid Oxisols

with various P fertilization histories, quantitative relationships

between Al-P, Fe-P, or Ca-P and Langmuir adsorption isotherms were not

found by Kurtz and Quirk (1965). They concluded from this that sorp-

tion maxima were not related to any particular P fraction in the soil.

In a study by Woodruff and Kamprath (1965), the P-sorption maximum

of a Typic Paleudult had to be completely saturated to obtain maximum

yields in comparison with P sorption of three Typic Hapludults which

had more clay but did not require that their P-sorption maxima be

completely saturated to obtain such yields. Olsen and Watanabe (1963),

stated that diffusion possibly becomes a limiting process in P uptake

more rapidly in sandy soils than in fine-textured soils. Therefore,

Woodruff and Kamprath (1965) concluded that a sandy soil requires a

much higher equilibrium solution P concentration to supply adequate

amounts of P for plant growth than does a fine-textured soil.

Efficiency of Phosphate Fertilizers

Phosphate Fertilizers

Conventional P fertilizers such as OSP and CSP were reported by

Hauck and Koshino (1971) to result in 5-25.' recovery of applied P in

the crop. This incomplete recovery of applied P fertilizers was attri-

buted by the authors to rapid dissolution and consequent release of

high concentrations of P primarily in unavailable forms.

To control the dissolution rates of P fertilizers, two primary

approaches have been followed: 1) use of compounds with limited water

solubility, such as rock phosphate (RP) (Ensminger, Pearson, and


Armiger, 1967), acidulated RP (McLean and Logan, 1970), or ammonium

phosphate (AP) (Webb and Pesek, 1958), and 2) modifying soluble mate-

rials, primarily by various coatings to slow their nutrient release to

the soil solution (Army, 1963). Nearly 40 such coatings have been

patented in the United States by O'Conner (1965), Formaini (1967),

Bidlack and Bidlack (1967), Blouin and Rindt (1967), and Goodale and

Frump (1968).

One of the coatings developed by Blouin and Rindt (1967) at the

Tennessee Valley Authority is an essentially impervious coat of ele-

mental S with a microcrystalline soft wax sealant, and microbicide,

usually coal tar. They reported that the sealant slowed water-vapor

transfer through the S coating, while the coal tar retarded microbial

decomposition of that sealant and possibly affected transformation of

the S coating. When used with CSP, this product is termed SCSP and has

various release characteristics. The P release through this coating

was stated by the developers to be dependent on coating thickness, in-

tactness of sealant, microbicide effectiveness, soil incorporation,

soil temperature, and soil water content. They found that dissolution

rate of P from the coated pellet was substantially slower than that of

an uncoated pellet, subsequently allowing a more even apportionment of

P over the growing season of crops.

Secondary Forms of Phosphate Fertilizers in Soil

When fertilizer containing principally Ca(H2PO4)2.H20 (MCP) was

applied to acid soils, Lindsay, Lehr, and Stephenson (1959) reported

that water vapor was adsorbed by the pellet establishing a wetting zone

and allowing water movement from the pellet to the soil, and due to the

low pH of 1.5 of the escaping solutions, appreciable quantities of Al,

Fe, Mn, K, and Ca were dissolved from the soil. After addition of MCP

fertilizers to a wide variety of Florida soils, among them Typic and

Rhodic Paleudults, Fiskell (1965) also reported dissolution of large

quantities of Al, Fe, and Mn. In the ionic environment of the ferti-

lizer pellet, a wide range of secondary P compounds were noted by Lindsay

et al. (1962) to be precipitated from solution. Identification of these

compounds and those in subsequent studies by other authors has employed

the use of X-ray diffraction. Included among identified reaction

products between water-soluble orthophosphate fertilizers and soils have

been dicalcium phosphates (DCP), CaHPO4'2H20 and CaHPO4 (Brown and Lehr,

1959; Lehr, Brown and Brown, 1959; Lindsay et al., 1962; Das and Datta,
1968; Bell and Black, 1970b), octocalcium phosphate (OCP), CagH2(P04)6

5H20 (Lehr and Brown, 1958; Lindsay and Taylor, 1960; Lindsay et al.,

1962; Bell and Black, 1970a, 1970b), hydroxyapatite (HAp), Ca10(OH)2(P04)6

(Lindsay and Taylor 1960; Murrmann and Peech, 1968), and fluoapatite

(FAp), Ca0F2(PO4)6 (MacIntire and Hatcher, 1942; Lindsay and Taylor,
1960; Murrmann and Peech, 1968). These transformations are considered

significant in that the solubility and availability of these phosphates

have been reported by Lindsay and Taylor (1960) to decrease in the same

order as their appearance in the soil, i. e.,

DCP > OCP > HAp > Fap.

In addition to crystalline secondary P fertilizer compounds,

amorphous Al-P and Fe-P compounds were always present as reaction prod-

ucts of P fertilizer and soil (Lindsay and Taylor, 1960; Lindsay et al.,

1962; Taylor, Gurney, and Moreno, 1964; Das and Datta, 1968). As dis-

cussed earlier, Juo and Ellis (1968) have found these amorphous Al-P and

Fe-P compounds to be relatively plant available.

The presence or addition of other soil compounds or fertilizers

such as KC1, IiH1103, or K2S04 near Ca(H2P04)2'H20 has been found to so

complicate the dissolution process that reaction products cannot be

described in a simple system (Taylor and Gurney, 1965b). Among the

most common compounds known to form in such situations are calcium ferric

phosphates, calcium aluminum phosphates (Lindsay and Taylor, 1960;

Lindsay et al., 1962; Taylor et al., 1964;Das and Datta, 1968), and

potassium and ammonium taranakites (Lindsay and Taylor, 1960; Lindsay

et al., 1962; Taylor and Gurney, 1965a, 1965b). Also present have been

found gypsum or related compounds (Bouldin, Lehr, and Sample, 1960;

Taylor and Gurney, 1965b).

Solid phases remaining in the residue of 60 to 80 mesh MCP granules

after 3 weeks of reaction time in a Typic Hapludult were identified by

Bouldin et al. (1960) to be almost exclusively dicalcium phosphate di-

hydrate (DCPD) with traces of anhydrous dicalcium phosphate (DCPA).

The fraction of added P remaining as DCPD was predicted by the authors

to be 21.7%, while the observed fraction was found to be 20.7%.

Allmaras and Black (1962) applied 32 to 60 mesh CSP in which P was

present almost entirely as orthophosphate to 14 different soils, includ-

ing two acid Paleudults. After a 2-week growth of sorghum in the

greenhouse, the authors found only 14% of the applied CSP was recovered

from the residual granules.

Plant Responses to Phosphate Fertilization

Field Response

Fertilizer placement

Corn yields increased with change from low to high water solubility

of banded P fertilizers applied to acid soils low in available P

(Webb and Pesek, 1958). On twenty field sites in Iowa from 1951 to

1956, these workers used CSP, among other fertilizers, with the fraction

of available P in water-soluble form ranging from 10-100%. With in-

creasing water solubility of phosphates applied at 13.4 kg P/ha on a

Floyd silt loam in 1955, corn yields were reported by the authors to

increase from 3519 to 4416 kg/ha. They concluded that for satisfactory

yield responses on the Iowa soils used, banded fertilizers should not

contain less than 60% of their P in water-soluble form. Similar evalu-

ation of corn yields as related to increasing water solubility was made

using broadcast P fertilizers, applied at 9-36 kg of available P/ha to

16 of the soils used in the previous study (Webb and Pesek, 1959). The

percentage of water-soluble P was found by the authors to be of no im-

portance in determining the effectiveness of fertilizers broadcast and

plowed under for corn. The differing responses of these two studies

were attributed by the authors to the fact that a minimum of soil-fer-

tilizer reaction occurred when a fertilizer was banded in comparison to

the mixing from incorporation of broadcast fertilizers. Therefore,

they stated that the effectiveness of the banded fertilizer was probably

determined to a large extent by the fertilizer characteristics rather

than by the nature of the soil-fertilizer reactions, while the opposite

was true of broadcast fertilizer.

The relative effectiveness of banded and broadcast P fertilizers

on three acid soils, among them a Typic Hapludult, and two alkaline

soils was investigated in the greenhouse by Terman, Bouldin, and Lehr

(1958). On the acid soils which were low in available P (< 1.5 ppm

extractable by reagent 4), P fertilizer applied in bands produced higher

yields of ryegrass and sudangrass dry matter than did broadcast and

incorporated fertilizer.

Residual effects

The effect on corn yields of large initial P applications followed

by smaller annual applications to a Rhodic Paleudult was investigated

by Robertson, Thompson, and Hutton (1966) for a 4-year period. Initial

applications of CSP were 123, 246, 469, and 592 kg P/ha; annual amend-

ments as OSP were 15 and 29 kg P/ha. Since the authors reported no

significant differences in corn yields between rates over the time period,

it was evident that the lowest P rate of 123 kg/ha was as efficient as

the higher rates in satisfying the demands of the crop. Phosphate at

0, 336, 672, and 1,344 kg P/ha was applied by Younge and Plucknett (1966)

to quench the P-fixation complex of a Hawaiian Typic Gibbsihumox. The

field experiment indicated that only the highest P rate supplied suffi-

cient P to produce adequate yields of pangolagrass-legume pasture over

the 6-year test period.

To a Typic Hapludult with a high P-fixation capacity of 719 kg P/ha,

Kamprath (1967) had added CSP in rates of 0, 177, 343 and 685 kg P/ha in

1956. The initial application of the highest P rate was reported by

the author to supply adequate P for growth in 1964. When 25 kg P/ha

were added to plots which had received the two intermediate rates, the

total P supply was noted to produce high corn yields. The author con-

cluded that the P added initially was not lost irreversibly, but was

available for plant growth in subsequent years. A similar such result

was reported by Fox, Plucknett and Whitney (1968). These workers

measured the efficiency of P applications made 9 years earlier to a

Gibbsihumox in Hawaii and concluded that 64 to 80% of the applied P could

be used by future crops grown on this soil. The residual effectiveness

of 10-year-old P applications to a Typic Hapludult was judged by Fox and

Kamprath (1970) to be 28-50%. They contended that this lower efficiency,

as compared to that reported by Fox et al. (1968), was due to the fact

that this Hapludult had been cropped annually with wheat or corn since

it was last fertilized, while the Gibbsihumox, with its grass-legume

mixture, had not been mixed since it was first fertilized. Fox and

Kamprath (1970) also reported a residual P response in a Typic Paleudult

fertilized 2 years previously with 0, 50, and 100 kg P/ha. To an acid

Aguic Plinthic Dystropept in Sierra Leone, West Africa, Brams (1973)

applied OSP at rates of 0, 30, 60, and 90 kg P/ha. After cropping with

corn and rice, peanuts and rice, and peanuts for 3 years following P

application, the author reported no differences in crop yields at any

of the rates. Although the check P level initially contained 55 kg P/ha,

the worker concluded that residual P could be utilized for extended

periods without soil P becoming a limiting factor.

Foliar Composition

In a review of nutrient concentration in plant tissue, Bates (1971)

stated that foliar composition was affected by soil properties, environ-

mental conditions, cultivar tissue sampled, and physiological age of

the plant. The relative water solubility of P fertilizers also has

been reported to influence nutrient content in foliage. In a study

previously discussed, Webb and Pesek (1958) reported that, as the frac-

tion of available P in water-soluble form in fertilizer increased from

10 to 100%, the P content of corn ear leaves at early silking also

increased from 0.217 to 0.228%. This same trend of increasing P content

in corn tissue with change from low to high water solubility of P

fertilizers was observed by McLean and Logan (1970) on a low P-fixation

capacity soil in the greenhouse trials. In a high P-retention capacity

soil, these workers did not note this trend and stated that the higher

water solubility of the P was a disadvantage to plant availability in

this case, since this soil had the capacity to tie up most of the water-

soluble P in forms initially unavailable to the corn seedlings.

Phosphorus concentrations in corn ear leaves, usually at silking,

have been noted by a number of workers to be correlated curvilinearly

with corn yields (Bates, 1971). Deficiencies of P were reported by

Hanway (1962) to result in greater differences in percent total P in

corn ear leaves and leaf sheaths than any other plant part. Although

differences in P content of corn leaves and sheaths were noted to

persist throughout the growing season, the author found they were

maximum near silking time and, at that time, P content changes with

leaf position on plant were small. In loamy soils, P concentrations in

leaf tissue of corn ears at silking have been reported to range from

0.08-0.41% P (Table 2).

Table 2. Phosphorus content reported in corn ear-leaf tissue
at early silking

P rates P range in corn ear leaf Reference
tissue at 3 months of age

various 0.08 0.33 Dumenil (1961)

0-134 0.15 0.37 Hanway (1962)

0-44 0.09 0.41 Peck, Walker, and
Borne (1969)

0-68 0.17 0.34 Voss, Hanway, and
Dumenil (1970)


Site Selection

The site chosen for the field experimentation was at the Agricul-

tural Research Center at Jay, Florida, and four years previously had

been cleared of pine (Pinus elliotti). The soil was dominantly Red Bay

fine sandy loam [Rhodic Paleudult], with the southern portion grading

into Dothan fine sandy loam [Plinthic Paleudult]l; both soils occur in

the northwestern areas of Florida and southern Alabama. This soil was

located nearly one mile from a previous long-term study reported by

other workers (Robertson, Thompson, and Hutton, 1966). They demonstra-

ted that corn grown on this soil responded to phosphatic fertilizer and

to residual effects of this fertilizer several years after phosphorus

fertilization was ceased.

Soil samples were taken at 0-36 cm and 36-56 cm depths on March 17,

1971. Thirty cores per depth sample were combined from the previously

uncultivated site. Deep plowing to 36 cm and addition of 6.72 metric

tons of dolomite/ha preceded by one month this soil sampling. Samples

were stored in sealed plastic bags to maintain approximate field mois-

ture prior to transport to the laboratory and subsequent analysis.

1 Soil Taxonomy of the National Cooperative Soil Survey, SCS, USDA,
Washington, D.C., December, 1970.

Laboratory Experiments

Native Soil Phosphorus

Total phosphorus

Soil samples from the experimental site were analyzed for total

soil P after Na2CO3 fusion according to the procedure of Jackson (1958).

Interference of soluble silica with P determination was prevented by

dehydration of soluble silica with 5N HC1 and choice of the ascorbic

acid-molybdate colorimetric procedure given by Watanabe and Olsen

(1965). Details are given in a later section.

Inorganic phosphorus fractionation

The modification by Petersen and Corey (1966) of the sequential P-

fractionation scheme previously proposed by Chang and Jackson (1957)

was employed. The P fractions sequentially extracted by 0.5N NH4F at

pH 8.2, 0.1N NaOH, 0.3M Na3C6H5072H20 and Na2S204, and 0.5N_ H2SO4 are

termed aluminum phosphate, iron phosphate, reductant soluble phosphate,

and calcium phosphate, respectively. The colorimetric determination

of P in these solutions is described in a later section.

Plant-available phosphorus

As has been presented in the literature review, soil extraction

with 0.03N NH4F in 0.1N HC1 (Bray and Kurtz, 1945) has been one of the

methods best correlated with plant uptake of P and plant yields on

Ultisols and Oxisols. For that reason, this extractant was chosen to

determine plant-available P in the soil being studied.

A preliminary study was made with 0. 03N NH4F in 0. N HC1 to evalu-

ate the effect of time of shaking on the amount of P extracted from the

soil using a 1:10 ratio of soil to solution. Shaking times used were 1,

5, 15, 20, 60, 120, 240, 360, 720, 1440, and 2880 minutes with two re-

plications. Following a 20-minute period for filtration, solution P was


A second preliminary study was made to determine the effect of

sequential extractions on extractable P using a 1:10 ratio of soil to

solution and shaking for 5 minutes. From these studies and the previous

literature, it was decided to extract the soil from field plots with

0.03N NH4F in 0.1N HC1 at a 1:10 ratio of soil to solution with 5 minutes

of shaking. Total time elapsed for shaking and filtration was 25


A third preliminary study was performed to determine the influence

of toluene on P retention by this Paleudult. Suspensions were prepared

using 75 and 200 ig P/g soil employing 5 g of soil and 50 ml of solu-

tion with and without addition of two drops of toluene. These solutions

were shaken twice per day for 30 minutes over a period of 6 days after

which time solution P was measured.

Soil Retention of Applied Phosphorus

To determine which of the various methods for defining soil P

retention gives the best description of P retention by this Paleudult,

studies were initiated by equilibrating several levels of P with moist

soil in the presence of 0.01M CaC12. On an oven-dry basis, 5 g of the

moist, uncultivated Red Bay sample from 0-36 cm depth, screened to < 1 mm,

were added to each of 72 polypropylene centrifuge tubes. As KH2PO4,

12 levels of P were applied: 0, 10, 50, 75, 100, 150, 200, 500, 1,000,

2,000, 3,500, and 5,000 pg P/g soil. Total solution volume in each

sample was 50 ml. Equilibration times of 1, 72, and 144 hours on a

reciprocal shaker were used. Two replications of each level at each

equilibration time were employed. To each of the samples that were to

be equilibrated for 72 and 144 hours, were added 2 drops of toluene to

deter microbial assimilation of added P. Upon completion of each equi-

libration, the suspensions were passed through a plastic assembly (Nalge)

with a 0.2 b-membrane filter to minimize the possibility of colloids

and microbial presence in phosphorus analysis immediately following the


Phosphate potential

It has been suggested that the availability of soil P is determined

primarily by the chemical potential of phosphate (Schofield, 1955). To

test this hypothesis, phosphate potentials (PP) for each equilibrium

solution of the preceding study were calculated by the following equa-


PP = ?pCa + pH2PO4 = -(logCca + logYCa(H2P4)2)

-(logCH2PO4 + logYCa(H2PO4)2) [5]

where log yi = (Az+z- [_ )/(i + c,8 ) [6]

in which constant A is 0.509, constant product a0 is 1.5, z+ and z- are

the respective valences of the cation and anion species in solution and

S= iCiZi2 [7]
where C is concentration and Z is the valence of each ion in solution.

For each addition of P, AP was calculated by the equation

AP = (Cp Cpf)/100 [8]
where Cpi and Cpf are the initial and final P solution concentrations,

expressed in lg/ml. Customarily, after determining PP and AP values,

these points have been plotted as a Q/I, or Quantity/Intensity, curve

(White and Beckett, 1964) which was then extrapolated to obtain both

PP at AP = 0 and replaceable P. The slope of the curve was referred

to by these authors as potential buffering capacity (PBC) and is herein

referred to as P-sorption maximum (Psm). In the present studies, values

obtained by graphical extrapolation of the Q/I curves were compared

with those given by linear regression equations, and it was found that

the equations presented a more accurate description of PP at AP = 0 and

replaceable P. For these equations, with PP employed as the Y variable

and AP as X, PP at AP = 0 was termed b in the equation Y = mX + b.

Where AP was Y and PP was X, Psm and replaceable P were noted to be the

respective m and b values in the linear equation.

Phosphate sorption isotherms

The amount of P held by the soil and the equilibrium concentration

of P in solution were used in several equations to assess which most

accurately described P retention in this soil. Though previously pre-

sented in the literature review, the employed equations are again defined

in a manner consistent with the materials and methods section.

Freundlich equation.--This equation has been found to define soil

P retention over a wide range of equilibrium P concentrations (Olsen

and Watanabe, 1957). This equation is

x/m = kcn [1]

where x is 4g of P held, m is grams of soil, c is equilibrium concentra-

tion of P in pg/ml in solution, and k and n are constants.

Langmuir equation.--As has been pointed out in the literature

review, this equation has been found to more closely define P retention

in some soils than does the Freundlich (Olsen and Watanabe, 1957). Also,

it is possible to calculate the P-sorption maximum with Langmuir, while

this is not possible with Freundlich. The Langmuir equation is
c = 1/kb + c/b [2]
where c, x, and m are as described in equation [1], b is the phosphate-

sorption maximum, and k is a constant.

Square-root inclusion in parabolic equation.--Gunary (1970) found

a better fit for P retention on a variety of soils by modifying the

parabolic equation. His equation is
S A + DJ- + Be [3]
xTm -
where c, x, and m are as described in equation [1] and B, A, and D are


Components Contributing to Soil Phosphorus Retention

As discussed in the literature review, soil Al, Fe, and organic

matter have been found to be active in P retention by Ultisols and

Oxisols. Accordingly, the approach taken in this study was to succes-

sively remove each of these components from the Rhodic Paleudult under

consideration and then to determine P retention after various additions

of P. The relative contributions of these factors to P retention should

then be separated. In this study, moist uncultivated Red Bay fsl, taken

at 0-36 cm and screened to < 2 mm, was used.

Organic matter removal

Organic matter removal by two reagents, hydrogen peroxide and

sodium hypochlorite, was compared.

Hydrogen peroxide treatment.--To each of a set of nine 90-ml poly-

propylene centrifuge tubes were added 10 g of soil on an oven-dry basis,

10 ml of distilled water, and two 5-mr increments of 30% H202. The

mixtures were stirred, left overnight, and placed in a water-bath at

65-70C for 30 minutes with an additional 5 ml of H202 added to each

sample after the initial reaction ceased. To maintain a 10% solution,

10 ml of distilled water were added to each sample upon cooling. In

succession, 50-ml washings of the soil were made with 1N Ca(OAc)2,

1N CaC12, and distilled water. After a second wash with distilled water,

the suspensions were shaken for 6 days and decanted. The stoppered

samples were stored for 2 months. Determinations of suspension pH were

made after storage. The washing were combined and analyzed for P, Al

and Fe.

Sodium hypochlorite treatment.--Another set of samples with sodium

hypochlorite was also used to remove organic matter. To a set of nine

90-ml polypropylene centrifuge tubes with 10 g of soil on an oven-dry

basis were added 20 ml of 5.25% NaOC1 adjusted to pH 9.5 immediately prior

to use (Lavkulich and Wiens, 1970). These samples were placed for 15

minutes in a boiling water bath. To maintain a Na-saturated system,

washings with 50 ml of 0.5N NaC1 and 50 ml of 0.5N NaC1 followed. A

second 0.05N NaC1 wash was shaken for 6 days and decanted. The stop-

pered samples were stored for 2 months after which time suspension pH

determinations were made. The washings were combined and analyzed for

P, Al, and Fe.

Citrate-dithionite-bicarbonate treatment

Six samples previously treated with H202 and six others with NaOC1

were twice-treated with citrate-dithionite-bicarbonate (CDB) according

to the method of Kunze (1965). To each of the above samples were added

40 ml of 0.3M Na3C6H507*2H20 and 5 ml of 1M NaHC03. These suspensions

were heated in a water bath to 750 with 2 g of solid Na2S204 being added

to each and then occasionally stirred over a 15-minute period. Ten ml

of saturated NaC1 were added to cause colloids to flocculate. Super-

natants from both CDB treatments were combined and analyzed for P, Al,

and Fe. The pH of the suspensions was determined.

Dissolution with hot sodium hydroxide

After 3 weeks of storage, three samples after the H202 and CDB

treatments and three with NaOC1 and CDB treatments were subjected to

50 ml of 0.5N NaOH at 100C for 2 minutes to permit dissolution of free

alumina and amorphous aluminosilicates (Jackson, 1965). The samples

were cooled rapidly by an ice-water bath to stop the reaction and then

centrifuged at 2,100 rpm for 15 minutes. The solutions were decanted

and analyzed for P, Al and Fe. The soils were washed twice with 50 ml

of iN Ca(OAc)2 at pH 5, once with 50 ml of iN CaCI2, and four times

with 50 ml of distilled water until a negative AgC1 test in the washing

indicated completion of salt removal. Measurements of electrical

conductivity also showed no salts to be present after the washings and

these washes were discarded.

Check treatments

The check treatments were thrice-replicated treatments of (1) soil

and distilled water, (2) soil Ca saturated with IN Ca(OAc)2 at pH 5,

(3) soil Ca saturated with iN CaC12, and (4) soil Na saturated by 0.51

NaCI. Samples subjected to check treatment (1) were treated exactly

as were the hydrogen peroxide-treated samples with the exception that

whenever H202 was added to the peroxide-treated samples, distilled water

was added to check treatment (1) samples. For example, in check (1),

10 g of soil, 10 ml of distilled water and two 5-ml increments of dis-

tilled water were stirred, left overnight, and placed in a water-bath

at 65-70C for 30 minutes with an additional 5 ml of distilled water added

to each sample after the initial reaction of the peroxide-treated samples

ceased. As in check (1) samples, check (2) samples were treated exactly

as were the peroxide-treated samples with the exception that whenever

H202 was added to peroxide-treated soils, iN Ca(OAc)2 at pH 5 was added

to check (2) soils.

To the samples of check (3) were added the same quantities of 1N

CaC12 as were added H202, CDB, and NaOH to the H202-CDB-NaOH-treated

samples. All other conditions were similar. In other words, 1N CaC12

was substituted for H202, while the samples were subjected to exactly

the same heating rates and times and washes as the peroxide-treated

samples. These samples of check (3) then received 47 ml of 1N CaC12,

instead of 40 ml of 0.31 Na2C6H507"2H202, 5 ml of iM NaHCO3, and 2 g

solid Na2S204, and were heated to 75C. Instead of dissolution with hot

0.5N_ NaOH, check (3) samples were then subjected to dissolution with

hot 1N CaC12. As check (3) substituted 1N CaC12 for H202, CDB, and NaOH,

check (4) substituted 0.5N NaC1 for NaOC1, CDB, and NaOH under similar


Phosphorus additions

After a month, each sample from the above treatments received two

washes with 0.01M CaCl2. This was followed by sequential additions of

0.31, 0.62, 1.24, 3.10, 6.20, and 31 pg P/g soil. Phosphorus was

applied as Ca(H2PO4)2 H20 in 50 ml of 0.01M CaC1Z. Each level of P added

was shaken for 30 minutes. The pH was determined and the suspensions

were then centrifuged for 15 minutes at 2,000 rpm. Supernatants were

analyzed for P. At all levels of applied P, except 31.0 4g/g, undetectable

amounts of P in solution required that the three replicates be combined

for each of the ten treatments. These solutions were evaporated to

dryness and taken up in 10 ml of 0.1N HC1 containing 0.05 4g P prior to

P analysis. All samples between sequential P additions and all super-

natants prior to analysis were kept refrigerated. Total time elapsed

in this study was 4 months.

Studies using 3sp

In a subsequent study of 2-weeks' duration, each of the above 30

samples received 50 ml of 0.01M CaCl2 and 2 ml of a solution containing

42.0 pg P as Ca(H2PO 4)2-0/g soil with 32P present in the same chemi-

cal form. These soil solutions were shaken for 30 minutes once a day

over a 3-day period. Following centrifugation at 2,000 rpm for 15

minutes and decantation, the amount of 3sP not sorbed by the soil was

determined. The procedures for this and subsequent 32P determinations

are presented in the section on analytical procedures. Duncan's new

multiple range test (Steel and Torrie, 1960) was used to determine which

treatment differences were significant in this study and in all subse-

quent studies with 32p.

In a concomitant study also of 2 weeks' duration, 10 g of the un-

treated soil were added to each of 15 centrifuge tubes. Treatments were

0, 1.55, 6.20, 15.5, and 31.0 4g 3AP as Ca(H2P04)2.H20/g soil added in

50 ml of 0.01M CaC12 per sample. The samples were shaken and allowed

to stand overnight. Each sample then received 2 ml of solution containing

21.0 4g P/ml labeled with sP and supplied as Ca(H2P04)2.H20. These

suspensions were shaken daily for 30 minutes over a 3-day period, then

centrifuged at 2,000 rpm for 15 minutes and subsequently analyzed for

32P not sorbed by the soil.

After both 32P studies, each soil sample was mixed with 50 ml of

1N NH4OAc at pH 4.8, since this reagent had been found to remove dical-

cium phosphate from soils (Fiskell and Spencer, 1964). These samples

were shaken for 15 minutes, centrifuged at 2,000 rpm for 10 minutes,

and decanted. The solutions were analyzed for 32p. Determinations of

pH and 31P were also made.

Following the above extraction, 50 ml of 1N NH4F at pH 8.2 (Fife,

1962) were added to extract the soils used in both 32P studies. Samples

were shaken for 15 minutes, centrifuged for 10 minutes at 2,000 rpm.

The supernatants were analyzed for 32P and Al. Since the pH values of

the soil suspension treated with 1N NH4F at pH 8.2 dropped to 6.0, super-

natants were also analyzed for Fe. Analysis for 3iP in the supernatants

was also made.

Fertilizer Phosphorus Dissolution

Dissolution of soluble salts from fertilizers has been reported by

Blouin and Rindt (1967) to be retarded by coating the pellet with a

combination of sulfur and wax. In a laboratory study, information was

sought on the dissolution rates of conventional non-coated P fertilizers,

such as ordinary superphosphate (OSP), concentrated superphosphate (CSP),

and acidulated rock phosphate (ARP), compared to sulfur-coated CSP (SCSP).

Fertilizer dissolution without soil

In separate, covered 0.2 1k-membrane filters, were placed 5 g each

of fertilizer granules of OSP (8.7% P), CSP (20.1% P), ARP (14.5% P,

5.8,4 water soluble), and SCSPa(15.5% P), and SCSPb (14.8% P). Every

week for 26 weeks each fertilizer was leached with 50 ml of distilled

water. Determinations of pH and P were made on each leachate. Since

the filter assemblies were covered to retard contamination and evapora-

tion, fertilizer pellets remained moist between leachings. Following

termination of the leachings, residual fertilizer pellets were analyzed

for P and subjected to X-ray diffraction.

Fertilizer dissolution and retention in soil

This study was identical to the above one except that each of

these P sources was mixed with 10 g, on an oven-dry basis, of the moist

uncultivated Red Bay soil (0-36 cm depth, < 2 mm diameter). As in the

concomitant study, weekly leachings with 50 ml of distilled water were

added to each soil-fertilizer mixture for a 26-week period. Phosphorus

and pH determinations were made on each leachate. Soil-fertilizer

mixtures remained moist between leachings. After the experiment, the

fertilizer pellets were removed manually from the soil, analyzed for P,

and subjected to X-ray diffraction.

Field Experiments

To allow for adequate comparison of SCSP and CSP used under field

conditions, six field studies were performed over a 3-year period on the

site previously discussed. Corn (Zea mays L.) was chosen as the indicator

crop. Wheat (Triticum sativa) was used as a cover crop in the fall and

was plowed under a month prior to planting the second and third years.

Phosphorus Sources, Rates, and Placements, 1971

Phosphorus sources, rates and placements

In the main experiment, the P sources were granular CSP (20.1% P)

and two SCSP fertilizers which were supplied by the Tennessee Valley

Authority. The SCSP fertilizers consisted of CSP coated with sulfur, 3%

wax, 0.25% coal tar microbicide, and 2% diatomaceous earth conditioner.

These sulfur-coated fertilizers differed only in percent P and amount of

S-coating present. Source 1 of SCSP pellets contained 15.5% P, coated

with 16.6% S based on total weight. Source 2 of SCSP contained 14.8% P

and 19.3% S as coating based on total weight of the pellets.

Rates of CSP used were 28, 56, 84, 112, 140, 168, and 196 kg P/ha.

Both SCSP fertilizers were used at the first five of the above rates.

For these sources, the two higher rates were excluded since space was

not available for the extra plots within the site. Check plots received

no phosphate fertilizer. Each rate of each fertilizer was incorporated

as broadcast and as row placement. Both placements are described in a

later section.

Experimental design

Phosphorus sources, rates, and placement were compared in a split-

plot arrangement. Main plots contained the various combinations of the

P sources and rates, while the split-plot treatment was fertilizer

placement. Main plots were arranged in a randomized complete block

design with each treatment combination replicated four times. This

study consisted of 144 plots, each 4.38 x 9.11 m. Each plot contained

four rows of corn with the two outside ones serving as border rows

between treatments and the inner two being those from which soil, plant,

and yield data were collected.


The soil was prepared for planting and stakes were placed along the

perimeter of the plot boundaries. On March 17, 1971, the area was to

receive a uniform broadcast application of 149 kg/ha of N as NH4NO3,

112 kg/ha of K as KC1, and of gypsum as a source of S. However,

the first replication received four times the above rate of fertilization

due to an equipment error. At each rate, one-half of each plot received

P fertilizer broadcast uniformly by hand; the other half-plot received

P fertilizer in two bands, each 5-6 cm wide and 7-8 cm from each of the

four planting rows in the plot. The fertilizer was incorporated care-

fully with a rotary hoe to minimize horizontal displacement of the P

fertilizer. On May 14, all plots received 75 kg/ha of N as NH4NO3 as

a sidedressing.

Planting and cultural practices

The field was planted with 'Pioneer' 3009B corn two days after

fertilization. The four rows per plot were 91 cm apart, with seed

spacing at 15-cm intervals. Where necessary, a uniform stand was ob-

tained by replanting two weeks later at the above spacing. When the

corn was 30-cm high, the stand was thinned to 30-cm spacing to give

approximately 25,860 plants/ha. Later in the season, the field was

sprayed for control of corn earworm. Irrigation was not used in this

study. Throughout most of the season, rainfall distribution (Appendix

Table 28) was sufficient to prevent evidence of wilting due to drought


Plant sampling

Entire corn plant tops, excised at soil level, were sampled from

the border rows of each plot on May 17, 1971. Samples consisted of

8 plants per plot where plants were small, and 4 plants per plot where

plants were large. These plants were desiccated in a drying room and

ground by a Wiley mill through a stainless steel sieve. One g of each

ground sample was ashed for 2 hours at 350C and for 4 hours at 600C.

Details of plant analyses are given in a subsequent section.

At silking stage on June 12 and 13, 1971, corn ear-leaf samples from

eight plants from the two inner rows of each plot were taken. For those

plants not silking, the fifth leaf below the whorl was taken, since

the nutrient composition of these leaves has been found to resemble

most closely that of the corn ear leaves at silking (Hanway, 1962). These

samples were dried and ground as described above.

Soil sampling

Soils were sampled from each plot on June 23, 1971, at which time

the corn growing on these plots was at the late-silking stage. Six

cores,0-15 cm, were taken between plants within the two inner rows of

each plot. Samples were air-dried on greenhouse benches to cease dis-

solution from P fertilizer pellets and to facilitate removal of any of

these pellets or their fragments prior to analysis. After having passed

the soil through a 2-mm sieve carefully to avoid crushing fertilizer

pellets, these pellets were removed manually before analysis. All soils

were extracted with 0.03N NH4F in 0.1N HC1 as described in an earlier

section. The subsequent P analysis is described in a later section.

Soils were sampled at harvest on September 16, 1971, in the same

manner as at the earlier date. All plots were sampled, except those with

band placement at rates of 56, 112, 168, and 196 kg P/ha. Soils were

air-dried, sieved, extracted, and analyzed for P as described above.

Fertilizer pellets manually removed from samples before extraction were

subjected to X-ray diffraction to identify compounds present in the

pellet residue.


Corn was hand-harvested from the two inner rows of each plot on

September 14 and 15, 1971. Ears from each plot were shucked and weighed.

Moisture percentage of the ears was determined and yield weights were

adjusted to 14% moisture since yields are reported at this commonly

accepted moisture level. Numbers of stalks and ears in these two inner

rows were recorded for each plot. Six ears from each plot were dried,

shelled, weighed, and kernel weight was determined as a percentage of

total ear weight.

Since the corn had remained on the stalks for approximately a

month after reaching maturity, some ears in several plots suffered

racoon damage. In such cases, the average ear weight for undamaged ears

on that plot was substituted for that of each damaged ear.

Statistical analysis of data

This split-plot field study was in actuality a 5 x 3 factorial in

rates and sources plus the three additional treatments of check, CSP

at 164 and 196 kg P/ha. Therefore it was decided to place the emphasis

of the analysis of variance on the factorial set of treatments. The sum

of squares associated with rate in each analysis was partitioned into

regression components by means of orthogonal polynomials. Responses

analyzed were P content of corn plant tissue at pre-silking stage, P

content of corn ear-leaf tissue at late silking stage, P content of soil

when plants were at late-silking stage, corn yield, average ear weight,

kernel weight as percentage of ear weight, number of ears, and ears/stalk.

Linear and curvilinear correlation coefficients were obtained from corn

yields against soil P levels at silking and at harvest, corn yields

against P, Al, Fe, Ca, K, Mg, Mn, Zn, and Cu content of corn plants prior

to silking, corn yields against P, Al, Fe, Ca, K, Mg, Mn, and Si content

of corn ear leaves at silking, P levels in corn plants prior to silking

against the other nutrient levels in corn plants prior to silking, and

P levels in corn ear-leaf tissue at silking against each of the other

nutrient levels in corn ear leaves at silking.

Fertilization with Mixtures of Coated and Non-coated Phosphorus
Fertilizers, 1971

Design and cultural practices

The purpose of this study was to investigate the effect on corn

yield of various combinations of SCSP and either pelletized ordinary

superphosphate (OSP) containing 8.7% P, or the pelletized CSP used in the

main experiment. Six different combinations of each uncoated source

with SCSP (15.5% P) were used, varying amounts of each from 0 to 100%

by 20% increments with P rate held constant at 84 kg/ha. This ferti-

lizer was applied broadcast. Zach treatment combination was replicated

four times using a randomized complete block design. This study con-

sisted of 48 plots, each 4.38 in x 9.11 m. Each plot contained four rows

of corn, with soil, plant, and yield data being collected from the two

inner ones.

Due to this study being located adjacent to the one previously

discussed, all fertilization other than P treatments, planting, and

cultivations were identical to those of the first study and were per-

formed on the same dates.

All plant and soil samplings and procedures and the corn harvest

were identical to those of the first study, with the exception that

soil samples were taken from this second study only at harvest time.

Statistical analysis of data

Analysis of variance was performed on the same responses analyzed

in the previous study. In order to maximize amount of information

recovered, the sum of squares associated with percentage SCSP in the

fertilizer mixture was partitioned into regression components by means

of orthogonal polynomials. Linear and curvilinear correlation coeffi-

cients were obtained for the same combination of factors as given in

the previous study.

Phosphorus Sources, Rates, and Placements, 1972


This study was a continuation on the same plots of the P source,

rate, and placement study previously discussed, with a major modifica-

tion in P fertilization. In the previous experiment,two placements of

P fertilizers were studied, whereas in the continuation, only plots

where P was applied broadcast were refertilized in 1972. Thus this

modification allowed for the study of the residual effect of coated and

non-coated P sources. Used as P sources on the refertilized plots were

CSP (20.1% P) and SCSP (15.5% P), with each being applied to the plots

where this same P fertilizer, coated or non-coated, was used the

previous year. Since the previous year the greatest response to P was

obtained at the lowest rate, the amount of P applied on the refertilized

plots was one-half that of the previous year, e. g., plots receiving

28 kg/ha in 1971 received 14 kg/ha in 1972. Thus the rates applied were

14, 28, 42, 56, and 70 kg P/ha. However, no further P fertilizer was

applied on the plots with the two highest levels of CSP in 1971. Again,

check plots received only the blanket N and K fertilization. Fertiliza-

tion of the field with 150 kg N as NH 4N3, 112 kg K as KC1, and 37 kg

gypsum/ha was applied broadcast on March 20, 1972. All P fertilizer

was manually broadcasted.

The field was planted at 30-cm intervals on rows 91 cm apart with

'Pioneer' 3369A corn on March 21, 1972. This spacing gave approximately

25,860 plants/ha. When the plants reached about 60 cm in height,

56 kg/ha of N as NH4NO3 were sidedressed. Cultural practices and pest

control were uniform across all experimental treatments and were con-

sistent with the practices of the Agricultural Research Center at Jay,

Florida. Weed control was excellent from the use of 'Bladex' herbicide.

Irrigation supplied 1.25 cm of water to the plants after any week when

natural rainfall (Appendix Table 67) was insufficient to supply this


Plant sampling

At late-silking stage on June 13, 1972, five corn ear-leaf samples

were taken from the two inner rows of each of the 144 plots. For any

plants not yet silking, the fifth leaf below the whorl was taken. Leaf

samples were treated in the same manner as those of the previous year

and analyzed for P, Al, Fe, Ca, K, Mg, and Mn.

Soil sampling

Prior to refertilization and planting, soils were sampled from all

plots which were to remain as residual plots. This sampling of March

20, 1971, consisted of six core samples per plot, taken at 0-15 cm.

Soils were dried, extracted, and analyzed employing the same procedures

as previously outlined.

Soil samples were taken from all plots on June 14, 1972 when the

corn was at the late-silking stage. Methods of sampling, drying, asking,

and extraction for subsequent P analysis were the same as those used

for previous soil samples.


The corn was harvested from this field study on August 23, 1972,

using the same procedures as had been used the previous year with one

exception. To determine kernel weight as a percentage of total ear

weight, three ears, rather than six, were taken from each plot.

Statistical analysis of data

The statistical approach for this 5 x 3 factorial study plus the

three additional treatments was alike that of the previous year. Again,

emphasis was placed on the factorial set of treatments in the analysis.

The sum of squares associated with each rate in each analysis was

partitioned into regression components by means of orthogonal polynomials.

Responses analyzed were P in corn ear-leaf tissue at late-silking stage,

corn yield, average ear weight, and kernel weight as percentage of ear

weight. Linear and curvilinear correlation coefficients were obtained

from corn yields against soil-P levels at preplanting, corn yields

against soil-P levels at corn silking, corn yields against soil-P levels

at harvest, corn yields against P content of corn ear leaves at silking,

and P levels in corn ear leaves at silking against each other nutrient

concentration in corn ear leaves at silking.

Various Combinations of Coated and Non-coated Fertilizers with and
without Strip Mulch, 1972

This study involved four factors each at two levels run in four

replicates of a randomized complete block arrangement. Due to site size,

only 12 of the 16 possible treatment combinations were used. The design

was essentially a 3/4-replicate of a 24 factorial with the 48 plots

being the same size as those of previous studies.

Used as fertilizer sources in this study were SCSP and CSP, urea

and S-coated urea (SCU), KC1 and S-coated KC1 (SKC1). The SCU supplied

by TVA, contained 34.8% N, 19.5% S, 3.0% wax sealant, and 0.25% coal

tar microbicide. The SKC1, also supplied by TVA, contained 22.9% K,

26.1% S, 3.0% wax sealant, and 0.25% microbicide. A black polyethylene

strip mulch, 25 cm wide (Hayslip and Iley, 1966), was used with various

combinations of the fertilizers. Treatments employed were CSP and SCSP

each in combination with urea and KC1 with and without plastic mulch,

CSP and SCSP each in combination with SCU and SKC1 with and without

plastic mulch, CSP and SCSP each with SCU and KC1 without mulch, and

CSP and SCSP each with urea and SKC1 also without mulch.

Fertilization, planting, and cultivation practices

From the various P, N, and K sources used, each plot received 42 kg

P, 168 kg N, and 140 kg K/ha. All fertilizer was applied manually as

a band between the corn rows on March 21, 1972. Plastic strip mulch

was placed by machine to cover the fertilizer bands on either side and

between the two inner rows of each plot. Approximately 7 cm of each

edge was covered with soil by the strip mulch applicator. Planting,

additional fertilization, and cultivation procedures were identical to

those of the previous study and were performed on the same dates. All

plant and soil samplings and the corn harvest were identical in all re-

spects to those performed in the residual P study previously discussed.

Statistical analysis of data

Since the design of the experiment was related to the quarter

replicates of the 24 factorial, the factorial effects were joined in

chains or sets of four. Treatment combination determined the four

members of each chain. By assuming one of the four members in each

chain to be negligble, it was possible to estimate the three remaining

effects. The factor with the greatest number of interactions, usually

three or greater, in each of the four chains was the one considered

negligible to allow analysis of the other factors. Analyses of variance

were carried out on corn yield, average ear weight, and kernel weight

percentage of ear weight. Linear and curvilinear correlation coeffi-

cients were obtained from the same combinations performed in the residual

P study.

Evaluation of Residual P from Previous Two Years, 1973

These two studies were continuations on the same plots of the two

previous studies of 1972, except that no P was applied in 1973. Blanket

fertilization of the field was alike with the previous year and applied

in mid-March, 1973. Planting with the same variety and spacings as

employed the previous year was one day after fertilization. All cultural

practices were uniform across the treatments and followed those of the

previous year. Irrigation was not employed. Rainfall was sufficient

throughout the season (Appendix Table 98) to prevent any appearance of

drought injury to the corn.

Plant sampling

At late-silking stage on June 17, 1973, five corn ear-leaf samples

were taken from the two inner rows of each of the 192 plots. For any

plant not silking, the fifth leaf below the whorl was taken. Leaf

samples were treated in the same manner as those of the previous year

and were analyzed for P, Al, and Fe.

Soil sampling

After the blanket fertilization, but prior to planting, soils were

sampled from all 144 plots of the main experiment, i. e. that study

which in 1972 evaluated residual P. This soil sampling in mid-March,

1973, consisted of six cores per plot taken at 0-15 cm, Soils were

stored in sealed plastic bags prior to the sieving and analysis proce-

dures as previously described for other soil samplings.


Corn was harvested on August 22, 1973, from all plots of both

experiments using the same procedures as were used the previous year.

Statistical analysis of data

The statistical approaches for both experiments were alike those

of the previous year. Responses analyzed were P in corn ear-leaf tissue

at late silking and corn yield. Linear and curvilinear correlations

were obtained from corn yields against soil-P levels at preplanting,

corn yields against P content of corn ear leaves at silking, and P

levels in corn ear leaves at silking against both Al and Fe concentra-

tions in corn ear leaves at silking.

Analytical Procedures

Clay Mineralogy and Associated Properties


Pretreatments are necessary before effective mineralogical analysis

can be performed (Kunze, 1965). However, since any pretreatment will

affect the clay somewhat, the mildest treatment which would accomplish

the purpose was chosen. Removal of soluble salts, di- and tri-valent

cations, and residual lime carbonates from Red Bay fsl, 0-36 cm, <1 mm,

was accomplished with acid sodium acetate according to the procedure of

Jackson (1956). Organic matter was removed from the sample by hydrogen

peroxide oxidation as described by Douglas and Fiessinger (1971). Free-

iron oxide removal from the sample was accomplished by two successive

citrate-dithionite-bicarbonate (CDB) treatments (Mehra and Jackson,

(1960). The CDB supernatants were analyzed for Al and Fe as described

in a later section. To disperse this treated soil while removing little

amorphous material, 1N Na2CO3 was used to bring the soil solution pH

to 9.5 as recommended by Rich and Zelazny.*

Washings and triturations with water at pH 10 through a 4.4 P sieve

were used to separate sand from silt and clay in the sample (Day, 1965).

Silt was separated from clay by centrifugation at 2,000 rpm for 5 minutes

in a 200-ml glass centrifuge bottle filled to the 10-cm level with the

above water. After decantation, water was again added to the 10-cm

mark, the suspension was stirred with a rubber-tipped stirrer at high

speed, centrifuged at 1,500 rpm for 10 minutes, and decanted. The above

procedure with centrifugation at 1,000 rpm for 2 minutes was repeated

until the supernatant was clear. The settled fraction was classed as

silt, while the supernatant was labeled clay. Sand, silt, and clay in

the sample were quantitatively determined by weight. Clay was not

fractionated. The supernatant was adjusted to pH 4 with 6N HC1 to

induce flocculation and subsequent concentration of the clay fraction

(Kunze, 1965).

*L. W. Zelazny, personal communication.

Clay mineralogical analyses

X-ray diffraction.--A 225-mg portion of the separated clay was

mounted by suction onto each of four ceramic tiles (Rich, 1969). Sub-

sequently, two tiles were saturated three times with 5 ml of iN MgCl2,

washed twice with distilled water, and solvated with 5 ml of 20%

glycerol solution. The remaining two tiles were saturated three times

with 5 ml of 1N KC1 and washed twice with distilled water. X-ray dif-

fractograms were obtained of Mg-saturated, glycerol-solvated samples

with and without 100C treatment. Also subjected to X-ray diffraction

were K-saturated samples without heat treatment and then with heating

in an oven at 110, 300, and 550C. Blank tiles were inverted over samples

in the oven to inhibit peeling. The diffractograms were run from 3 to

30 degrees 20 on a General Electric XRD-700 instrument with Ni-filtered

CuKa radiation. Radiation was detected with a proportional counter.

Instrument settings were a potential of 25 KV and a 25-ma current. The

respective goniometer scan and strip-chart drive speeds were 20 and

1.27 cm per minute. Scale expansion and zero offset were 775 and 8,

respectively, while range was 10K. Angles were converted to angstrom

units (A) from a prepared chart.

Differential thermal analysis.--The clay fraction was saturated

twice with 1N MgCl2, washed six times with distilled water to a negative

AgC1 test, dried at 40C, ground to pass a 4.4 tp sieve, and desiccated

over Mg(N03)2 for 72 hours at 56% relative humidity (Barshad, 1965).

A 100-mg portion of the sample was then sandwiched between previously

fired (1000C) asbestos around the chromel-alumel thermocouple according

to the procedure of Reneau and Carlisle (1971). The sample holder was

then connected to a transistorized circuit and heated in an inconel

block in a Deltatherm D2000 DTA instrument. Settings were 25, sensi-

tivity using a heating rate of 10C per minute from room temperature

(23C) to 1000C. A standard of Ward's Brazilian gibbsite, < 4.4 4,

was prepared, packed, and heated in the same manner as was the clay

sample. Endothermic reaction areas were measured by tracing the endo-

thermic peaks on tracing paper of uniform texture and weight, cutting

out the peak area, and weighing these traces. Slope ratios of endo-

thermic peak sides as an indication of kaolinite and halloysite were

made as described by Bramao et al. (1952).

Infrared analysis.--One mg of the air-dried clay remaining from

DTA preparation and 400 mg of spectroscopic-grade KBr were dried at 105C

and ground for 30 seconds in a plastic "Wig-L-Bug" with a stainless

steel ball. The mixture was then placed in a stainless steel die in a

Carver Laboratory Press, Model M, with a vacuum of 8,000 psi applied

for 5 minutes (Mortensen, Andersen, and White, 1965). Infrared patterns

of the nearly transparent pellets were obtained from 4,000 to 650 cm-1

on a Model 700 Perkin-Elmer IR-Spectrophotometer at a scan speed of 8

minutes. Instrument calibration was made with standard polystyrene

film. Ward's Brazilian gibbsite was used similarly for a reference


Selective dissolution with hot NaOH.--Dissolution of the clays by

boiling 0.5N NaOH was performed (Hashimoto and Jackson, 1960). To a

previously dried (105C) and weighed polypropylene centrifuge tube was

added a previously dried (105C) 75-mg sample of the Red Bay clay and

80 ml of 0.5N NaOH. The mixture was boiled rapidly for 2.5 minutes,

cooled, centrifuged at 2,000 rpm for 5 minutes and decanted. A second

8-ml aliquot of 0.51 NaOH was added, mixed, centrifuged, and decanted,

To the soil was added 80 ml of 1N HC1. The suspension was mixed, cen-

trifuged, and supernatant was discarded. Successive 50-ml increments

of water and acetone were added, mixed, centrifuged, and decanted. The

centrifuge tube with the sample was dried overnight at 105C and weighed.

Weight loss is normally attributed to non-crystalline material and

organic matter dissolved by the treatment; however, Dixon (1966) showed

that up to 98% of gibbsite can be dissolved by hot NaOH dissolution.

Accordingly, a Ward's Brazilian gibbsite standard was treated in the

same manner as the clay. Weight loss of clay after treatment was at-

tributed to the non-crystalline material, organic matter, and gibbsite


Selected clay properties

Specific surface area.--The ethylene glycol monomethyl ether

(EGME) method of Carter, Heilman and Gonzalez (1965) was employed to

determine surface area of the clay. A 1.2-mg sample of Red Bay clay

was placed in a tared aluminum dish and dried to a constant weight over

P205 in an evacuated dessicator. Three ml of EGME were added to the

sample forming a mineral-adsorbate slurry which was subsequently equi-

librated in a CaC12-monoglycolate dessicator until constant weight

was obtained. Surface area was determined according to the relation-


A = Wg/(Ws x 2.86 x 10"4) [9

where A is specific surface in m2/g, Wg is weight of EGME retained in
sample, Ws is weight of dried sample in g, and 2.86 x 10 is the amount
of EGME required to form a monolayer on a 1-m sample.

Clay-bound organic matter.--The Walkley-Black (1934) method for

easily reduced organic matter was used to determine the amount of clay-

bound organic matter. To both a blank and 0.5 g of oven-dried (105C)

clay were added 10 ml of 1N K2Cr207, followed by gentle swirling and rapid

addition of 20 ml of concentrated H2S04. After standing for 30 minutes,

200 ml of H20 and 5 drops of orthophenanthroline-ferrous complex indi-

cator were added immediately prior to titration. Titration with 0. F

FeSO4 to a red-brown end-point followed and organic matter content was


Clay chemical analyses

Cation exchange capacity.--To 50 mg of :~g-saturated, salt-free clay,

dried at 105C, were added 200 ml of 1N NH4NO3. The mixture was shaken

for 4 hours, centrifuged at 2,000 rpm for 5 minutes, and the supernatant

was analyzed for Mg. From the Mg present in solution, CEC was deter-


CD?-extractable cations.--Supernatants of CDB extraction were ana-

lyzed for Fe, Al, and Si. To prevent Na2S204 interference with the Fe

analysis, 5 ml of the CDB supernatant plus 30 ml of H20 were heated at

75C for 15 minutes. Dropwise addition of 30% H202 was made until the

full ferric yellow color persisted and 1 ml of excess H20 was added.

After continued heating for 15 minutes, the solution was cooled to 25C

and 3 drops of 61 HC1 were added. Iron in this modified CDB extract

was determined on a Perkin-Elmer Model 303 DU Atomic Absorption Spectro-

photometer with an acetylene-air oxidizing flame at a wavelength of

2,483A. Aluminum and silicon were determined directly from the CDB

extracts on the same spectrophotometer with a nitrous oxide-acetylene

flame at wavelengths of 3,093 and 2,516A, respectively. These elemental

L. W. Zelazny, personal communication.

analyses were performed by the Soil Science Analytical Service Laboratory,

University of Florida.

Soil Particle-Size Distribution

Particle-size distribution was measured by two separate methods.

After dispersing 25 g of air-dried soil in j5 sodium hexametaphosphate

(Calgon), particle-size distributions were determined using the Bouyoucous

hydrometer method as described by Day (1965). Particle-size distribu-

tion was also quantitatively measured by weight after the pretreatments

were performed in the above clay mineralogy study.

Soil Chemical Analyses


All soil extractants were analyzed for P using the Murphy and Riley

(1962) ascorbic acid reductant technique as modified by Watanabe and

Olsen (1965). This technique has been found to be ideally suited for

use in soil P determination (John, 1970, Alexander and Robertson, 1970).

Eight ml of the ascorbic acid reductant reagent were used per 50 ml

containing appropriate aliquots of either sample or standard solutions.

Standards were prepared in the same reagents as the soil extracts, at

0, 0.04, 0.1, 0.2, 0.4, and 0.6 ppm P. The transmittance was measured

at 882 mU by a Beckman B Spectrophotometer. Phosphorus concentrations

in the extracting solution were obtained from the standard curve, which

obeyed Beer's law, and from the appropriate dilution factors.

Prior to analysis of all extracts containing NH4F, it was necessary

to add 15 ml of 0.8M H3B03 per 50 ml of total volume to prevent F from

interfering with the molybdenum blue color in the P determination

(Jackson, 1958). Neither the boric acid nor the fluoroborate formed

interfered with P color development by the above procedure.

Soils sampled on June 14, 1972 from the main field experiment were

extracted with IN NH4OAc at pH 4.8 by the University of Florida Exten-

sion Soil Testing Lab. Subsequent P analysis was by the chlorostannous-

reduced molybdophosphoric blue color method as presented by Jackson


Radioactive 3aP.-Determination of 3ap was made by adding 1 ml

the extract to 5 ml of PPO-POPOP (Bray, 1960) and recording counts by

a Packard Tri-Carb Liquid Scintillation Counting System Model 314 EX

with setting at 950 V and 104 count maximum or 100 minutes for accu-

racy. Internal quenching did not pose a problem in any of the samples.

Radioactive decay of samples was corrected by use of 32P standards in

each group analyzed.

Other analyses

pH.--Soil pH determinations were made on 10-g samples of unculti-

vated Red Bay fsl and soils sampled on June 14, 1972, from the main field

experiment. A 1:1 soil to distilled water ratio was used. After 1 hour

of intermittent stirring, the pH was read potentiometrically by inserting

the glass electrode in the soil sediment and the reference electrode in

the suspension (Peech, 1965). The meter was standardized with buffer

solutions at pH 4 and 7.

Determinations of pH were also made on soil samples after H202,

NaOC1, and CDB treatments for various component removal in the above P

retention study. These readings were made using an Orion single-cell

electrode with the meter standarized as above.

Organic matter.--Soil organic matter in 0.5 g of soil dried at 105C

was determined by the Walkley-Black (1934) method as previously described.

Aluminum.--Aluminum analysis of most soil extractants was by atomic

absorption spectroscopy as described above. The exceptions were the two

sets of 32p samples extracted by 1.1 NH4F which were analyzed for Al by

the aluminon procedure by Yuan and Fiskell (1959).

Iron.--Iron determination in most soil extractants was by atomic

absorption spectroscopy as described above. The exceptions were the two

set of 32P samples extracted with 1N NH4F. These extractants were

evaporated to dryness in 50-ml breakers. To each beaker 25 ml of dis-

tilled water were added and vigorously mixed. A 5-ml aliquot of each

solution was mixed well with 1 ml of 0.02% orthophenanthroline reagent,

1 ml of 10M hydroxylamine solution, and 10 ml of 1N NH4OAc at pH 4.8.

Standards containing 0, 5, 10, 15, and 20 pg Fe and samples were read at

490 m, on a Beckman B Spectrophotometer. Iron concentrations in the

samples were obtained from the standard curve and appropriate dilution


Calcium, magnesium, and potassium.--The June 17, 1972 samples

which had been extracted by iN NH4OAc at pH 4.8 were analyzed for Ca and

K using a Beckman B spectrophotometer and for Mg by a Beckman DU by the

University of Florida Extension Soil Testing Laboratory.

Plant Chemical Analyses

Ashed samples were digested in 1N HC1 and each was brought to 50 ml

volume with IN HC1 before analysis for P, Al, Fe, Ca, K, Mg, Mn, Zn, Cu,

and Si.


Phosphorus in the digested plant tissues was determined by the

same ascorbic acid reductant technique (Watanabe and Olsen, 1965) as

previously described.

Other elements

Measurements of Al, Fe, Ca, Mg, Mn, Zn, Cu, and Si in the plant

ash were by atomic absorption spectrophotometry. Determinations of K

in these samples were by flame photometry. These analyses were per-

formed by the Soil Science Analytical Service Laboratory, University of


Fertilizer Analysis


Phosphorus concentrations in all samples of CSP, OSP, and SCSP

pellets were determined by the ascorbic acid reductant technique

(Watanabe and Olsen, 1965). Five mg of a composite sample of fertilizer

pellets analyzed were dissolved in 50 ml of IN HC1. Measurement of P

concentrations, after suitable dilution, was made by the colorimetric

procedure described above.

X-Ray diffraction

Fertilizer granules of CSP, OSP, and SCSP before and after labora-

tory tests and those recovered from field studies were subjected to X-

ray diffraction in order to identify the fertilizer compounds present.

Powder mounts of the ground and sieved, < 4.4 p, fertilizer pellets were

used to obtain the X-ray diffractograms which were run from 3 to 60

degrees 20 on a General Electric XRD-700 instrument with Ni-filtered

CuKa radiation. A proportional counter was used to detect the radiation.

Instrument settings were as previously presented, with the exception

that the range, scale expansion, and zero offset varied with the sample.

X-ray data of fertilizer compounds as presented by Lehr et al. (1967)

were used to identify the fertilizer compounds present.


The composite sample Red Bay fsl taken to 36-cm depth was found by

both the Bouyoucos hydrometer and quantitative weight methods to be

composed of 68.2% sand, 13.7% silt, and 14.2% clay. Soil organic matter

was found to be 1.14%. Weight loss upon drying to 105C was 4.55% and

was attributed to organic matter and "water of constitution" (Jackson,

1956) associated with the sesquioxides in the soil. Selected chemical

properties of this soil (Table 3) are similar to those reported from

other locations of this soil series. Although dolomite at 6.72 metric

tons/ha had been applied to this soil, the small reaction time of one

month prior to sampling explains the high acidity and low Ca levels


Clay Mineralogical Analysis

Since clay colloids of soils have been shown in the literature to

influence P retention, a detailed study of the mineralogy of the clay

fraction in this soil was undertaken. Complementary techniques of X-ray

diffraction, differential thermal analysis, and infrared analysis were

used to define the clay minerals present.

X-Ray Diffraction

X-ray diffractograms of the clay fraction of the Red Bay Ap horizon

confirmed the presence of a vermiculite-chlorite intergrade [a vermicu-

lite with aluminum hydroxyl interlayers], a true vermiculite [one without

interlayers], kaolinite, gibbsite, and quartz. Potassium saturation

Table 3. Selected chemical properties of Red Bay fsl

N KC1 NH40Ac N HC1
pH CEC Al Al Fe Ca Al Fe Ca
H20 meq/10Og ------------------- ppm ------ -------

4.90 6.58 406 665 6.75 110 3570 935 545

*At pH 4.8.

produced sharp peaks at 14.6, 7.19, 4.87, 3.57, and 3.35A. A sharp, but

skewed to the higher 2Q side, peak was found at 4.37A. The skewedness

was attributable to a quartz peak at 4.28A which became evident upon

heat destruction of gibbsite peak at 4.37A. Upon heat treatment of the

K-saturated clay to 550C, the sharp 14.6A peak was observed to shift to

a broad peak with maxima at 10.4 and 11.7A. The 14.6A shift to a 10.4A

d-spacing denoted a vermiculite without interlayers, while the shift to

11.7A signified that hydroxy-aluminum remained interlayered in the

vermiculite (Whittig, 1965). The disappearance of the 7.19 and 3.57A

peaks following heat treatment of the K-saturated clay was indicative

of kaolinite (Jackson, 1964). Gibbsite presence was confirmed in the

clay by heat destruction of the 4.87 and 4.37A peaks (Whittig, 1965).

The d-spacings of quartz at 4.28 and 3.35A were not affected by heat

treatment. The diffractogram of the magnesium-saturated, glycerol-

solvated clay showed a small 19.2A peak which could indicate montmoril-

lonite presence (Whittig, 1965).

Differential Thermal Analysis

Differential thermal analysis of the clay produced a broad endotherm

at 120C, a sharp endotherm at 310C, a smaller, but well-defined symmet-

rical endotherm at 530C, one at 970C, and an 880C exotherm. The slope

ratio of the endotherm at 530C was found to be 1.04, thereby indicating

kaolinite rather than halloysite was present (Barshad, 1965). However,

it should be mentioned that electron micrographs of this clay did show

some tubular halloysite to be present. Comparing DTA endotherms of 10

and 30 mg of standard gibbsite with the soil clay endotherms, 50.0% of

the clay was identified as gibbsite. In a similar comparison with 5 and

10 mg of kaolinite standards, it was found that 13.6% of the clay was


Infrared Analysis

The presence of gibbsite in this clay also was confirmed by infrared

analysis. In agreement with previous work with this mineral (Van Der

Marel, 1966; White, 1971), adsorptions bands due to OH groups of gibb-

site were found at 3580, 3490, 3420, and 3370 cm-1 in both the reference

gibbsite and Red Bay clay samples. The observation that the peak at

3420 cm-1 of the clay is deeper than that of the reference gibbsite

possibly is caused by kaolinite, which also contains a peak from 3420 to
3440 cm1. The Al-OH lattice vibration peaks at 1020, 800, 740, and

670 cm-1 were noted in both the reference gibbsite and the clay. It is

possible that the gibbsite peak at 1020 cm-1 was masked by the lattice

vibrations of kaolinite from 1034 to 1010 cm-1 (Van Der Marel, 1966).

The Si-0 lattice vibrations of quartz at 1080 and the "quartz doublet"

at 800 and 780 cm-1 (Farmer, 1971) were observed in the IR pattern of

this clay. Lattice vibrations of kaolinite possibly could have con-

tributed to bands recorded at 910 and 700 cm1.

Non-crystalline Components

Sequential dissolution with hot 0.51 NaOH dissolved 12.0J' of the

clay; this included some dissolved gibbsite since 10.7% of the gibbsite

reference was dissolved by the same treatment and up to 93,8 of gibbsite

has been shown to be dissolved by this treatment (Dixon, 1966). The

amount of amorphous material, after adjustment for dissolution of gibb-

site and clay-bound organic matter, was determined to be 3.2,4.

Selected Clay Properties

Clay-bound organic matter was 3.45, of the total clay weight. The

specific surface area (SA) of the clay was 76.3 ma/g, while cation

exchange capacity (CEC) was 6.58 meq/100g. The bulk of these two para-

meters was attributed to the vermiculites, amorphous material, and clay-

bound organic matter, since kaolinite, gibbsite, and quartz have quite

small surface areas and CEC values. Found to be extractable by CDB

treatment of the clay were 2770, 4290, and 216 ppm of Fe, Al, and Si,


In this clay, percentages of gibbsite, kaolinite, and clay-bound

organic matter were 50.0, 13.6, and 3.5% respectively. By use of the

observed CEC and SA values and the relative contribution of each of the

clay components to these values, the clay also contained 5.5% inter-

layered plus true vermiculite, 3% montmorillonite, 23.3% quartz, and

1.1% amorphous material. In previous mineralogical studies of this

soil by Fiskell and McCaleb (1953), Yuan et al. (1960), and Ferraz,

Robertson, and Hutton (1968), gibbsite also was reported to be the

dominant clay mineral with lesser amounts of the other above minerals

also mentioned.

Native Soil Phosphorus

Fractionation of the total native P (70.5 p.g P/g soil) in this soil

demonstrated that more than one-half of this P was held in the 0.1N

NaOH extractable or Fe-P form (Table 4). The Al-P, Ca-P, and reductant

soluble-P, as defined by Petersen and Corey (1966), accounted for lesser

amounts of native P. Greater than 50% Fe-P of the native P in other

Red Bay samples was reported by Yuan et al. (1960) and Robertson et al.


As an indication of available P, 0.03N NH4F in 0. N HC1 was used to

extract the uncultivated soil (Table 4). As defined by the extractant

Table 4. Native soil P in Red Bay fsl


Na2CO3 fusion


0. 5 NH4F

0. N NaOH

0. 5 H2SO4


0. 03N NH4F + 0.1N HC1

P removed
p.g/g soil % of total


0 0

12.1 17.2

38.6 54.8

6. 8.5

9.0 12.8
65.7 93.3

5.2 7.4

used, approximately 12 kg P/ha of the native P were available to crops.

Using this same extractant, preliminary studies determined the

effect of various shaking times and sequential extractions on P extracted

(Tables 5 and 6). Although there was variation between P extracted at

the various shaking times, the greatest difference found was due to a

marked readsorption of P released at the 48-hour extraction period.

By sequentially extracting the uncultivated Red Bay sample with 0.03N

NH4F in 0.1N HC1, successively decreasing amounts of P were released

from the soil. This sequence of extractable P demonstrated that "avail-

able P" as defined by this reagent indicates a continuing supply of P

rather than only an initial amount. Total P extracted by the five

sequential treatments was equivalent to 28 kg P/ha.

Soil Retention of Applied Phosphorus

A preliminary study revealed that the presence of commonly employed

toluene had no effect on P retention in this soil (Table 7). However,

for consistency with previous investigations of P retention, toluene

was used in these studies.

Phosphate retention data for this soil conducted at various equi-

libration times were treated according to the P-sorption isotherms of

the Freundlich and Langmuir equations (Olsen and Watanabe, 1957), the

square-root inclusion method of Gunary (1970), and the phosphate poten-

tial (PP) method of Schofield (1955). Percentage of P sorbed by the

soil decreased as P rates increased from 10 to 5,000 ig P/g soil and

increased as equilibrium times increased from 1 to 144 hours (Table 8).

Phosphorus Adsorption Isotherms

When dealing with P sorption in relation to theoretical equations,

Table 5. Effect of various shaking times on P extracted by
0.03N NH4F in 0. i HC1 from uncultivated Red Bay

Shaking time

Shaking time












P extracted*
Ig P/g soil












*Values not followed by the same letter
are significantly different.

Table 6. Effect of sequential extractions with 0.031 NH4F in
0.i1N HC1 on extractable P from uncultivated Red Bay







pg/g soil






Total 12.60

*No significance among extractable
P values was noted.

Table 7. Toluene influence on P retention by Red Bay


2 drops/50 ml





lg/g soil





P remaining in solution
after 6 days





At same P

amendments no difference between solution P was

Table 8. Phosphorus retention by uncultivated Red Bay fsl as influenced
by P rates and equilibration times

P P sorbed after
applied 1 hr 72 hr 144 hr 1 hr 72 hr 144 hr
---- -------- g/g soil ---------------------- -----

0 -0.03 -0.11 -0.01 -

10 9.84 9.88 9.97 98.4 99.8 99.7

50 49.1 49.4 49.8 98.1 98.7 99.6

75 73.0 73.3 74.5 97.3 97.7 99.3
100 95.2 96.4 98.9 95.2 96.5 98.9

150 133.0 138.0 146.0 88.6 92.1 97.3

200 166.0 181.0 191.0 83.1 90.4 95.3

500 287.0 367.0 400.0 57.3 73.4 80.0

1,000 358.0 563.0 600.0 35.8 56.3 60.0

2,000 455.0 810.0 840.0 22.7 40.5 42.0

3,500 570.0 1,045.0 1,040.0 16.3 29.9 29.7

5,000 535.0 1,150.0 1,250.0 10.7 23.0 25.0

it is well to keep in mind the work of Hsu and Rennie (1962) which

showed that data obtained from P precipitation by exchangeable Al on

resin conformed to both Langmuir and Freundlich adsorption isotherms.

Close agreement between P-fixation data and these isotherms, they

concluded, did not necessarily imply an occurrence of an adsorption

reaction in soils. In principle, as Hsu (1965) stated, precipitation

and adsorption result from the same chemical force.

The pertinent data obtained by the various isotherms defining P

sorption in Red Bay are presented in Table 9. For purposes of brevity,

the methods employed and previously defined are referred to as Freundlich,

Langmuir, and Gunary. The P-sorption rate maximum (Psrm) in the present

studies is the amount of P retained at the highest rate in each range

of P applied to the soil. In Table 9, r is the correlation coefficient

of observed and theoretical P sorption, and Psm is the P-sorption

maximum estimated by the various methods.

Phosphate retention as defined by several theoretical approaches

was highly correlated with P sorption in Red Bay soil (Table 9). However,

the correlations of some approaches with P sorption were better under

certain equilibrium times and concentrations. At the rates of applied

P from 500-5000 [g P/g soil at equilibrium time of 144 hours, the

Freundlich estimation (r = 0.9989) of sorbed P more accurately described

P sorption than did the Langmuir equation (r = 0.9893). This finding

is in agreement with the statement of Olsen and Watanabe (1957) that the

Freundlich relationship generally applied for large amounts of sorbed P.

Freundlich estimation of sorbed P was poorest at the lowest equilibrium

times and concentrations. With increase in both equilibrium time and

concentration, Freundlich description of P sorption usually improved and











* q
H 0-




-P +









r CbD 01 OS

1 cr 0 N

0 0
O 4

ho 0- 0

0 Id
Cf0 a4 a4l 04 QL
~ ~-30
i~-1 CT

0 r-i hf
OT -PIi- ~~

co ON ON i

0o c 0 \C




H 4-

r-- --i z (
OCC0 0o

14 I
r4 l 0 L4 i





ErM 4)







Cn ,






C-oo 0 -4qi C-O
Snc \

N\d --,



-4 -l O -l co
I 0n 0\\aO0


---I I

0 i-I hO
OH b

O tD

bD =1 4
hfl .H .r-H
1. 0 -4+
be r..!
O O --

P4 p
0 1 0

MHaa -1 4i
oii S6
oA A B)i TD

\O q C,- oo
co \O 0 co
O 0 O\Nc

S0 00\0
* a

(0 NO\


0 O c0\ 0\
0\ 0\ O G


H P9*

S*dH 0

h aj o a

D SlP4 0








t )

11 g







n WH

OH t*R f
)0 r-1 bO
0 b0O

bDtfl ) R

.rl 4 0 a-I
bO .H

00 4 r4 -r

o -
o*Hl ul
a4 04>-'-

CO- 01 0 NO
-l r'cn o -i
0 crc 0 -4

0 N N


I W\ -04 0
S-t a a n


d*~\0- O 0
1 -a-0 N (
c^ T-Itc 1 ( -3

O\cO (O\'
0000 N

co\co o0\

ON C\ o\ c\

M 0 C)C

0 ON 0\ )




r0000 P
Mnot- &

* 0









approached 'r' values found for the Langmuir equation.

Langmuir description of P sorption by this soil was at its best at

equilibrium concentrations less than 20 jg P/ml. Previous workers

(Olsen and Watanabe, 1957; Rennie and McKercher, 1959; Weir and Soper,

1962) also reported close agreement between P sorption and Langmuir

description of this sorption when equilibrium solution concentration

was less than 20 jg P/ml. These above workers used respective equili-

bration times of 24, 6, and 6 hours. In the present study, as equili-

bration time was increased from i to 44 to 172 hours, correlations

between P sorption and Langmuir decreased. This occurrence was attri-

buted to the fact that P sorption with time increased and that Langmuir,

having been theoretically derived for monolayer gas adsorption, did not

accurately define this increased sorption.

Only a linear Langmuir relationship was reported by Olsen and

Watanabe (1957), Rennie and McKercher (1959), Hsu and Rennie (1962),

and Gunary (1970) over equilibrium ranges up to 14 Lg P/ml. At a similar

low-P range, Syers et al. (1973) found P sorption in some Brazilian

Ultisols to be defined by two slopes of linear Langmuir relationships.

The difference between single and double Langmuir slopes, according to

Syers et al. (1973), was that insufficient data points were collected

by the previous workers in this low-P range. Syers et al. (1973) used

eight measurements, while the maximum used by the former workers was

four. The data of the present study, using from five to six points in

this low-P range, also plotted a single linear Langmuir curve at each

equilibration time of 1, 72, and 144 hours (Fig. 1). It is possible

that the two Langmuir curves reported by Syers et al. (1973) were attri-

butable to their longer equilibration time of 72 hours as compared with



200- 1 HOUR

Om 150

E 100 72 HOURS

50 144 HOURS

0 10 20 30 40

Cpg P/ml

Fig. 1. Phosphorus sorption data for Red Bay fsl
plotted according to Langmuir equation.

0.5 to 24 hours of the other studies. In the present study, as time

of equilibration was increased, deviations from Langmuir linearity

occurred at lower equilibrium concentrations (Fig. 1) and are reflected

in the first two 'r' values at each equilibrium time (Table 9). After

an hour of equilibration, linearity of the Langmuir sorption curve

extended to applied P between 17 and 34 pg P/ml. After 72 hours, the

deviation from linearity occurred between 11.9 and 19.2 pg P/ml.

Following 144 hours of equilibration, the Langmuir curve became non-

linear between 4.0 and 9.4 g P/ml. Non-linearity of the Cambai soil

of Syers et al. (1973) occurred at 2.0 ug P/ml. Phosphate-sorption

maximum of the Cambai soil was 3.5 times that of Red Bay. It seems

that two linear portions of the Langmuir curves of a soil can result

at equilibrium concentrations less than 20 tg P/ml, if equilibrium time

is at least for a period of 72 hours, or if the soil has a high P-

retention capacity of at least 150 1.g P/g soil, or a combination of

both these factors.

Although the Langmuir equation more accurately described P sorption

in Red Bay soil than did the Freundlich one at lower equilibrium concen-

trations and times, use of the parabolic equation by the method of Gunary

(1970) gave the best estimation of sorbed P at all times and concentra-

tions tested. Gunary (1970) also reported similar good fit on the 24

soils he investigated. Over all times and concentrations, his equation

accounted for more than 99.5% of the variation in P sorption by Red Bay.

Although no theoretical foundation for this equation exists, by defining

a saturation maximum, the implication was that the soil will absorb only

a given amount of P. At the various equilibrium times and concentrations,

Psm values as defined by Gunary ranged from 1.1 to 3.7 times those

obtained by Langmuir maxima. Gunary (1970) also noted for his soils

that Psm as defined by his equation was 1.4 to 2.4 times those defined

by Langmuir equation. In the Red Bay soil, though P retention defined

by the Langmuir equation increased with time, 96. 0% of all P which was

retained after 144 hours was already fixed by soil components after

only 72 hours. Phosphorus retention as determined by Psrm was almost

identical to Psm obtained from the Langmuir equation (Table 9). Over

all ranges of equilibrium concentrations and all equilibrium times,

Psrm was 93.8% of Psm by Langmuir. As compared with Psrm and Psm by

Langmuir, Psm by Gunary appeared to overestimate P retention in this


Phosphate Potential

The relevant variables obtained by treatment of the data according

to phosphate potential (PP) technique of Schofield (1955) are presented

in Table 9. As defined by Schofield (1955), PP is defined as 0.5pCa +

pH2PO4. For these data, r values are correlation coefficients of ob-
served and theoretical P sorption, and Psm is the sorption maximum

estimated by this method at the indicated equilibrium conditions. Phos-

phate potential for AP = 0 is located at the point where no P sorption

or desorption occurs. The PP data presented in Table 9 are taken from

linear regression equations employed as described in the materials and

methods section.

Variation in P sorption by Red Bay soil was best estimated by PP
values (R = 99.1) when the lowest equilibrium time and concentration

were used. Estimation of this variation by Langmuir was also observed

to be best (R2 = 99.8) at these lowest times and concentrations. As

time and concentration were increased, correlations between PP and P

sorption in Red Bay soil decreased. The poorest correlations with P

sorption obtained by any of the methods investigated were those of PP

over the widest range tested. In these instances, PP only accounted

for 71 to 76,o of the variation in P sorption. Phosphorus sorption had

been observed to be well correlated with PP at both ends of the entire

concentration range; yet, over the entire range, these correlations

decreased considerably. This occurrence was attributed to the expo-

nential increase of AP which must be divided into several linear rela-

tionships to fit accurately linear regression equations (Fig. 2).

The Psm values as defined by PP, were noted to exceed those for

Langmuir maxima by a factor of 2.4 to 8.6. At 1-hour equilibration,

the ratio of Psm by PP to Psm by Langmuir at the lowest concentrations

was found to be 2.6. With time increase to 72 or 144 hours, this ratio

remained constant. This meant that neither P sorption as defined by

PP nor that by Langmuir changed in relation to each other as time was

increased. Over the wider concentration ranges, as defined by PP tech-

nique, 89.9$ of all P retained after 144 hours was sorbed after only 1

hour. Similar values for Langmuir and Gunary were 45.9 and 25.2%,

respectively. At the concentration range of 0.16-17 4g P/ml, Psm by

Langmuir and PP at 1 hour were calculated to be 91.5 and 91.2,, respec-

tively, of that sorbed after 144 hours. Therefore, when 150 pg P/g

soil were applied to Red Bay under conditions of the present study,

a rapid initial reaction and a slower subsequent reaction were noted to

occur. As discussed in the literature review, two such reaction rates

have been observed by numerous workers in various soils under similar

laboratory conditions.

However, the low equilibrium concentration range at which Langmuir

U 1000


S 0 -7.0- 6.0 5.0 4.0 3.0 2.0 1.0

I PP= (0.5pCa+ PH2PO4)


Fig. 2. Quantity/Intensity curve of P additions'to
Red Bay fsl after 1 hour equilibrium time.