Aggregate Formation in Soils with Special Referenceto Cementing Substances1

E. M. KROTH AND J. B. PAGE2

MANY investigations (i, 2., 3, 5, 25)3 have shownthat organic matter is an important factor in-

fluencing the aggregation of soils. The microorgan-isms using this organic matter as a source of energyhave been considered by some to be the responsibleaggregating agents. Kanivetz and Korneeva (9)showed that bacterial contamination increased soilaggregation and Peele (22) prepared water-stableaggregates from quartz sand with bacterial mucous.Martin and Waksman (15) found fungus myceliahad a binding effect on soil particles. Martin andAnderson (16) also showed that fungi would formsoil aggregates.

A direct binding action of clay particles by polarorganic materials was found by Sideri (24) andMyers (20). Ensminger (7) and Ensminger andGieseking (6) showed that montmorillonitic claysabsorbed proteins within the crystal lattice and theseproteins to some extent were shielded from the actionof bacterial enzymes. These investigations indicatedthat micoorganisms may be more important as in-direct agents in being responsible for the production,through the decomposition of fresh organic matter,of polar molecules which act as direct binding agentsof the soil particles through physico-chemical bonds.This assumption was strengthened by Martin (12)who found that fresh materials caused greater aggre-gation than composts when incorporated with the soil.McCalla (17) reported similar results with fullydecayed straw. The investigation of Myers and Mc-

' Calla (21), in which organic materials were, in-cubated in contact with soil materials, showed thatmaximum aggregation lagged behind maximum mi-crobial activity. This also indicated that metabolicproducts were more influential as binding agentsthan the microoganisms themselves.

At the time this study was undertaken the bene-ficial effects of organic matter as regards soil aggre-gation were attributed to the direct binding action ofthe microorganisms and to the .action of polar mate-rials produced by them as metabolic products. Thelocation of these substances with respect to the in-terior or exterior of aggregates had not been deter-mined, although Kubiena (10) stated that a humiclayer formed on or near the aggregate surfaces, thusforming a protective shell. These metabolic productswere not described in the literature other than aspossessing di-pole moments. It was the purpose ofthis study to investigate aggregation processes in aneffort to characterize further the active organic sub-stances and to locate their position within the ag-gregates.

.During the course of this investigation work wasreported by Martin (13, 14) and McHenry and Rus-sell (19) which showed that products of microbialactivity were effective aggregating agents. Martin(13) identified some of them as microbially synthe-sized polysaccharides. Also a study by McCalla (18)showed that gums, fats, and waxes were capable ofbinding soil particles into water stable, secondarygroups.

The investigation reported in this paper was di-vided into three parts, vis., electron microscopestudies of natural and synthetic microaggregates, soilstabilization by decomposition products of varioustypes of organic matter, and a study of selectednatural aggregates by chemical and physical means.

ELECTRON MICROSCOPE STUDIESThe electron microscope* was used in an attempt to dis-

cover the nature and location of the binding agents presentin microaggregates and whether there was any definiteorientation of particles.

Natural aggregates were obtained from a Brookston siltyclay loam from Ohio and a Davidson clay from North Caro-lina. Small portions of these soils were shaken in distilledwater, allowed to settle for 12 hours, and a few millilitersof the suspensions removed from which specimens were pre-pared.

Synthetic aggregates were produced by flocculating ben-tonite with dilute solutions of iron and aluminum chloridesor suspensions of gelatin and humus preparations. The humuswas secured from composts of alfalfa, corn stover, and a i: imixture of corn stover and alfalfa, by electrodyalizing nitratesfrom portions of these materials which had been soaked inwater for 3 days. The bentonite suspension (particles 100 to160 //.) which received Fe+++ ions had a pH of 3.8 and thesuspension which received the Al+++ ion, a pH of 4.2. Theselow pH values were needed to prevent the formation of ironand aluminum hydroxides. The flocculated material waswashed twice with distilled water, dried at 45° C, and placedin distilled water. The bentonite-gelatin system was loweredto pH 4.7 before flocculation occurred. The bentonite-humussystems did not flocculate over a wide pH range, but, onevaporation to dryness, water-stable aggregates were formed.A kaolinite-humus system did not produce stable aggregateswhen similarly dried.

All specimens for the electron microscope were preparedby the method described by Humbert and Shaw (8).

NATURAL BROOKSTON AGGREGATES

Fig. i — i, 2, 3, are electron micrographs of,natural Brookston aggregates. Organic matter ispresent throughout, coating the individual clay crys-tals. No particular type, of crystal orientation is inevidence. The organic matter does not appear incapsular form but is thought to be combined with theclay particles by a physico-chemical bond and notby physical interlacing of different kinds of material.

'Contribution from the Department of Soils, Ohio State University, Columbus Ohio

?o°"LESr cSdV?as00™611 UniVerSit7'IthaCa' "• Y

'Electron Microscope Radiation Laboratory, Ohio State University, Columbus, Ohio, Dr. A. P. Prebus in Charge.

SOIL SCIENCE'SOCIETY PROCEEDINGS 1946

SYNTHETIC AGGREGATES —— GELATIN

Gelatin was used since it was thought to resemblethe nitrogenous materials produced during the de-composition of complex proteins in plant residues.Nitrogenous substances decompose rapidly in thesoil and as rapidly produce a binding effect on soilparticles. Since Ensminger and Gieseking (6) found ,gelatin to be adsorbed by bentonite clays there ap-peared to be a definite relationship between it andthose substances which cause aggregation when or-ganic matter decays in the soil.

In a bentonite-gelatin system above a pH of 4.7there was no flocculation. Below this value wheregelatin had a positive charge flocculation occurredrapidly. Fig. 3 — 2 shows a gelatin-bentonite floc-cule made into a specimen without previous drying.TJie crinkled effect is attributed to rapid dehydrationin the specimen chamber of the microscope which isunder a vacuum of io~3mm of mercury. As water waswithdrawn the plates were pulled closer togethercausing buckling of these thin crystals. The gelatinis thought to be distributed between the plates. Thedense appearing areas are interpreted as being ac-cumulations of gelatin.

Fig. 3 — i shows a gelatin-bentonite aggregatethat was dried before being prepared as a specimen.Slow drying permitted the crystals to orient them-selves so as to eliminate buckling. Again the gelatin

FIG. i.—Brookston electron micrographs, i, 2, and 3, naturalaggregates; 4, soil minerals after treatment with hydrogenperoxide. White line represents ip.

Fig. i, 4 'shows no evidence of aggregation afterremoval of organic matter by hydrogen peroxide.

NATURAL DAVIDSON AGGREGATES

Fig. 2 — I, 2, 3, shows the natural Davidson ag-gregates to be different from those in the Brookstonsoil. The aggregates are smaller and crystal edgesare shown with less detail. Individual crystals areapparently • heavily coated with iron and aluminumoxides. Removal of these oxides by the method ofDrosdoff (d), Fig. 2 — 4, destroyed the aggregates,thus indicating that aggregation in this soil is due tothe physical binding forces of these compounds.

SYNTHETIC AGGREGATES —— IRON AND ALUMINUM

Fig. 3 — 3, 4, shows bentonite floccules formed ,byFe+++ ions and Al+++ ions and stabilized by drying.The crystal plates are oriented in the horizontalplane and appear to be covered with some materialwhich exists homogenously throughout the aggre-gate. This material is interpreted as being hydrousiron and aluminum oxides acting as the bindingagents. During flocculation Fe+++ and Al++* ions wereadsorbed in excess of the base exchange capacity asexplained by Lutz (n). On dehydration the oxidesformed, covering each crystal and binding themphysically in a manner similar to the iron and alumi-num oxides in the Davidson soil.

FIG. 2.—Davidson electron micrographs, i, 2, and 3, natural' aggregates; 4, soil minerals after removal of free iron and

aluminum oxides. White line represents i/i.

KROTH AND PAGE : AGGREGATE FORMATION IN SOILS

FIG. 3.—Electron micrographs of synthetic aggregates. I,Na-bentonite, 100 ,̂ flocculated with gelatin and evaporatedto dryness; 2, H-bentonite, ioo-i6o/i, flocculated with gela-tin, not evaporated; 3, Na-bentonite, IOG/J, flocculated withFeCls, dried at 45° C; 4, H-bentonite, 100-160 ,̂ flocculatedwith AlCls, dried at 45° C. White line represents i/j,.

is distributed throughout the aggregate preventingredispersion by physico-chemical forces.

around soil particles was obtained. It is concludedthat organic matter is effective in aggregation not byphysically surrounding and holding soil particles butrather by changing the cohesive forces between soilparticles a"s a result of its adsorption.

SOIL STABILIZATIONWhen fresh organic matter is allowed to decom-

pose in the soil, the effective aggregating agents maybe microbial mucous, mycelia, polar active syntheticand decomposition products, and more resistant sub-stances such as ligno-proteins, waxes, fats, and resins.This portion of the study was undertaken to singleout the more active agents.

A Miami silt loam containing 2% organic matterwas highly dispersed by- running it through a MikroPulverizer from which all screens had been removed.Four-thousand-gram samples of this soil were thor-oughly mixed with 80 grams of pulverized fresh al-falfa, fresh corn stover, composted corn stover, anda composted I : i alfalfa-corn stover mixture, respec-tively, moistened, and incubated for i month incolanders at 28° C. The samples were initially mois-tened capillarity by placing the colanders, in a shallowpan of water. They were air dried and rewet in the

SYNTHETIC AGGREGATES —— HUMUS

The humus suspensions' were used to representpolar active substances occurring in the more re-sistant portions of the soil organic matter. Fig. 4 —I, 2, 3 shows aggregates prepared from bentonite andcolloidal humus. They appear to be similar to thoseprepared with other aggregating agents. The crystalsare in the horizontal plane and all the organic matterappears to be homogenously distributed throughoutthe 'aggregates. Physico-chemical bonds are con-sidered to be the binding forces. Where mutualphysico-chemical forces are not present only a phy-sical mixture results as is shown in Fig. 4 — 4.Kaolinite-humus systems did not produce stable ag-gregates because of the low surface activity of thisclay mineral.

This part of the study gives evidence that aggre-gating agents are homogenously distributed through-out the aggregate mass and that polar organic sub-stances act as binding agents of soil particles possess-ing surface activity through physico'-chemical bonds.No evidence that organic matter exists as a capsule

FIG. 4.—Electron micrographs of synthetic aggregates. I, H-bentonite, 100—i6o/u, evaporated at 45° C with alfalfa hu-mus; 2, H-bentonite, ioo-i6o/u, evaporated at 45° C withcorn humus; 3, H-bentonite, unfractionated, evaporated at45° C with alfalfa-corn humus; 4, ungractionated H-kaolin, evaporated at 45° C with alfalfa humus. White linerepresents IM.

SOIL SCIENCE SOCIETY PROCEEDINGS 1946

same manner three times during the period. A checksample was similarly treated except for the additionof organic matter.

After incubating 3 days the fresh alfalfa had pro-duced a profuse growth of mold on the surface of thesoil. At the end of 5 days the relative mold growthfor the group was fresh alfalfa, heavy; fresh corn,light; composted corn, very -light; composted corn-alfalfa, very light; check, trace.

Fig. 5 — 6 gives the appearance of the samples atthe close of the incubation period. The cracks shownappeared when the soil was first moistened. Fig. 5 —6, parts 3, 4, 5, show less cracking than the check.Fig. 5 — 6, part 2, which contained the fresh alfalfa,shows marked binding effects due to water-solubleaggregating substances, as explained by Martin (12).

STABILITY TESTS

At the close of the incubation period the sampleswere air dried and their stability tested by placingmedium sized clods in distilled water and slaking for12 hours. Fig. 5 — 1-5 shows the results. The differ-ences appeared immediately. Fig. 5 — 4 and 5 showthat the soil treated with fresh corn stover had a

Fiq. 5.—Slaking tests on a treated Miami soil, i to 5 showcondition after 12 hours in distilled water; i, check; 2,composted,corn; 3, composted corn-alfalfa; 4, fresh corn;S, fresh alfalfa.Six shows the same soil after incubating for I month, i,

check; 2, fresh alfalfa; 3, fresh corn; 4, composted corn-alfalfa; 5, composted corn.

slight increase in stability over that receiving thefresh alfalfa.

An additional test of stability was made by placingduplicate samples of large and small clods in theupper screen of a wet-sieving aggregate analysis ap-paratus (26) and slaking for 30 minutes by sub-merging the screens in water. The clods were thenwet sieved for 30 minutes and the quantities remain-ing on the upper 5-mm screen determined. The re-sults are given in Table i. These data show that thestability pictured in Fig. 5 — 4 and 5 is real. A great-er stabilizing effect for fresh corn stover is againnoted over that of fresh alfalfa. The clods from thecheck sample and those receiving the compostedmaterial completely disintegrated.TABLE i.—Results of wet sieving clods from soil incubated with

different types of organic matter added at the rate of 2% of dryweight of soil.*

Treatment

% of clods 2 in. ormore in cross

% of clods yi to iin. in cross sec-tion. . . . . . . . . . .

Check

0.0

Freshcorn

stover

98.7

80.6

Com-postedcorn

2.3

o.o

Freshalfalfa

77.6

73-8

Com-postedcorn-.

alfalfa

1-5.

o.o*Samples slaked for 30 minutes in distilled water and wet sieved for 30

minutes. Figures are percentage of sample' remaining on 5-mm screen.

DISCUSSION OF STABILITY TESTS

These stability tests show that the decomposition'offresh organic matter in-close proximity to soil par-ticles has a binding effect on Miami silt loam. Martin(14) considered the agents responsible for soil aggre-gation resulting from applying readily decomposiblematerial'to a soil to be of four types, viz., (a) cellsand filaments of micoorganisms, (b) water-solublesubstances in original materials, (c) decompositionproducts of microbial metabolism, and (d) productsof microbial synthesis. Examination with a light

. microscope revealed no gummy areas or filaments.The original water-soluble materials would seem tohave been destroyed after i month. Therefore, theeffective agents are considered to be products direct-ly concerned with microbial activity. Many of thedecomposition products are expected to be polar sub-stances resulting from enzymatic action on complexproteins occurring in the fresh alfalfa and to a' lessextent in the corn stover. Products of microbialsynthesis are soil binders also (14) and some havebeen identified as polysaccharides. These polysac-charides, though polar, do not contain nitrogen as amolecular component. The possibility therefore ex-ists that the binding agents in the soil receiving thealfalfa were predominately gelatin-like while those inthe soil treated with fresh corn stover resembledpolysacchatides.

The composted materials were ineffective in pro-ducing aggregation since they did not serve as energymaterial for organisms, being themselves residues of

KROTH AND PAGE : AGGREGATE FORMATION IN. SOILS

intensive decomposition. The experience with humuspreparations described above shows such substancesto be relatively inactive. They contain polar com-plexes, but their polarity is not of such a degree asto react with the less surface-active colloids, such asthose in Miami soil, to form stable aggregates. Thesamples treated with the composts were a physicalmixture of relatively inert substances. The compostsmay have contained waxes and resins, but they wereineffective when applied in this manner. This studygives strong support to the conclusion drawn fromthe electron microscope studies, vis., that organicmatter is effective in aggregation largely because ofactive polar materials produced during its decompo-sition, and that these materials are uniformily dis-tributed throughout the bound or aggregated portionof the soil.

CARBON-NITROGEN RATIOS

The work of Salter (24) showed that most agri-cultural soils have a carbon-nitrogen ratio near 10: i.When organic materials of wider ratio (such as cornstover) are added, decomposition occurs in such amanner that carbon is eliminated as CO2. On appli-cation of residues with a ratio narrower than 10: i,such as alfalfa, decomposition widens the ratio, elimi-nating nitrogen as ammonia. When residues of wideratio are added to a soil, competition for the availablenitrogen occurs between the microorganisms andcrop plants. This must be kept in mind when apply-ing organic matter to a soil to improve its physicalcondition. Because of their value in this connection,the data in Table 2 are presented.

AGGREGATE ANALYSISThe purpose of this part of the study was to in-

vestigate the nature and location of cementing agentsfound in stable aggregates. Aggregates were obtainedfrom bluegrass areas of a Brookston silty clay loamand a Miami silt loam on the University farm, Co-lumbus, Ohio. Samples were air dried and an aggre-gate fraction ranging from 3.36 to 3.96 mm wasseparated by dry sieving. Foreign matter was re-moved by hand.

One-hundred-gram portions of these aggregateswere refluxed in soxhlet extractors for 16 to 20 hours

with ethyl alcohol, acetone, and ether. Also. 100 gramswere placed in 400 milliliters of distilled water andallowed to soak for 8 hours, then removed and airdried. The organic solvents were expecte^ to removefats, waxes, and resins that might act as soil bindingagents.

STABILITY TESTS

Fifty-gram samples of extracted and untreated ag-gregates were shaken dry on a g-mesh screen, by trieshaker described below, for given intervals of time.At the end of each interval, the quantity of materialremaining on the 6-mesh screen was determined.Results are given in Table 3. These data give evi-dence that the fats, waxes, and resins soluble in or-ganic solvents are active in influencing the stabilityof selected Brookston aggregates to abrasion but haveno appreciable effect in the Miami samples. TheMiami soils are naturally well drained and oxidationhas reduced organic residues to a small amount ofresistant materials. Poor natural drainage during theformation of the Brookston soils has favored the ac-cumulation of organic matter so that sufficient quan-tities of fats, waxes, and resins are present to act asphysical binding agents in Brookston soils.

DESCRIPTION AND OPERATION OF SHAKER

, To study the chemical nature of the organic matterthroughout entire aggregates, arbitrary shells fromthese selected aggregates were obtained by abrasion.This was done by use of the shaker, illustrated inFig. 6, -so designed that the aggregates were shakenat 30 oscillations per. minute on a g-mesh wire screen.Fine material from the aggregates and any fracturedportion fell through the screen into a beaker asshown. To get successive shells, the aggregates wereshaken on a g-mesh screen and at intervals removedand sieved on a screen of the desired size. For ex-ample, the aggregates as selected would not pass a6-mesh screen. To obtain a 6-mesh shell, or shell No.i, these aggregates were shaken on a g-mesh screenand then placed on the 6-mesh sieve. Any aggregateswhich did not pass the 6-mesh sieve were returnedfor additional shaking. In this manner 6-, 7-, 8-, andg-mesh shells were obtained. These were designatedas shells, Nos. i, 2, 3, and 4, respectively. The sizesof the meshes were 3.36, 2.79, 2.36, and 1.98 mm,

TABLE 2.—Data on incubation studies (4,000 grams of soil+8o grams of organic matter).

p H . . . . . . . . . . . . . . . . . . . . . . . .

Total nitrogen added as or-

Unincubated soil

Check

5-68i tR13°

0-135

5.400

Freshcorn

6.00153

0.7306.130

posted

6.20261

2.9447.S6S

Freshalfalfa

5 ft T

193

2.165

8.4.44.

Com-postedcorn-alfalfa

6.80239

3-5658.965

Incubated soil

Check

5.6s152

0.135

5.400

Freshcorn

6-37

0.151

posted

5-62309

0.1905

v

Freshalfalfa

6-45153

6.840

Com-postedcorn-alfalfa

6.50261

SOIL SCIENCE SOCIETY PROCEEDINGS 1946

TABLE 3.—Aggregate stability after extraction with ethyl alcohol,ether, acetone, and distilled water*

Treatment

Ethyl alcohol. . . . . . . .E the r . . . . . . . . . . . . . . .Acetone . . . . . . . . . . . . .Distilled water. .......Av. of 4 check samples

Ethyl alcohol.Check I . . . . . .

Ether... .Check II.

Acetone. .Check III.

Brookston aggregates

Afteri hr.

15-017-317.623.724-5

After2 hrs.

348.96.2

15-6

After3 hrs.

0.63-01.28.18.2

After4 hrs.

o.o1.2O.O

4.0

Miami aggregates

After30 min.

15-715.0

2I.OI9.I

15-315-9

After60 min.

3-0

5-85-2

3-22.8

After90 min.

O-30-5

"0.9i.o

0.60-5

*Grams of Original 50 grams of 6-mesh aggregates remaining on 6-mesh screen after various periods of dry shaking.

respectively. The spheres which passed the g-meshscreen were considered to be the aggregate centers.Fractured portions falling into the beaker were re-'moved by a i6-mesh sieve and discarded. In thismanner only successive shells from' stable aggregateswere obtained.

NITROGEN STUDIES OF AGGREGATE SECTIONSShells and centers of untreated and extracted ag-

gregates were analyzed for nitrogen by a semi-microKjeldahl method. Samples of 200 to 300 milligramswere used. The quantity of nitrogen found was ex-pected to give an indication of the amount of protein-aceous materials present in the various sections ofthe aggregates. Results are given in Table 4.

In the case of the Brookston aggregates, data forthe check samples show no significant differences

FIG. 6.—Shaking apparatus used in nitrogen and stabilitystudies on selected aggregates.

TABLE 4.—Nitrogen in shells of Brookston and Miami aggregatesafter extraction with organic solvents.

Treatment

Percentage nitrogen

ShellNo. I

ShellNo. 2

ShellNo. 3

ShellNo. 4

Av. ofshells

foreachtreat-ment

Aggre-gate

centers

BrookstonEthyl alcohol.Ether. .'......Acetone......

Check I . . . . . .Check I I . . . . .Check I I I . . . .Av. of each un-

treated shell

0-3490.3560.353

0.3300.3370-343

0.337

0-3550.3960.369

0.3380.345

0.3480-3950.358

0.333

0-355

0-339 0.343Miami

0-355

0.356

P-3420-3340.336

0-337

0.3520.3850-359

0.3360-345

0-339

0.3360-3330-3530.3190.3230.336

0.326

Ethyl alcohol.Ether. . . . . . . .

Checkl... . ' . . .Check II . . . . .

Av. of each un-treated shell

0.213

0.201

0.2270.225

0.226

O.2O9

0.204

0.2320.225

0.228

O.2O40.2030.203

0.2350.228

0.231

0.1940.1970.205

0.228

O.224

0.205

0.203

0.229

O.227

0.1900.189

0.215

0.217

between the nitrogen content of the shells of eachsample. However, the nitrogen content of the aggre-gate centers is less by an average of 3.8% than theouter portions. This may be too small to be signifi-cant, although in every case the trend is the same.

The higher nitrogen content of the extractedBrookston aggregates than the check samples isthought to be due to the irreversible dehydration ofthe organic matter by the extracting solvent. In gen-eral, the same relationship holds here as with theuntreated • samples, viz., individual shells show noappreciable difference in nitrogen content but arehigher than the aggregate centers with the exceptionof the acetone-treated samples.

The Miami data show no concentration of organicmatter in the arbitrary shells. No dehydrating effectis noted on the organic matter. Aggregate centersshow an average decrease of 4.4% in nitrogen con-tent from'that of the shells.

The statement above on the significance of thesedata applies in this case as well.

In this study, the data showed no evidence ofshells of organic matter concentration. A probablereduction in nitrogen content within aggregate cen-ters is indicated. However, the organic matter is be-lieved to be distributed throughout both kinds ofaggregates in such manner as to exert uniformbinding forces. These results support those obtainedwith the electron microscope. If capsular organicmatter had been found, it might have been expectedthat aggregates would have shown higher concentra-tions of organic matter in the outermost shells. Since

KROTH AND PAGE: AGGREGATE-FORMATION IN SOILS 33

this was not found to be the case the original con-clusion that the binding is physico-chemical, result-ing from adsorption on crystal surfaces, is supported.

DISCUSSION AND CONCLUSIONSAll parts of this investigation give evidence that

the aggregating agents under study were distributedcontinuously throughout the aggregates and in con-tact with each particle. The electron microscopestudies gave no visible sign of capsules or coatingswhich could act in a protective capacity, but in allcases give indications that the binding agents camein contact with each crystal surface. The stabilizationstudy also showed the benefit of an intimate mixtureof soil and organic matter, for when decompositionproducts appeared they were in contact with the soilparticles and were able- effectively to produce water-stable units.

The chemical analysis of aggregate shells alsogave evidence that organic matter exists uniformlythroughout the aggregates with the possible exceptionof the aggregate centers. It is believed that the slight-ly higher concentration, found in the four arbitraryoutside layers is a luxury accumulation and does notinfluence aggregate stability to an appreciable extent.These aggregates were formed under bluegrass sodthat had been in place for several years. When thebluegrass was first established, residues of roots andtops soon accumulated and, on decomposition, pro-vided aggregating substances in sufficient quantity tobind the soil into stable aggregates. On continuedgrowth additional materials were produced and werecarried into the aggregates when in solution or as acolloidal suspension in the soil water. Since the activeareas of the clay crystals had previously been satis-fied, these additional substances became an accumula-tion which would not necessarily have a binding orprotective function.

This study has shown that, in general, there aretwo types of forces that are responsible for bindingsoil particles into secondary groups and are exempli-fied by polar organic substances, on the one hand,and by the iron and aluminum oxides, fats, waxes,and resins, on the other. The polar substances formphysico-chemical bonds with the surface-active claysthat are not easily disrupted on rehydration. Thesesubstances may be the immediate products of mi-crobial activity or portions of the more resistant andinert humus. The inorganic oxides, fats, waxes, andresins provide a continuous matrix which binds thesoil particles into secondary units by-physical forcesalone.

In soils where organic matter is the main aggre-gating agent the polar materials are probably themost important and those produced on the initialdecay of fresh organic matter are thought to be themost active in cultivated soils. This type was shownto be responsible for the binding of the Miami soilsin the second part of this study. It was also probablyresponsible for the binding which was observed in theselected Miami aggregates in the aggregate analysis

studies. In this soil the grass residues decayed rapid-ly, but some polar materials were constantly beingadded. The waxes, fats, and resins did not accumu-late sufficiently to be effective since they were quick-ly decomposed in this well-aerated soil.

In the less readily drained Brookston soil, how-ever, the waxes, fats, and resins did accumulate andexert a stabilizing effect on these aggregates, as wasfound in the aggregate analysis studies. These sub-stances would not be effective where conditions donot favor organic matter accumulation. However, themore effective agents in this soil were probably polarcompounds from the decomposition of grass residues.

The composted materials contained polar-activecomplexes that were able to stabilize bentonite butwere ineffective when incubated with the Miami soil.This further confirms the work of Martin (12) thatcomposts do not improve soil aggregation.

The superiority of polar substances produced bythe decay of fresh organic matter as the aggregatingsubstance is indicated in all parts of this study. Theyare temporary in their action and their chemical com-positions are variable, as pointed out by Martin (14)and McHenry and Russell (19). They all have oneproperty in common, the ability to form physico-chemical bonds with surface-active clay minerals. Inorder to keep cultivated soils in the best physicalcondition, a sufficient quantity of rapidly decompos-ing organic matter to produce these active compoundsneeds to be present at all times. Additional researchis needed to characterize these substances further, todetermine the amount necessary for a given quantityand type of clay, and to ascertain the best method ofassuring their presence in the soil.

SUMMARY

Investigations of natural and synthetic aggregates.and of incubation studies in which fresh and com-posted organic matter were incorporated with thesoil were made. All aggregating agents were foundto be uniformly distributed throughout the aggre-gates. Polar substances resulting from decompositionof fresh organic matter were explained as being themost effective in aggregating cultivated soils. Moreresistant humus, fats, waxes, and resins were foundto be effective also. Additional research is pointed outas necessary so that good soil management can insurea constant supply of polar-active materials in orderto keep a given soil in optimum physical condition.

34 SOIL SCIENCE SOCIETY PROCEEDINGS 1946