DIVISION S-3-SOIL MICROBIOLOGY & BIOCHEMISTRY

Nitrogen Transformations in Surface-Applied Poultry Litter:Effect of Litter Physical Characteristics

M. L. Cabrera,* S. C. Chiang, W. C. Merka, S. A. Thompson, and O. C. Pancorbo

ABSTRACTPassing poultry litter through a fine sieve (<0.83 mm) generates a

fine fraction that is higher in N concentration than the whole litterand cheaper to transport per unit of N. This fine fraction can bepelletized to facilitate handling, but changing the physical character-istics of the litter may change the amount of N loss or the rate atwhich N mineralizes. The objective of this work was to evaluate theeffect of physical characteristics of the fine poultry litter fraction (pel-letized or fine particles) on net N and C mineralization, NH, volatil-ization, and denitrification resulting from surface applications of thefine fraction to Cecil loamy sand (clayey, kaolinitic, thermic TypicKanhapludult) and Dothan loamy sand (fine-loamy, siliceous, thermicPlinthic Kandiudulf) soils. The soils were adjusted to 52% water-filledporosity, treated with either pelletized or fine-particle poultry litterat 30.7 g N m~2, and incubated at 25 °C for 35 d. Humidified air wascirculated over each sample (15 chamber volumes min~') and the NH,evolved was trapped in 0.025 M H2SO4. Inorganic N contents andrates of denitrification and respiration were measured at 1, 3, 7, 14,21, and 35 d after application. The physical characteristics of the litterdid not affect total amounts of net N mineralized and NH, volatilizedin 35 d. However, total denitrification losses were significantly higherfor pelletized (6.2% of the applied N in Dothan and 7.9% in Cecil)than for fine-particle litter (0.2% in Dothan and 0.8% in Cecil). Thus,surface application of pelletized litter may result in increased deni-trification losses compared with fine-particle litter.

POULTRY LITTER, a byproduct of the poultry indus-try, is a mixture of excreta, bedding material,

feathers, and some waste feed. This litter is usuallyapplied as fertilizer to cropland or pastures. Poultrylitter can be passed through a 0.83-mm (or smaller)sieve to generate a fine fraction with a higher N con-centration than the whole litter and, therefore, cheaperto transport per unit of N (Ndegwa et al., 1991). Thefine fraction can be used in this form (fine particles)or can be pelletized to facilitate application. Use of apelletized fine fraction may allow more uniform ap-plications, but it may also affect N transformationssuch as N mineralization, NH3 volatilization, and den-itrification. Nitrogen mineralization is of extreme im-portance because it converts organic N into NH^ ,which is available for plant uptake and microbialprocesses. Ammonia volatilization and denitrificationare of economic importance because they may lead tosignificant losses of N (Thompson et al., 1990; Vander Molen et al., 1990; Egginton and Smith, 1986;

M.L. Cabrera, Dep.of Crop and Soil Sciences/Institute of Ecol-ogy; W.C. Merka, Extension Poultry Science; S.A. Thompson,Agricultural and Biological Engineering Dep., Univ. of Georgia,Athens, GA 30602; S.C. Chiang, Agricultural Research and Ed-ucation Center, Univ. of Florida, Ona, FL 33865; and O.C. Pan-corbo, Massachusetts Dep. of Environmental Protection, Lawrence,MA. Received 5 Mar. 1993. * Corresponding author.

Published in Soil Sci. Soc. Am. J. 57:1519-1525 (1993).

Paul and Beauchamp, 1989). In addition, denitrifi-cation is of environmental importance because it canrelease N2O, a trace gas believed to be involved inthe destruction of the ozone layer (Crutzen, 1976; Liuet al., 1976) and in global climate warming (Wangand Molnar, 1985). The release of CO2 is also im-portant because of its possible effect on global climatewarming (Dickinson and Cicerone, 1986). The objec-tive of this work was to study the effect of physicalcharacteristics of the fine poultry litter fraction (pel-letized or fine particles) on net N and C mineraliza-tion, NH3 volatilization, and denitrification resultingfrom surface applications of the fine fraction to soil.Surface applications are common in pasture and no-till fields.

MATERIALS AND METHODSThe soil samples used in the study were collected moist from

the upper 15 cm of areas mapped as Cecil and Dothan loamysands, which are representative of the Piedmont and CoastalPlain regions of Georgia, respectively (Table 1). The sampleswere passed through a 4-mm sieve and stored at 4 °C untiluse. Just before use, the soils were wetted to 0.102 kg kg-1

with a N-free nutrient solution (100 mg Ca L-1, 24 mg MgL-1, 113 mg S L-1, 0.5 mg P L-1, and 4 mg K L-1 addedas KH2PO4, K2SO4, MgSO4, and CaSO4; pH = 7). The studywas first carried out with Cecil and subsequently with Dothansamples.

Poultry litter from a broiler house was passed through a0.83-mm sieve to obtain a fine fraction that was pelletizedusing a steam pelletizer. The fine material (hereafter calledfine-particle litter) to be compared with pelletized litter wasobtained by crushing the pelletized litter and passing it througha 0.83-mm sieve. Both pelletized and fine-particle litter werestored in plastic containers at room temperature until use. Thetotal N and C concentrations were equal in both materials (56.9g N kg-1 and 351 g C kg-1) but the inorganic N was higherin the fine-particle litter (6.48 g NH^-N kg-1 and 4.70 gNO3--N kg-1) than in the pelletized litter (4.84 g NH4

+-N kg-1

and 3.79 g NOj -N kg-1). These differences in inorganic Ncontent could have been due to a slightly higher biologicalactivity in the fine-particle litter than in the pelletized litterduring laboratory storage. The water content was 65.7 g kg-1

of dry material for fine-particle litter and 74.8 g kg -1 forpelletized litter.

The treatments consisted of pelletized or fine-particle litterapplied on the soil surface, plus a control treatment with noadditions. The experimental units were Plexiglas cylinders (4.45-cm i.d., 10 cm long) with one end plugged by a rubber stopper.An amount of soil equivalent to 150 g of dry soil was packedin each cylinder to a bulk density of 1.74 g cm-3 to achieve52% water-filled-porosiry (WFP). Bulk densities similar to thisare commonly found in the consolidated Ap horizons of thesesoils (Perkins, 1987), and WFP values close to 60% are con-sidered optimum for aerobic processes (Linn and Doran, 1984).Water-filled porosity was calculated as follows: WFP =[(gravimetric water content x soil bulk density)/total soil po-

1519

1520 SOIL SCI. SOC. AM. J., VOL. 57, NOVEMBER-DECEMBER 1993

Table 1. Selected properties of Cecil and Dothan soils.Cation-

exchangeSoil capacity

Total

Sand Silt Clay pHt ~C N~ NH^

CecilDothan

cmolc kg"1.651.72

-g kg-1- mgNkg-79.9 14.4 5.7 6.96 3.25 0.45 0.381.6 14.4 4.0 5.04 4.11 0.47 0.5

4.710.3

15 g soil/10 mL water.j Includes NOi and NO 3.

rosity] x 100, where soil porosity = 1 — (soil bulk density/2.65). In the pelletized litter treatments, three pellets (5.5-mmo.d., 7 mm long, 0.28 g oven-dry weight each) were placedon the soil surface, whereas in the fine-particle litter treatmentsan amount equivalent to 0.84 g of oven-dry material was uni-formly applied on the soil surface. The rate used was approx-imately equivalent to 319 mg N kg-1 or 30.7 g N m~ 2 (on anarea basis). The treatments were arranged in a completely ran-domized design.

Sufficient experimental units were prepared to allow for de-structive sampling of three replicates of each treatment at 1, 3(4 for Cecil), 7, 14, 21 and 35 d after application. After thetreatments were applied, each cylinder was closed with a rub-ber stopper containing inlet and outlet tubings so that a flowof humidified air (300 mL min-1 chamber-1) could be estab-lished over the surface of the soil. This flow rate achieved anair removal equivalent to 15 chamber volumes min-1, whichshould have provided optimum conditions for NH3 volatiliza-tion (Kissel et al., 1977). All cylinders were placed in anincubator at 25 °C.

During the first 21 d, the NH3 evolved from six replicatesof each treatment was trapped by bubbling the air through 50mL of 0.025 M H2SO4. During Days 21 through 35, NH3 wastrapped in three replicates only. Air was continuously circu-lated over all samples, including those whose evolved NH3was not being trapped. Sulfuric acid in the traps was changedevery 3 to 4 d to ensure complete NH3 trapping. All of theexperimental units were weighed every 2 to 4 d and, whenneeded, the soil was rewetted to the initial water content byinjecting water below the soil surface with a hypodermic needleand syringe. Dothan samples were rewetted more often thanCecil samples to avoid the fluctuations observed in the latter(Fig. I)-

At each extraction time, every sample (including those notto be extracted) was rewetted to its original water content andeach replicate to be extracted was placed inside a 0.95-L glassjar. The jar was closed with a lid fitted with a rubber septum,10% of the air volume was replaced by acetylene to inhibit thereduction of N2O to N2 (Davidson et al., 1986), and the sam-ples were incubated in the dark at 25 °C for 9 h. Air samples(5 mL) for N20 and CO2 analyses were taken at 1 and 9 hafter acetylene application. After the incubation period, 300mL of deionized water was added to the jar and the pH wasmeasured. Subsequently, 450 mL of 1.67 M KC1 was addedto achieve a final volume of 750 mL of 1 M KC1, which wasused to extract inorganic N from the soil by shaking for 30min at 120 oscillations min-1. An aliquot of the extract wascentrifuged, subsampled, and frozen until inorganic N analyseswere performed.

Nitrous oxide concentration in the gas samples was deter-mined with a Tracer 550 GC (Trametrics Analytical Div., Aus-tin, TX) equipped with an electron capture detector at anoperating temperature of 340 °C. A mixture of Ar/CH4 (95:5)was used as carrier gas at 25 mL min-1, and two 1.8-m Po-rapak Q 80/100 columns (Altech Associates, Deerfield, IL) at50 °C were used to separate N2O. Carbon dioxide concentra-tions were determined using a Varian 3700 GC (Varian Ana-lytical Instruments, Sugarland, TX) equipped with a thermalconductivity detector operating at a temperature of 200 °C.

60

50

£ 40

53 30

aPH

caW

20

50

40

30

20 DOTHAN

0 5 10 15 20 25 30 35

DAYS AFTER APPLICATIONFig. 1. Average water-filled porosity during incubation of Cecil

and Dothan soil samples (mean ± standard error).

Helium at 20 mL min-1 was used as carrier gas, and a 1.8-mPorapak Q 80/100 column at 40 °C was used to separate CO2.All values were corrected for the solubility of N2O and CO2in water, as described by Tiedje (1982). Statistical analysis ofN2O and CO2 measurements was done on log-transformed data([Iogi0 (x + 1)]. Respiration and denitrification rates betweenmeasurements were estimated by linear interpolation and rateswere integrated with the Romberg method using Mathcad 3.1(MathSoft, 1992).

The total amount of inorganic N released from the litter ateach extraction time was estimated by adding cumulative gas-eous N losses and soil inorganic N released from the litter.Cumulative gaseous C and N losses from the litter were cal-culated by subtracting the cumulative gaseous losses of thecontrol treatment from the cumulative gaseous losses of thelitter treatments. Similarly, soil inorganic N released from thelitter was estimated by subtracting soil inorganic N in the con-trol treatment from soil inorganic N in the litter treatments.The amount of organic N mineralized from the litter was cal-culated by subtracting initial inorganic N in the litter from thetotal amount of inorganic N released from the litter at eachextraction time. To express the organic N mineralized as apercentage, the amount of N mineralized was divided by theinitial amount of organic N in the litter, which was estimatedby subtracting initial inorganic N from total N in the litter.

Soil and litter materials were ground to pass a 0.149-mmsieve and their C and N contents were determined by dry com-bustion (Nelson and Sommers, 1982) with a Carlo Erba Ana-lyzer (Carlo Erba Instruments, Milan, Italy). Particle-sizeanalysis was accomplished by the pipette method (Gee andBauder, 1986). Inorganic N in the litter was extracted by plac-ing 1 g of oven-dry equivalent litter into a 50-mL centrifugetube, adding 40 mL 1 M KC1, shaking for 30 min, and sub-sampling the supernatant. Nitrite + NO3-N in extracts wasdetermined by the Griess-Ilosvay technique (Keeney and Nel-son, 1982) with previous reduction of NO3- to NO2 in a Cdcolumn, and NH4-N was determined by the salicylate-hypo-chlorite method (Crooke and Simpson, 1971). An analysis of

CABRERA ET AL.: POULTRY LITTER NITROGEN TRANSFORMATIONS 1521

Table 2. Cumulative N mineralized and CO2-C released fromfine-particle and pelletized poultry litter applied on the surfaceof Cecil and Dothan soils and incubated for 35 d at 25 °C.

Table 3. Nitrogen balance at 35 d for fine-particle and pelletizedlitter applied on the surface of Cecil and Dothan soils andincubated at 25 °C.

N mineralizedtTime

d

147

142135

Fine— % of organic

O.lb§34.6a57.8a65.5a73.5a72.5a

PelletsN ——Cecil

4.2a29.4a58.8a70.0a73.3a76.6a

CO2-C released:):Fine

—— %

l.Oa10.2a21.3a36.1a41. 4a44.7a

Pelletsof total C ——

O.lb5.4b

14.3b25.9b31.1a36.1a

Dothan137

142135

Mb17.5a51.5a57.9a57.5a60.4a

4.3a8.5b

44.5a59.0a66.4a70.8a

1.3alO.la26.3a35.0a38.4a41. 3a

O.lb2.5b

12.0b22.9b27.0b30.6b

t Organic N applied was 24.7 g N m"2 for fine-particle litter and 26.0g N m~2 for pelletized litter.

t Total C applied was 189.6 g C m-2.§ Means within a row followed by the same letter are not significantly

different according to Fisher's LSD at a 0.05 probability level.

variance was conducted for all variables measured at each ex-traction time as well as for cumulative values of N and Cmineralized, NH3 volatilized, and N denitrified (SAS Institute,1985).

RESULTS AND DISCUSSIONNitrogen and Carbon Mineralization

Cumulative N mineralized (expressed as a percentageof organic N) at the end of Day 1 was significantly higherfor pelletized than for fine-particle litter (Table 2). Ifcumulative N mineralized were expressed as the per-centage of total N applied, the differences between littertreatments would be 3.5% for Cecil and 2.8% for Do-than. Because these differences are somewhat similar tothe initial difference in inorganic N content between fine-particle and pelletized litter (4.5% of total N = differ-ence in inorganic N/total N x 100), it is quite possiblethat the extra N mineralized from the pelletized littercorresponded to an organic N pool that had already min-eralized (before starting the experiment) in the fine-par-ticle litter. The mineralization of N from that pool in thepelletized litter was accompanied by a very small releaseof CO2 (Table 2).

By the next sampling period there were no significantdifferences between litter treatments in cumulative Nmineralized in Cecil soil. However, fine-particle litterhad released a significantly higher proportion of the or-ganic N (17.5%) than pelletized litter (8.5%) in Dothansoil. This difference between litter treatments in Dothansoil could have been due to a lower enzymatic activityin Dothan than in Cecil soil, probably a result of thelower pH of Dothan soil (Table 1). A significant pro-portion of the readily mineralizable N in poultry litter isbelieved to consist primarily of uric acid, which is con-verted to urea by many aerobic bacteria (Schefferle, 1965).Urea is in turn hydrolyzed to NH^ and HCOj by theenzyme urease (Kissel et al., 1988). A low soil enzy-matic activity would tend to slow down the mineraliza-tion of the litter organic N, the effect being more noticeable

SoilSoil Litter NH4

+ NO3-f Denit.t Volat.§ Total

Cecil

Dothan

Fine-particlePelletizedFine-particlePelletized

0.3a#O.la2.3a4.2a

- % of N applied in16.5a 0.8b12.0b 7.9a16.5a15.3a

0.2b6.2a

lilted -60.3a60.2a49.2a49.6a

77.9a80.2a68.2a75.3a

t Includes NO-2 and NOi.t N lost through denitrification.§ N lost through NH, volatilization.H Total N applied in litter was 30.7 g N m~2.# For each soil, means within a column followed by the same letter

are not significantly different according to Fisher's LSD at a 0.05probability level.

when there is no intimate contact of the litter with soil,as in the case of pellets. Nevertheless, the differencebetween litter treatments decreased rapidly and by theend of the study both types of litter had similar propor-tions of their organic N mineralized in both soils (Table2).

The proportion of the litter organic N that was min-eralized during the 35 d of the study varied between 60.4and 76.6%. Sims (1986) found that between 40 and 60%of the added organic N was mineralized within 90 to 150d after poultry litter was mixed with samples of Evesboroloamy sand (mesic, coated Typic Quartzipsamment). Ina similar study with 20 poultry litter samples mixed withKalmia sandy loam (fine-loamy over sandy or sandy-skeletal, siliceous, thermic Typic Hapludult), Bitzer andSims (1988) found that, on the average, 66.6% of theorganic N was mineralized in a 140-d incubation at 23 °C.

It should be noted that our calculations of N miner-alized from litter treatments do not include losses throughNO emission. If it is assumed, however, that NO lossescorrespond to approximately 1.7% of the nitrified N(Tortoso and Hutchinson, 1990), it can be concluded thatNO losses must have amounted to < 0.3% of the N ap-plied because the maximum amount of nitrified N (es-timated from accumulated NOj and denitrification losses)represented 14.8% of the applied N (data not shown).

Similar to the results obtained with cumulative N min-eralized, at the end of the study there were no significantdifferences between pelletized and fine-particle litter inthe percentage of the litter N in inorganic form (presentin the soil and evolved in gaseous form); the values var-ied between 68.2 and 80.2% (Table 3). Westerman etal. (1988) found that 59 to 66% of the N applied in threepoultry litter samples was in inorganic form after a 39-wk incubation at 25 °C, whereas Hadas et al. (1983)reported that 42 to 50% of poultry litter N was in inor-ganic form after 60 to 90 d of incubation at 25 °C. Partof the differences with our results may be due to differ-ences in poultry litter composition and incubation time,and to the fact that the values reported by Westerman etal. (1988) and by Hadas et al. (1983) do not includegaseous losses of N.

The emission of CO2 was initially faster with fine-particle litter than with pelletized litter (Table 2). Thiseffect was probably caused by the larger surface areaexposed by the fine particles, compared with pellets.

1522 SOIL SCI. SOC. AM. J., VOL. 57, NOVEMBER-DECEMBER 1993

O FINE PARTICLES• PELLETIZEDV CONTROL

10 15 20 25 30 35 0 10 15 20 25 30 35

DAYS AFTER APPLICATIONFig. 2. Cumulative NH, loss and NHj-N concentrations in Cecil and Dothan control soil samples, and in samples that received

surface applications of fine-particle and pelletized poultry litter and were incubated at 25 °C for 35 d (mean ± standard error).

Differences between litter types tended to disappear withtime in Cecil soil, but still persisted at the end of thestudy in Dothan soil (Table 2). The total amount of CO2-C released represented 30.6 to 44.7% of the litter C,which agrees with values obtained by Gale and Gilmour(1986) in a 34-d incubation (25 °C) of poultry litter sur-face applied to samples of Captina silt loam (fine-silty,siliceous, mesic Typic Fragiudult).

Ammonia VolatilizationInitial NH^-N in the litter amounted to 8.5% of the

total N in pelletized and 11.4% in the fine-particle litter(see above). This difference could be noticed in theNH^-N content of the soil 1 d after application, whenlitter mineralization and NH3 volatilization were still pro-ceeding at a relatively low rate (Fig. 2).

In general, NH3 volatilization was faster from fine-particle than from pelletized litter in both soils, but thedifference between materials persisted longer in Dothansoil. Part of the faster volatilization from the fine-particlelitter may have been due to its higher initial NH^-Ncontent, although this alone cannot explain the magni-tude of the differences in NH3 volatilized. A significantproportion of the faster volatilization from the fine-par-ticle litter was probably due to the faster N mineraliza-tion observed during the first few days, particularly inDothan soil (Table 2).

Between 66 and 80% of the total NH3 volatilized waslost during the first week and NH3 volatilization waspractically complete by Day 14 in both soils. The totalamount of NH3 lost was similar for pelletized and fine-particle litter, and amounted to =50% of the applied Nin Dothan and =60% in Cecil (Table 3). In contrast,Hadas et al. (1983) found that more inorganic N was

lost from poultry manure pellets than from ground ma-nure when mixed with soil and incubated at 14 or 35 °Cfor 90 d. The amounts lost from pelletized manure rep-resented 12% of the litter N at 14 °C and 22% at 35 °C.They did not measure gaseous losses but hypothesizedthat inorganic N may have been lost in the form of NH3because the decrease in soil NHf —N observed in theirsamples was not accompanied by a parallel increase insoil NOj -N. The higher N loss from the pelletized litterin their study was apparently due to the fact that thepelletizing process increased the rate of mineralizationof the litter; their ground material had not been exposedto the pelletizing process and, therefore, released N at aslower rate. Because in our study both pelletized andfine-particle litter had been exposed to the pelletizingprocess, our results mainly reflect the effect of the phys-ical characteristics of the litter. Our results indicate thatthe physical characteristics of the pellet did not affecttotal NH3 volatilized.

The high NH3 losses measured in this study were duein part to the optimum environmental conditions for NH3volatilization (25 °C, high air flow rate) and to the chem-ical characteristics of the soils used. Both soils have verylow cation-exchange capacity (Table 1) and, therefore,a low capacity for adsorbing NH^ ions. In addition, bothsoils have low H+ buffering capacity and presented alow resistance to the increase in pH caused by the hy-drolysis of urea from the litter (data not shown). Anincrease in pH causes a larger proportion of the ammo-niacal N to be present in the form of NH3, which wouldtend to enhance volatilization (Koelliker and Kissel, 1988).The pH of litter-treated samples decreased to values sim-ilar to the control treatment between Days 7 and 14 inCecil soil, and between Days 21 and 35 in Dothan soil(data not shown).

CABRERA ET AL.: POULTRY LITTER NITROGEN TRANSFORMATIONS 1523

3.0 -

»2.0

— 1.5

S 1-0K. 0.5

H5 o.oa 70

-P 60i t** 50

1 40

sT 3020

"o"S3 10

•3- °

O FINE PARTICLES• PELLETIZEDv CONTROL

- CECIL

I I 1 I I I

- DOTHAN

5 10 15 20 25 30 35 0 10 15 20 25 30 35

DAYS AFTER APPLICATIONFig. 3. Denitrification rates and (NOj + NOj)-N concentrations in Cecil and Dothan control soil samples, and in samples that

received surface applications of fine-particle and pelletized poultry litter and were incubated at 25 °C for 35 d (mean ± standarderror).

The total amount of NH3 lost was lower in Dothanthan in Cecil soil, probably partly due to the lower pHof Dothan soil. In addition, Cecil samples showed moreevaporative water losses (Fig. 1), which may have en-hanced NH3 volatilization by increasing the concentra-tion of NH| in the soil solution (Kucey, 1988).

The NH3 volatilization observed (50 and 60% of ap-plied N) indicates a large potential for NH3 loss whenfine poultry litter is surface applied in these soils. Datafrom other studies with poultry litter with which to com-pare these results are scarce. Giddens and Rao (1975)found that surface application of poultry litter (=15 g Nm~2) to a Cecil sandy loam resulted in 11% of the ap-plied N lost as NH3 during 28 d of incubation at 28 °C.The much lower NH3 volatilization observed in their workwas probably due to the lower air exchange rate used(=0.08 chamber volumes min"1) compared with the 15chamber volumes min"1 in our study.

A few studies have measured NH3 volatilization fromfresh poultry manure, but it should be noted that theresults from those studies are not directly comparable tothose obtained with poultry litter because fresh poultrymanure has higher water and ammoniacal N contentsthan poultry litter. In addition, fresh manure does notcontain bedding material (Gale et al., 1991). Crane etal. (1981) surface applied fresh hen manure at rates of22 to 105 g N m-2 to samples of Norfolk loamy finesand (fine-loamy, siliceous, thermic Typic Kandiudult)and Davidson clay loam (fine-loamy, mixed, mesic AquicCalciustoll) soils. The estimated NH3 losses (expressedas a percentage of the N applied) after 5 d of incubationat 24.5 °C ranged from 58 to 75% for Norfolk soil, andfrom 49 to 70% for Davidson soil. Wolf et al. (1988)surface applied hen manure at 43.5 g N m~2 to samples

of Bowie sandy loam that were subsequently incubatedat 23 °C for 10 d. They circulated air over the samples(=2 chamber volumes min-1) and measured a total lossof 37% of the N applied. In an 11-d field study on aCaptina silt loam, Wolf et al. (1988) also found that 37%of the surface-applied N in hen manure was lost as NH3with applications of 18.4 and 36.7 g N m~2.

DenitrificationThe rates of denitrification from poultry-litter-treated

soils reached a maximum during the first week and de-creased thereafter (Fig. 3). The same trend was observedfor rates of C02 emission (data not shown), indicatingthat the availability of easily decomposable C createdconditions appropriate for denitrification. In general, therate of denitrification from the pelletized litter was sig-nificantly larger than that from fine-particle litter duringDays 3 to 7. As a result, the concentration of soil(NOj + NO3-)-N in the pelletized litter treatment de-creased significantly during the first week of incubation(Fig. 3). In contrast, soil (NOj + NOf)-N concentra-tions in the fine-particle litter treatment remained rela-tively stable until nitrification started, =1 wk afterapplication.

The high rate of denitrification observed with pelle-tized litter may have been due to the development ofanaerobic conditions caused by high respiration rateswithin the pellet. The pellets absorbed water immedi-ately after they were placed on the soil surface, whichmay have enhanced microbial activity and decreased gasdiffusion within the pellet. Although the initial rates ofCO2 emission tended to be higher for fine-particle thanfor pelletized litter (data not shown), slow diffusion of

1524 SOIL SCI. SOC. AM. J., VOL. 57, NOVEMBER-DECEMBER 1993

CO2 out of the pellet and slow movement of 02 into thepellet may have created anaerobic microsites appropriatefor denitrification (Parkin, 1987). In addition, the pel-letized litter contained NOj and easily available C, whichare the other two requirements needed for denitrifiers tofunction adequately (Tiedje, 1988).

Integration of the rates of denitrification for the 35 dof the study indicated that samples with fine-particle lit-ter lost significantly less N than those with pelletizedlitter (Table 3). It should be noted that the integration ofdenitrification rates may have overestimated true cu-mulative values because denitrification rates were mea-sured after the samples were rewetted to their initial watercontent. Differences in denitrification rates before andafter rewetting the samples may have been larger forCecil than for Dothan soil because of the larger WFPfluctuations in Cecil samples (Fig. 1).

The low N losses found with fine-particle litter (0.2%of the applied N for Dothan and 0.8% for Cecil) agreewith results of Aulakh et al. (1991), who measured neg-ligible denitrification when crop residues were appliedon the surface of Nicollet loam (fine-loamy, mixed, mesicAquic Hapludoll) samples (60% WFP) and incubated at25 °C for 35 d. Apparently, the high CO2 emission ratesobserved in the fine-particle litter (data not shown) werenot sufficient to generate adequate conditions for deni-trification.

The N losses found with pelletized litter (6.2% of theapplied N in Dothan and 7.9% in Cecil) are comparableto the initial NO3~-N in the litter (6.7% of total N), whichsuggests that most of the N lost may have been initiallypresent in the litter. The proportion lost during the firstweek amounted to 45% of the total N loss in Dothan and60% in Cecil.

The relatively high denitrification losses observed withpelletized litter at a relatively low WFP (52%) are par-ticularly interesting because denitrification losses areusually negligible below 60% WFP (Linn and Doran,1984; Aulakh et al., 1991). Rice et al. (1988) also foundrelatively large denitrification rates from surface appli-cation of an organic waste to samples of Aubarque loam(coarse-loamy, mixed [calcareous] mesic Aerie Hapla-quept) adjusted to =43% WFP. These results emphasizethe importance of O2 consumption as a factor regulatingthe relative degree of anoxia in an environment, andindicate that significant denitrification losses may occurat <60% WFP when the characteristics of a surface-applied organic material favor the development of ade-quate conditions for denitrification.

CONCLUSIONSIn conclusion, the physical characteristics of the poul-

try litter fine fraction (pellets or fine particles) are notlikely to affect net N mineralized and NH3 volatilizedfrom surface applications to soil. But, surface applica-tion of pelletized litter may lead to lower CO2 losses andlarger denitrification losses than surface applications offine-particle litter. The decision on whether to use pel-letized or fine-particle litter will depend on whether thecapability to achieve more uniform applications with pel-letized litter can offset potentially greater N losses throughdenitrification from the pellets.

ACKNOWLEDGMENTSWe thank John Rema for laboratory assistance. This work

was supported by a grant from the USDA Low-Input Sustain-able Agriculture Program.

CABRERA ET AL.: POULTRY LITTER NITROGEN TRANSFORMATIONS 1525

Myths and Science of Soils of the TropicsSSSA Special Publication Number 29

There are several misconceptions about soils of the tropics. These misconceptions and myths are based on inadequate informa-tion on principal soils of the regions, interaction between soils and prevalent climate, soil physical and mineralogical properties,soil chemical and nutritional characteristics, soil biota and their effects on productivity. Myths are propagated by perpetual foodcrisis, agrarian stagnation, severe problems of soil and environmental degradation and resultant economic and socio-political instability.

It is time that myths regarding soils of the tropics are replaced by scientific realities. We need to strengthen the database sothat land capability can be assessed, ecologically compatible soil and crop management systems can be developed and validated,and long-term planning can be made to adopt strategies for sustaining agricultural growth and preserving productive potential ofthe soil resource. It is these concerns that led to the publication of Myths and Science of Soils of the Tropics.

Myths and Science of Soils of the Tropics. R. Lal and P. A. Sanchez, editors. Published by the Soil Science Society of Americaand American Society of Agronomy. SSSA Special Publication Number 29. Softcover, 204 pages, 1992. ISBN 0-89118-800-2. Price:$24.00 (members' first copy $20.00).Please send me ________Society Membership NumberMethod of payment: ____

copy(ies) of Myths and Science of Soils of the Tropics.

Card NumberSignature __Name ____Address ___

_Check or money order enclosed ____Credit card (check one): D Visa D MasterCard______________ Expiration Date __

_Bill me ($2.00 handling charge)D Discover

City ______Zip/Postal Code

State/ProvinceCountry

All payments must be in U.S. funds drawn on a U.S. bank or add $20 U.S. to the total amount due. Advance payment and 10 percentper book for postage is required on all orders outside the United States. Wisconsin residents add appropriate sales tax. Send yourorder to: SSSA, ASA Headquarters Office; Attn: Book Order Department; 677 South Segoe Road; Madison, Wisconsin 53711-1086 USA.

FAX Your Order 24 Hours a Day 608-273-2021