ARTICLE IN PRESS

Pedobiologia 50 (2006) 243—256

0031-4056/$ - sdoi:10.1016/j.

�CorrespondE-mail addr

(M.B. Postma-B

www.elsevier.de/pedobi

Earthworm species composition affects the soilbacterial community and net nitrogenmineralization

Maria B. Postma-Blaauwa,�, Jaap Bloemb, Jack H. Faberb,Jan Willem van Groenigenb, Ron G.M. de Goedea, Lijbert Brussaarda

aDepartment of Soil Quality, Wageningen University, 6700 EC Wageningen, The NetherlandsbAlterra Green World Research, Wageningen, The Netherlands

Received 5 November 2005; accepted 3 February 2006

KEYWORDSSoil;Earthworms;Biodiversity;Species interactions;Nutrients;Functional groups

ee front matter & 2006pedobi.2006.02.001

ing author. Fax: +31 0 31ess: maria.postma@gmlaauw).

SummaryKnowledge of the effects of species diversity within taxonomic groups on nutrientcycling is important for understanding the role of soil biota in sustainableagriculture. We hypothesized that earthworm species specifically affect nitrogenmineralization, characteristically for their ecological group classifications, and thatearthworm species interactions would affect mineralization through competitionand facilitation effects. A mesocosm experiment was conducted to investigate theeffect of three earthworm species, representative of different ecological groups(epigeic: Lumbricus rubellus; endogeic: Aporrectodea caliginosa tuberculata; andanecic: Lumbricus terrestris), and their interactions on the bacterial community,and on nitrogen mineralization from 15N-labelled crop residue and from soil organicmatter.

Our results indicate that L. rubellus and L. terrestris enhanced mineralization ofthe applied crop residue whereas A. caliginosa had no effect. On the other hand,L. rubellus and A. caliginosa enhanced mineralization of the soil organic matter,whereas L. terrestris had no effect. The interactions between different earthwormspecies affected the bacterial community and the net mineralization of soil organicmatter. The two-species interactions between L. rubellus and A. caliginosa, andL. rubellus and L. terrestris, resulted in reduced mineral N concentrations derivedfrom soil organic matter, probably through increased immobilization in the bacterialbiomass. In contrast, the interaction between A. caliginosa and L. terrestris resultedin increased bacterial growth rate and reduced total soil C. When all three specieswere combined, the interaction between A. caliginosa and L. terrestris was

Elsevier GmbH. All rights reserved.

7 483766.ail.com

ARTICLE IN PRESS

M.B. Postma-Blaauw et al.244

dominant. We conclude that the effects of earthworms on nitrogen mineralizationdepend on the ecological traits of the earthworm species present, and can bemodified by species interactions. Knowledge of these effects can be made useful inthe prevention of nutrient losses and increased soil fertility in agricultural systems,that typically have a low earthworm diversity.& 2006 Elsevier GmbH. All rights reserved.

Introduction

The importance of soil biota for sustainableagricultural production is increasingly acknowl-edged (Brussaard et al., 1997; Fragoso et al.,1997; Bardgett and Cook, 1998; Wardle, 2002).However, many questions remain on the impor-tance of the diversity within soil biota groups forcontinued ecosystem functioning in agriculture(Swift et al., 2004). Earthworms are known toincrease nitrogen mineralization (Blair et al., 1997;Cortez et al., 2000), and thereby can increasenutrient availability in systems with reduced humaninfluence and low nutrient status, i.e. no tillage,reduced mineral fertilizer use and low organicmatter content (Ruz-Jerez et al., 1992; Doubeet al., 1997; Brown et al., 1998; Brown et al., 1999;Cortez and Hameed, 2001). The effect of earth-worms on nitrogen mineralization and crop produc-tion may, however, depend on earthworm speciesand species interactions present in the system(Brown et al., 1999).

Earthworm species have been found to interact bycompetition for food (Abbott, 1980; Dalby et al.,1998; Lowe and Butt, 1999; Baker et al., 2002; Loweand Butt, 2002) and by affecting each other’s burrowsystem (Capowiez et al., 2001; Jegou et al., 2001).These interactions directly affected the abundancesof the earthworm species populations, but mayindirectly also affect nitrogen mineralization andcrop production. There are indications that anincreased species diversity of earthworms has apositive effect on crop production (Brown et al.,1999). Few studies, however, have directly addressedsuch effects of interactions between earthwormspecies on ecosystem processes (Wardle, 2002).

Earthworms can be classified as epigeics (living inlitter or top soil layers where they forage primarilyon plant residues), anecics (living in permanentdeep vertical burrows in which they store plantresidues that are collected from the soil surface)and endogeics (living in the soil and foraging on soilorganic matter) (Bouche, 1977). Competition forfood may lead to resource partitioning and a highertotal abundance of the competing earthwormspecies compared to the single-species community,

cf. general ecology (Begon et al., 1990) and datafor nematodes (Postma-Blaauw et al., 2005). Thismay in turn result in increased incorporation ofcrop residues into the soil. Alternatively, byincorporating organic matter into the soil, epigeicand anecic species may provide an extra foodsource for endogeic species (Hughes et al., 1994;Jegou et al., 2001), and thus stimulate the feedingactivity of endogeic species. Earthworms affectnitrogen mineralization through direct and indirecteffects on the microbial community. By incorporat-ing organic matter into the soil and by grazing onthe bacterial community, earthworms have beenfound either to enhance or decrease bacterialbiomass (Ruz-Jerez et al., 1992; Scheu and Parkin-son, 1994; Bohlen and Edwards, 1995; Blair et al.,1997; Brown et al., 1998; Cortez et al., 2000), andto stimulate bacterial activity (Daniel and Ander-son, 1992; Ruz-Jerez et al., 1992; Wolters andJoergensen, 1992; Bohlen and Edwards, 1995). Thismay result in enhanced nitrogen immobilization ormineralization depending on species characteristicsand substrate quality.

We performed a mesocosm experiment withthree earthworm species from different ecologicalgroups (Bouche, 1977; Perel, 1977) (Lumbricusrubellus, Aporrectodea caliginosa tuberculata andLumbricus terrestris). We studied the bacterialcommunity and soil mineral N concentrations in thepresence and absence of these species and everyspecies combination. To distinguish between miner-al N derived from crop plant residues (residue-derived Nmin) and from soil organic matter (SOM-derived Nmin) as a result of litter feeding or soilfeeding activities by the earthworms, we used 15N-labelled plant materials (Brown et al., 1998; Bohlenet al., 1999). We hypothesized that earthwormspecies of different ecological groups would speci-fically affect nitrogen mineralization of the cropresidue and soil organic matter, and that interac-tions between the earthworm species would affectnitrogen mineralization, through competition andfacilitation effects. We hypothesized further thatthe effects on nitrogen mineralization could beexplained from effects on the bacterial biomassand growth rate.

ARTICLE IN PRESS

Earthworm species and interaction effects 245

Materials and methods

Experimental set-up

A mesocosm experiment was performed in PVCcolumns with sterilized soil amended with organic15N-labelled potato crop residue. We tested theeffect of earthworm species and species interac-tions on soil mineral N concentrations by manipula-tion of three earthworm species: L. rubellus(Hoffmeister, 1843), A. caliginosa tuberculata(Eisen, 1874) and Lumbricus terrestris L., indifferent combinations. Three single-species treat-ments were established, four multi-species combi-nations (L. rubellus+A. caliginosa, L. rubellus+L. terrestris, A. caliginosa+L. terrestris and L.rubellus+A. caliginosa+L. terrestris) and a controlfree of earthworms. We analysed earthwormspecies biomass and species interactions effects,using a regression analysis. The entire experimentincluded eight species combinations, with earth-worm biomass varying over four mesocosms withineach treatment, equalling 32 experimental units(Table 1).

MesocosmsMesocosms were constructed of PVC columns

(20 cm diameter, 45 cm height), which were split intwo longitudinal halves and fitted together withtape. By separating the halves, we could takesamples from specific depths. On 12 and 13 March2003, mesocosms were filled with the equivalent of

Table 1. Earthworm numbers and biomass (g) added per m

Treatment Mesocosm Numbers Biomass (g)

R C T R C T

Co 1 — — — — — —Co 2 — — — — — —Co 3 — — — — — —Co 4 — — — — — —R 5 2 — — 1.9 — —R 6 4 — — 3.7 — —R 7 6 — — 5.5 — —R 8 10 — — 9.8 — —C 9 — 2 — — 1.7 —C 10 — 6 — — 4.0 —C 11 — 8 ——— — 5.8 —C 12 — 12 — — 8.7 —T 13 — — 1 — — 5.5T 14 — — 1 — — 5.7T 15 — — 2 — — 16.0T 16 — — 2 — — 9.8

Co ¼ control, R ¼ L. rubellus, C ¼ A. caliginosa, T ¼ L. terrestris,CT ¼A. caliginosa+L. terrestris, RCT ¼ L. rubellus+A. caliginosa+L.

12 kg dry weight soil taken from a Fimic Anthrosol(FAO-UNESCO, 1988) situated in ‘De Bovenbuurt’(511590N, 51400E), The Netherlands. Soil materialwas collected (July 2002) from 0 to 30 cm depth inan arable field (sandy soil, under intensive arableculture for the last 20 years, organic mattercontent 3.28%). The soil was sieved over a 5-mmmesh to remove roots and coarse materials, andg-irradiated (25 kGy) to eliminate all living organ-isms. The soil was air-dried and remoistened to 60%WHC. During the experiment, soil moisture waskept constant gravimetrically with tap water, everyweek. After sterilization, drying and remoistening,pH-H2O equalled 6.0. Mesocosms were filled to36 cm to ensure a constant bulk density approx-imating 1.11 g cm�3.

Rhizon installationRhizons (miniature porous cups with a 2.5-mm

diameter and 10-cm long hydrophilic porouspolymer tube – Rhizon Soil Solution Samplers,art. number 19.21.25, Eijkelkamp AgrisearchEquipment, Giesbeek, The Netherlands) wereinstalled in the mesocosms in order to re-peatedly sample soil moisture. Two replicaterhizons were installed through holes in the sideof each mesocosm, at 12 and 15 cm soildepth. Rhizons were installed in horizontalposition in a crossed way above each otherwith a distance of 3 cm to minimize the influenceof the replicates on each other’s extractionefficiency.

esocosm

Treatment Mesocosm Numbers Biomass (g)

R C T R C T

RC 17 3 1 — 2.5 0.9 —RC 18 1 3 — 1.3 2.3 —RC 19 5 4 — 4.8 3.1 —RC 20 2 6 — 2.2 3.8 —RT 21 1 — 2 0.9 — 9.7RT 22 2 — 1 2.1 — 5.7RT 23 3 — 2 3.4 — 12.5RT 24 5 — 1 5.4 — 5.0CT 25 — 1 2 — 0.8 8.6CT 26 — 3 1 — 2.2 6.2CT 27 — 4 2 — 2.4 9.8CT 28 — 6 1 — 3.7 3.6RCT 29 3 1 2 3.3 0.7 8.3RCT 30 1 3 1 0.9 2.5 6.3RCT 31 5 4 2 5.4 2.9 12.5RCT 32 2 6 1 1.9 3.9 4.4

RC ¼ L. rubellus+A. caliginosa, RT ¼ L. rubellus+L. terrestris,terrestris.

ARTICLE IN PRESS

M.B. Postma-Blaauw et al.246

MicrobiotaThe sterilized soil was reinoculated with soil

microbiota. For this purpose we sieved unsterilizedair-dried soil (from ‘De Bovenbuurt’ fields) with a2mm mesh to remove earthworm cocoons. In eachmesocosm, 500 g of this inoculum soil was mixedthrough the sterilized soil before filling. Prior tothe experiment, the mesocosms were incubated for18 days in a climate chamber in the dark witha day/night temperature of 20/15 1C (‘day period’16 h) and air humidity of 80%. The mesocosmswere covered with a plastic sheet with some holes,to prevent excess evaporation. After incubation,total soil N was 1.01 g kg�1, total soil C was19.2 g kg�1, N–NO3 was 39.9mg kg�1 and N-NH4

was 1.2mg kg�1.

15N-labelled crop residueAfter preincubation, potato (Solanum tuberosum

L.) leaf and stem material (crop residue) was addedas food for the earthworms. The potato plants hadbeen labelled with 15N (Van Groenigen et al.,2005). The total N percentage of the crop residuewas 1.56%, and the atomic excess of the potatocrop residue for 15N was 1.236870.0430%. The cropresidue was dried, ground to approximately 5 cmand remoistened before addition to the mesocosmson top of the soil. We added 75 g dry weight of cropresidue to each mesocosm. This quantity wasadministered as ample food for the earthworms toremain vital during experimentation (Jos Bodt,personal communication, 2003).

EarthwormsAfter preincubation, the experiment was started

by inoculating the mesocosms with earthworms.Earthworm guts were not voided since this wouldhave reduced their vitality. The earthworm specieswere collected in the field prior to the experiment:L. rubellus and L. terrestris from a productiongrassland near Nijkerk, The Netherlands, A. caligi-nosa from a production grassland near Rheden, TheNetherlands. The earthworms were kept at 15 1Ctemperature on a turf medium with alder leaves(Alnus glutinosa) as a food source until experimen-tation. L. rubellus was introduced ranging from 1 to10 specimens per mesocosm, with a maximumtotal biomass of 9.8 g per mesocom. A. caliginosanumbers ranged from 1 to 12 specimens with amaximum total biomass of 8.7 g per mesocosm.L. terrestris numbers ranged from 1 to 2 specimenswith a maximum total biomass of 16.0 g permesocosm (Table 1). In the two- and three-speciescombinations, we added a maximum number of 11earthworms per mesocosm (equivalent to 350 in-d.m�2), and a maximum biomass of 20.8 g per

mesocosm (equivalent to 663 gm�2). Earthwormdensities ranged from densities comparable tothose in arable soil, in the mesocosms with 1, 2 or3 earthworms (Lee, 1985), to densities comparableto those in temperate grassland ecosystems, in themesocosms with the maximum number of earth-worms (Lee, 1985). This range was chosen withrespect to the possibility of increasing earthwormdensities in arable systems to densities that may befeasible under ideal circumstances.

After addition of earthworms, the mesocosmswere covered with a nylon mesh to prevent animalsfrom escaping. The mesocosms were kept in aclimate chamber with a temperature of 20 1C and adaylight period of 12 h. The experiment wasconducted from March 31 until June 2 in 2003.

Measurements

Mineral N sampling with rhizonsThe soil solution was sampled 14, 42 and 56 days

after introduction of earthworms. The concentra-tions of N-NO3 and N-NH4 at day 14 and 42 weremeasured using Skalar (Breda) continuous flowanalysis, those at day 56 were calculated on thebasis of the data obtained from the Stable IsotopeFacility (see below), due to a measuring artefactwith continuous flow analysis. To distinguish be-tween SOM-derived Nmin and residue-derived Nmin,the samples were analysed for 15N atomic excess ofmineral N. Mineral 15N was determined afterisolation of N-NO3 and N-NH4 using the micro-diffusion method of Brooks et al. (1989), modifiedby Sorensen and Jensen (1991). The method wasidentical to that described by Van Groenigen et al.(2005), except that 15N-NO3 and 15N-NH4 weredetermined simultaneously, rather than only15N-NO3, by adding 0.4 g Devarda’s alloy and 0.4 gMgO simultaneously. The acidified filters, packed inTeflon, were incubated for 7 or 8 days, dried andanalysed for 15N atomic excess on an isotope ratiomass spectrometer (ANCA-IRMS, Europa ScientificIntegra, UK) interfaced with a CN sample converterat the UC Davis Stable Isotope Facility, withatmospheric N2 as a standard (0.3663% atomicexcess). The 15N enrichments of the samples werecorrected for background abundances in the matrixsolution and filter paper.

Based on the data obtained from the StableIsotope Facility, we calculated the total mineral Namounts on day 56. Since the recovery of nitrogenafter preparation of the samples for 15N analysisaveraged 90%, we corrected the mineral N dataobtained from the Stable Isotope Facility with afactor 100/90.

ARTICLE IN PRESS

Earthworm species and interaction effects 247

Harvesting of the mesocosms, total soil N, totalsoil C

Mesocosms were harvested randomly 62 and 63days after the start of the experiment. Due to afailure of the climate chamber on day 61, theambient temperature had increased to 40 1C for8 h. The mesocosms were cooled at 4 1C immedi-ately afterwards. Earthworms had died in alltreatments, and final biomass could therefore notbe determined. However, in all treatments, re-mains of earthworms of the different species werefound, suggesting that earthworms had been alivein the treatments until day 61 of the experiment.We expect that a short period with high tempera-ture did not cause major changes in bacterialbiomass and diversity since 40 1C is not high enoughto kill bacteria, and it usually takes more than 8 hto induce significant changes in bacterial biomass.Bacterial growth rate (thymidine and leucineincorporation) may be more sensitive to an in-creased temperature. Nevertheless, we assumethat the differences between treatments werenot affected by the temperature incident. MineralN data were not affected, since all mineral Nmeasurements were taken before the accident.Total soil N and C amounts are relatively stableparameters, that are probably not affected by ashort temperature increase.

The remainders of the applied crop residue ontop of the mesocosms were removed by hand. Sincedead earthworm tissue was found mainly in thelower part of the soil mesocosm, soil from theupper 13 cm of the mesocosm was separated fromthe lower part and analyses were restricted to theupper parts of the mesocosms. Sub-samples weredried at 40 1C, ball-milled and approximately 40mgwas weighed out exactly in tin cups. The sampleswere analysed for total soil N. Total soil C wasanalysed by wet oxidation with K2Cr2O7 (Walingaet al., 1992).

Bacterial biomass, growth rate and DNA analysisBacterial numbers and cell volumes at day 62

were measured by confocal laser scanning micro-scopy and automatic image analysis, and biomasswas calculated from biovolume (Bloem et al.,1995a, 1995b). The bacterial growth rate wasdetermined by thymidine incorporation, a measurefor DNA synthesis (growth rate) and leucineincorporation, a measure for protein synthesis(growth rate and biomass turnover) (Michel andBloem, 1993). These analyses were carried out foreach earthworm species combination, for threeout of four randomly chosen mesocosms. Froma selection of the treatments viz. the control,L. rubellus in single-species culture, L. rubellus and

A. caliginosa in mixed-species culture andL. rubellus, A. caliginosa and L. terrestris inmixed-species culture, DNA was extracted (VanElsas and Smalla, 1995) and analysed by denaturat-ing gradient gel electrophoresis (DGGE) (Muyzer etal., 1993). This technique yields a banding patternwhere the number of DNA bands reflects thenumber of ‘species’ (genotypes) of abundantbacteria, and the band intensity reflects therelative abundance of the species (Bloem andBreure, 2003; Dilly et al., 2004).

Calculations and statistical analysesTotal mineral N was calculated as the sum of N-

NO3 and N-NH4. Residue-derived Nmin was calcu-lated as follows:

(1)

total excess 15N (mg l�1) ¼ total mineralN (mg l�1)� at% 15N excess/100,

(2)

residue-derived Nmin (mg l�1) ¼ total excess 15N(mg l�1)� 100 /at% 15N excess of the cropresidue.

The SOM-derived Nmin was calculated as the totalmineral N minus the residue-derived Nmin. Calcula-tions of at%15N excess in mineral N were based on anatural background abundance value of 0.3663 at%excess.

Statistical analyses were carried out using thestatistical package SPSS version 11.0. Datawere analysed using backward regression withthe three single species and all two-way andthree-way interactions as input variables. Inthis procedure, only significant parameters wereretained in the model. Since the relationshipbetween earthworm species biomass and severaloutput parameters is regarded to be a square-rootrather than a linear function, square-roottransformations were applied in several cases, toincrease model fit and to meet assumptions ofhomogeneity of variances and normality of data.For five variables, one or two outliers wereremoved based on the Cook’s Distance. The effectof the selected earthworm treatments on thenumber of bacterial DNA bands was analysed witha one-factor ANOVA, followed by a Tukey post-hocanalysis.

Results

Mineral N derived from crop residue

The concentrations of residue-derived Nmin in thesoil solution generally increased during experimen-

ARTICLE IN PRESS

Table

2.

Statistica

lresultsof

thebac

kwardregression

analysis

ofea

rthw

orm

spec

iesbiomass(g)an

dspec

iesinteractioneffectson

system

variab

les:

mod

elR2

and

sign

ifica

nce,

andtheco

nstant

andregression

coefficien

tsof

theparam

etersinclud

edin

themod

el

Mod

elR2

Mod

elP-va

lue

Con

stan

tMod

elco

efficien

tsof

param

eters

includ

edin

themod

el

Mineral

Nde

rive

dfrom

thecrop

residu

e(m

gl�

1)

Day

140.46

5o0.00

12.41

+0.50*R+0

.20*C

Day

420.52

9o0.00

17.74

+0.94*R+0

.44*T

Day

560.68

6o0.00

18.78

+1.23*R+0

.63*T

Mineral

Nde

rive

dfrom

soilorga

nicmatter(m

gl�

1)

Day

14n.s.

328

Day

42n.s.

332

Day

560.74

o0.00

133

5+3

4.19

*OR+1

5.70

*OC�23

.66*ORC

�10

.44*ORT

-11.21

*OCT+9

.30*ORCT

TotalC(g

kg�1)

0.25

80.00

420

.5�0.12

*CT

TotalN(g

kg�1)

n.s.

1.08

Bac

terial

biomass(mgCg�

1)

0.54

6o0.00

174

.8+7

.73*ORC+1

.02*RT

-1.26*RCT

Thy

midineinco

rporation(pmolg�

1h�

1)

n.s.

31.6

Leuc

ineinco

rporation(pmolg�

1h�

1)

0.27

70.00

919

2+2

.53*CT

Abbreviations:R¼

L.rube

llus,C¼

A.caligino

sa,T¼

L.terrestris,RC¼

L.rube

llus

*A.caligino

sa,RT¼

L.rube

llus

*L.

terrestris,CT¼A.caligino

sa*L.

terrestris,RCT¼

L.rube

llus

*A.caligino

sa*L.

terrestris.

M.B. Postma-Blaauw et al.248

ARTICLE IN PRESS

Earthworm species and interaction effects 249

tation, with final concentrations ranging between 8and 20mg l�1. L. rubellus induced an increase inresidue-derived Nmin from day 14 of the experimentonwards, with a 140% increase per 10 g earthwormbiomass on day 56 (Po0.001 on day 14, on day 42and on day 56, Table 2, Fig. 1). L. terrestrisenhanced residue-derived Nmin from day 42 of theexperiment onwards, with a 72% increase per 10 gearthworm biomass on day 56 (Po0.001 on day 42and on day 56, Table 2, Fig. 1). A. caliginosa did notaffect residue-derived Nmin, except for a smallpositive effect on day 14 (Table 2). No interactioneffects of the earthworm species on residue-derived Nmin were found.

Figure. 1. Partial residual plots of the effect of (a) L. rubemineral N derived from the crop residue (mg l�1) on day 56 oanalysis. For the model R2 and significance, and the coefficie

Mineral N derived from soil organic matter

The concentrations of SOM-derived Nmin in thesoil solution were approximately 300mg l�1, withlittle variation over time. No effects of the earth-worms on SOM-derived Nmin were observed up today 42. On day 56, L. rubellus and A. caliginosaenhanced SOM-derived Nmin. L. rubellus enhancedSOM-derived Nmin by 32% per 10 g earthwormbiomass (Po0.001), A. caliginosa enhanced SOM-derived Nmin by 15% per 10 g earthworm biomass(P ¼ 0:005) (Table 2, Fig. 2a,b).

All two-species interactions negatively affectedSOM-derived Nmin on day 56: SOM-derived Nmin in

llus (g) and (b) L. terrestris (g) on the concentrations off the experiment, as determined by backward regressionnts of the parameters included in the model, see Table 2.

ARTICLE IN PRESS

Figure 2. Partial residual plots of the effect of (a) L. rubellus (g), (b) A. caliginosa (g), and the interaction effects of (c)L. rubellus and A. caliginosa (g), (d) L. rubellus and L. terrestris (g), (e) A. caliginosa and L. terrestris (g), and (f) L.rubellus, A. caliginosa and L. terrestris (g) on the concentrations of mineral N derived from soil organic matter (mg l�1)on day 56 of the experiment, as determined by backward regression analysis. For the model R2 and significance, and thecoefficients of the parameters included in the model, see Table 2.

M.B. Postma-Blaauw et al.250

these mixed-species treatments was lower thancould have been expected based on the single-species treatments (Po0.001 for the interactionbetween L. rubellus and A. caliginosa, Po0.001 forthe interaction between L. rubellus and L. terres-tris and P ¼ 0:001 for the interaction betweenA. caliginosa and L. terrestris, Table 2, Fig. 2c–e).When all three species were present together,however, we found a positive interaction effect onSOM-derived Nmin (P ¼ 0:002, Table 2, Fig. 2f).

Total N and total C in the soil

At the end of the experiment, total soil Naveraged 1.08 g kg�1 and the amounts of total soilC ranged between 17 and 23 g kg�1. None of theearthworm species had a main effect on totalsoil N or total soil C. The interaction betweenA. caliginosa and L. terrestris, however, negativelyaffected total soil C (P ¼ 0:004, Table 2, Fig. 4b).

Bacteria

At the end of the experiment, bacterial biomassranged between 58 and 121 mg C g�1 soil. None ofthe three earthworm species had a main effect onbacterial biomass. The two-species interactionsbetween L. rubellus and L. terrestris (Po0.001)and between L. rubellus and A. caliginosa(P ¼ 0:002), however, positively affected bacterialbiomass (Table 2, Fig. 3a,b). In contrast, when allthree species were present together, we found anegative interaction effect on the bacterial bio-mass (Table 2, Po0.001, Fig 3c): the bacterialbiomass in this three-species treatment was lowerthan expected based on the single-species treat-ments. The interaction between A. caliginosa andL. terrestris did not affect bacterial biomass.

Bacterial growth rate measured as thymidineincorporation was on average 31.6 pmol g�1 h�1,and measured as leucine incorporation it rangedbetween 146 and 261 pmol g�1 h�1. None of thethree earthworm species had a main effect on

ARTICLE IN PRESS

Figure 3. Partial residual plots of the interaction effects of (a) L. rubellus and A. caliginosa (g), (b) L. rubellus andL. terrestris (g) and (c) L. rubellus, A. caliginosa and L. terrestris (g) on the bacterial biomass (mg C g�1), as determinedby backward regression analysis. For the model R2 and significance, and the coefficients of the parameters included inthe model, see Table 2.

Earthworm species and interaction effects 251

bacterial growth rate. The two-species interactionbetween A. caliginosa and L. terrestris, however,positively affected leucine incorporation (Table 2,Po0.009, Fig. 4a).

The number of bacterial DNA bands in theselected treatments ranged between 58 and 70bands. Earthworms had no effect on, or decreasedthe number of DNA bands. The number of DNAbands in the single-species treatments withL. rubellus (70 bands) was similar to the controltreatments (70 bands). The number of DNA bands inthe mixed-species treatments with L. rubellus andA. caliginosa (58.7 bands), and with all threespecies (60.5 bands), was lower than in the control(one-factor ANOVA, Po0.001).

Discussion

We hypothesized that earthworm species ofdifferent ecological groups would affect nitrogenmineralization differently, and that earthwormspecies interactions would affect nitrogen miner-alization through competition and facilitationeffects. Our hypothesis was confirmed by theresults showing that earthworm species had differ-ent effects on mineral N concentrations, agreeingwith the characteristics of the ecological groupsthey represented. Earthworm species interactionsalso affected mineral N concentrations and thebacterial community, the effects being specificallydependent on the combination of species present.We will henceforth consider the effects of thedifferent earthworm species and species combina-tions on the soil and bacterial parameters sepa-rately.

Earthworm species effects

L. rubellus and L. terrestris enhanced residue-derived Nmin indicating that mineralization of theresidue was enhanced. These results can beexplained from the incorporation of plant residuesinto the soil by L. rubellus and L. terrestris(Bouche, 1977). A. caliginosa forages mainly onsoil organic matter (Bouche, 1977), which explainsthe lack of an effect of A. caliginosa on residue-derived Nmin. L. rubellus and A. caliginosa also hada positive effect on SOM-derived Nmin, indicatingthat the mineralization of soil organic matter wasenhanced by the burrowing and feeding of thesespecies. Although statistically significant, the ef-fect of A. caliginosa on SOM-derived Nmin was smallcompared to the effect of L. rubellus (Fig. 2a,b),and the ecological relevance seems limited underour experimental conditions. We found no signifi-cant effect of L. terrestris on SOM-derived Nmin.Possibly, the lower burrowing activity in thepermanent burrow system of L. terrestris may nothave resulted in increased SOM-derived Nmin, incontrast to L. rubellus and A. caliginosa. Positiveeffects of earthworms on nitrogen mineralizationare frequently found for earthworm species ingeneral (Blair et al., 1997; Cortez et al., 2000)and were specifically described for L. rubellus(Brown et al., 1998; Whalen et al., 2000),and endogeic species (Villenave et al., 1999;Whalen et al., 2000). Both positive and negativeeffects of L. terrestris on nitrogen mineralizationhave been found (Devliegher and Verstraete,1997; Bohlen et al., 1999; Whalen et al., 2000;Tiunov and Dobrovolskaya, 2002; Wilcox et al.,2002).

ARTICLE IN PRESS

Figure 4. Partial residual plots of the interaction effectsof A. caliginosa and L. terrestris (g) on (a) the bacterialgrowth rate measured as leucine incorporation(pmol g�1 h�1) and (b) total soil C (g kg�1), as determinedby backward regression analysis. For the R2 and sig-nificance of the models, and the coefficients of theparameters included in the models, see Table 2.

M.B. Postma-Blaauw et al.252

We expected that the incorporation of organicmatter by L. rubellus and L. terrestris would resultin increased total soil N and C amounts, and thatthe increased nitrogen mineralization would concurwith effects on the bacterial community. We foundno effects of the three earthworms species on theamounts of total soil N and C or on the bacterialcommunity, however. In other studies a positiveeffect of earthworms on total soil N and C wasusually observed on a longer term than ourexperiment lasted (Gilot, 1997; Villenave et al.,1999). Earthworms have been found to stimulatebacterial activity (Daniel and Anderson, 1992; Ruz-

Jerez et al., 1992; Li et al., 2002) and to eitherreduced or enhance bacterial biomass (Ruz-Jerezet al., 1992; Scheu and Parkinson, 1994; Bohlen andEdwards, 1995; Blair et al., 1997; Devliegher andVerstraete, 1997; Brown et al., 1998; Bohlen et al.,1999; Cortez et al., 2000). We agree with Bohlenet al. (1999) that no simple relationship existsbetween bacterial biomass and earthworms.

Species interactions

The combination of L. rubellus and A. caliginosaresulted in a reduced SOM-derived Nmin and anenhanced bacterial biomass, indicating that im-mobilization of nitrogen into bacterial biomassoccurred. Possibly, burrowing activity by concurringspecies had increased, leading to an increasedavailability of mineralized nitrogen for incorpora-tion into the bacterial biomass. The combination ofL. rubellus and A. caliginosa also led to a reductionin bacterial genotypes compared to L. rubellus insingle-species culture, suggesting that the induc-tion of bacterial biomass by this treatment wasaccomplished by an increase in a limited number ofdominant species. A similar result was found for theinteraction between two nematode species (Post-ma-Blaauw et al., 2005).

Since crop residue incorporated into the soil byL. rubellus may serve as an additional food sourcefor A. caliginosa (Hughes et al., 1994; Villenaveet al., 1999), we had expected an increasedmineralization of the crop residue in the presenceof the two species. The absence of an interactioneffect on residue-derived Nmin despite incorpora-tion of crop residue into the soil by L. rubellus maybe explained by the assumption that A. caliginosadid not feed on the incorporated residue after all.Alternatively, A. caliginosa may have affected grossmineralization of the crop residue, but net miner-alization may have been limited due to immobiliza-tion of mineralized nitrogen into the bacterialbiomass.

Similar to the interaction effect of L. rubellusand A. caliginosa, the combination of L. rubellusand L. terrestris resulted in a reduced SOM-derivedNmin and enhanced bacterial biomass, indicatingthat immobilization of nitrogen into the bacterialbiomass occurred. This effect may have beencaused by an increased incorporation of the cropresidue, and/or an increased burrowing activity bythe concurring species. The absence of an interac-tion effect on residue-derived Nmin may indicatethe absence of a direct competition effect betweenthe species. Alternatively, mineralized nitrogenfrom the crop residue may have been incorporated

ARTICLE IN PRESS

Earthworm species and interaction effects 253

into the bacterial biomass. Evidence for interactionbetween L. rubellus and L. terrestris was given byLowe and Butt (1999), who observed competitionfor food between these species, resulting inreduced abundances of the species. A competitioneffect was also observed for other earthwormspecies (Abbott, 1980). Our results indicate thatbesides affecting population densities, the interac-tion between L. rubellus and L. terrestris canaffect the bacterial community and nutrientcycling, leading to immobilization of nitrogen.

Contrary to the previous interactions, the interac-tion between A. caliginosa and L. terrestris led toincreased bacterial growth rate (leucine incorpora-tion), and reduced total soil C, indicating thatbacterial mineralization increased in the presenceof L. terrestris and A. caliginosa. These results,however, contradict the finding that SOM-derivedNmin decreased in the presence of this speciescombination. This decrease in SOM-derived Nmin wassmall compared to the effect of the other speciescombinations (Fig. 2c, d), however, and may haveresulted from increased leaching of nitrogen alongthe permanent vertical burrows of L. terrestris,or from increased denitrification in the presence ofL. terrestris (Parkin and Berry, 1999).

Although the effects of the combination ofL. terrestris and A. caliginosa on bacterial growthrate and total C were small, these effects can beexplained by the ecological characteristics of thespecies. Jegou et al. (2001) indicated thatA. caliginosa may preferably use the burrows ofL. terrestris and Aporrectodea giardi, and may usecastings and burrow linings of L. terrestris as foodsources. In our case, feeding of A. caliginosa on theburrow linings of L. terrestris may have resulted in

Figure 5. Earthworm species effects (solid lines) and twconcentrations of mineral N derived from the crop residue, cmatter, total soil N and C, and the bacterial community. ThNegative, positive and strongly positive effects are indicated

increased bacterial growth rate. Alternatively,A. caliginosa may have affected the burrowingactivity of L. terrestris. Jegou et al. (2001) foundthat L. terrestris increased its burrow length in thepresence of A. caliginosa, and burrowing ofL. terrestris was more superficial than whenincubated alone. Possibly, an increased burrowingof L. terrestris may have led to increased bacterialgrowth rate and mineralization of total soil C.

The three-species interaction enhanced SOM-derived Nmin and reduced bacterial biomass, sug-gesting that bacterial activity increased under thethree-species combination. Indeed bacterial growthrate was highest in the three-species treatment, andtotal soil C second lowest. The combination ofA. caliginosa and L. terrestris showed a similar highbacterial growth rate and low total C, while thetwo-species combinations with L. rubellus (withA. caliginosa, and with L. terrestris) showedopposite effects. We therefore suggest that theeffect of the three-species combination was mainlydetermined by the simultaneous presence of A.caliginosa and L. terrestris, while the contributionof L. rubellus was minimized.

Conclusions

A synthesis of the main effects of the threeinvestigated earthworm species and the effectsof species is presented in Fig. 5. We found thatL. rubellus and L. terrestris enhanced mineraliza-tion of the crop residue, whereas no indication of aneffect was found for A. caliginosa. On the otherhand, L. rubellus and A. caliginosa enhancedmineralization of the soil organic matter, whereas

o-way species interactions effects (dotted lines) ononcentrations of mineral N derived from the soil organicree-way interaction effects were described in the text.by -, + and ++, respectively.

ARTICLE IN PRESS

M.B. Postma-Blaauw et al.254

we found no indication of an effect of L. terrestris.The interactions between different earthwormspecies affected the bacterial community and soilmineral N concentrations. Our results led to thesuggestion that the interaction between L. rubellusand L. terrestris, and L. rubellus and A. caliginosaled to increased immobilization of mineral Ninto the bacterial biomass. On the other hand, theinteraction between A. caliginosa and L. terrestrisled to increased bacterial growth rate and increasedmineralization of total soil C. When all threespecies were combined, the interaction effect ofA. caliginosa and L. terrestris dominated.

Our results indicate that the effects of earth-worms on nitrogen mineralization depend on theecological traits of the earthworm species and aredependent on earthworm community composition.Similar conclusions were reached for fungal grazingcollembolans (Faber and Verhoef, 1991; Cragg andBardgett, 2001) and bacterivorous nematodes(Mikola and Setala, 1998; Wardle, 2002; Postma-Blaauw et al., 2005). Together with these findings,our results support the hypothesis that there is nogeneral rule on the effect of diversity within soilbiota groups on ecosystem functioning, but that theoutcome depends on the ecological traits of thespecies present (Heemsbergen et al., 2004; Swiftet al., 2004).

The comparison of earthworm species diversityranging from one to three species is relevant inagricultural systems where earthworm speciesdiversity is usually below five species (Lee, 1985),and can be restricted to one species (Curry et al.,2002). The notion that earthworm communities canspecifically affect soil fertility may be of greatimportance to increase sustainable land use inagroecosystems, as proper earthworm managementmay sustain crop yields whilst fertilizer inputs canbe reduced (Brown et al., 2004).

Acknowledgements

We are very grateful to Guido Heijmans and JosBodt for collecting the earthworms and assistingwith the set-up and harvesting of the experiment,and to An Vos for analysing the bacterial commu-nity. Furthermore, we are grateful to EduardHummelink for producing and harvesting thelabelled potato plants. We thank Jaap Nelemansand Willeke van Tintelen for assistance in thelab, Wim Didden for valuable comments on theexperimental set-up and results, and KlemensEkschmitt (Justus Liebig University, Giessen, Ger-many) and Jouke Postma for help with the

statistical set-up and analyses. This study wasfunded by the Stimulation Programme Biodiversityof NWO, the Netherlands Organisation for ScientificResearch. J. Bloem and J.H. Faber were supportedby the Dutch Ministry of Agriculture, Nature andFood Quality through DWK programme 352 onAgrobiodiversity. We thank Stefan Scheu and ananonymous reviewer for their valuable commentsto an earlier version of the manuscript.

References

Abbott, I., 1980. Do earthworms compete for food? SoilBiol. Biochem 12, 523–530.

Baker, G., Carter, P., Barrett, V., Hirth, J., Mele, P.,Gourley, C., 2002. Does the deep-burrowing earth-worm, Aporrectodea longa, compete with residentearthworm communities when introduced to pasturesin south-eastern Australia? Eur. J. Soil Biol. 38, 39–42.

Bardgett, R.D., Cook, R., 1998. Functional aspects of soilanimal diversity in agricultural grasslands. Appl. SoilEcol. 10, 263–276.

Begon, M., Harper, J.L., Townsend, C.R., 1990. Ecology:Individuals, Populations and Communities, secondedn. Blackwell scientific publications, Boston.

Blair, J.M., Parmelee, R.W., Allen, M.F., Mccartney, D.A.,Stinner, B.R., 1997. Changes in soil N pools in responseto earthworm population manipulations in agrocosys-tems with different N sources. Soil Biol. Biochem. 29,361–367.

Bloem, J., Breure, A.M., 2003. Microbial indicators. In:Markert, B.A., Breure, A.M., Zechmeister, H.G. (Eds.),Bioindicators/Biomonitors – Principles, Assessment,Concepts. Elsevier, Amsterdam, pp. 259–282.

Bloem, J., Bolhuis, P.R., Veninga, M., Wieringa, J.,1995a. Microscopic methods for counting bacteriaand fungi in soil. In: Alef, K.P.N. (Ed.), Methods inApplied Soil Microbiology and Biochemistry. AcademicPress, London, pp. 162–173.

Bloem, J., Veninga, M., Shepherd, J., 1995b. Fullyautomatic determination of soil bacterium numbers,cell volumes and frequencies of dividing cells byconfocal laser scanning microscopy and image analy-sis. Appl. Environ. Microbiol. 61, 926–936.

Bohlen, P.J., Edwards, C.A., 1995. Earthworm effects onN dynamics and soil respiration in microcosms receiv-ing organic and inorganic nutrients. Soil Biol. Bio-chem. 27, 341–348.

Bohlen, P.J., Parmelee, R.W., Allen, M.F., Ketterings,Q.M., 1999. Differential effects of earthworms onnitrogen cycling from various nitrogen-15-labeledsubstrates. Soil Sci. Soc. Am. J. 63, 882–890.

Bouche, M.B., 1977. Strategies lombriciennes. In: Lohm,U., Persson, T. (Eds.), Soil Organisms as Components ofEcosystems. Ecol. Bull., Stockholm, pp. 122–132.

Brooks, P.D., Stark, J.M., McInteer, B.B., Preston, T., 1989.Diffusion method to prepare soil extracts for automatedN-15 analysis. Soil Sci. Soc. Am. J. 53, 1707–1711.

ARTICLE IN PRESS

Earthworm species and interaction effects 255

Brown, G.G., Hendrix, P.F., Beare, M.H., 1998. Earth-worms (Lumbricus rubellus) and the fate of 15N insurface-applied sorghum residues. Soil Biol. Biochem.30, 1701–1705.

Brown, G.G., Pashanasi, B., Villenave, C., Patron, J.C.,Senapati, B.K., Giri, S., Barois, I., Lavelle, P.,Blanchart, E., Blakemore, R.J., Spain, A.V., Boyer,J., 1999. Effects of earthworms on plant production inthe tropics. In: Lavelle, P., Brussaard, L., Hendrix, P.(Eds.), Earthworm Management in Tropical Agroeco-systems. CABI Publishing, Oxon, UK, pp. 87–147.

Brown, G.G., Edwards, C.A., Brussaard, L., 2004. Howearthworms affect plant growth: burrowing into themechanisms. In: Edwards, C.A. (Ed.), EarthwormEcology, second ed. CRC Press, Boca Raton, pp. 13–49.

Brussaard, L., Behan-Pelletier, V.M., Bignell, D.E., Brown,V.K., Didden, W., Folgarait, P., Fragoso, C., Freckman,D.W., Gupta, V.V.S.R., Hattori, T., Hawksworth, D.L.,Klopatek, C., Lavelle, P., Malloch, D.W., Rusek, J.,Soderstrom, B., Tiedje, J.M., Virginia, R.A., 1997.Biodiversity and ecosystem functioning in soil. Ambio26, 563–570.

Capowiez, Y., Monestiez, P., Belzunces, L., 2001. Burrowsystems made by Aporrectodea nocturna and Allolo-bophora chlorotica in artificial cores: morphologicaldifferences and effects of interspecific interactions.Appl. Soil Ecol. 16, 109–120.

Cortez, J., Hameed, R.H., 2001. Simultaneous effects ofplants and earthworms on mineralization of 15N-labelled organic compounds adsorbed onto soil sizefractions. Biol. Fert. Soils 33, 218–225.

Cortez, J., Billes, G., Bouche, M.B., 2000. Effect of climate,soil type and earthworm activity on nitrogen transferfrom a nitrogen-15-labelled decomposing material underfield conditions. Biol. Fert. Soils 30, 318–327.

Cragg, R.G., Bardgett, R.D., 2001. How changes in soilfaunal diversity and composition within a trophicgroup influence decomposition processes. Soil Biol.Biochem. 33, 2073–2081.

Curry, J.P., Byrne, D., Schmidt, O., 2002. Intensivecultivation can drastically reduce earthworm popula-tions in arable land. Eur. J. Soil Biol. 38, 127–130.

Dalby, P.R., Baker, G.H., Smith, S.E., 1998. Competitionand cocoon consumption by the earthworm Aporrec-todea longa. Appl. Soil Ecol. 10, 127–136.

Daniel, O., Anderson, J.M., 1992. Microbial biomass andactivity in contrasting soil materials after passagethrough the gut of the earthworm Lumbricus rubellusHoffmeister. Soil Biol. Biochem. 24, 465–470.

Devliegher, W., Verstraete, W., 1997. The effect ofLumbricus terrestris on soil in relation to plantgrowth: effects of nutrient-enrichment processes(NEP) and gut-associated processes(GAP). Soil Biol.Biochem. 29, 341–346.

Dilly, O., Bloem, J., Vos, A., Munch, J.C., 2004. Bacterialdiversity during litter decomposition in agriculturalsoils. Appl. Environ. Microbiol. 70, 468–474.

Doube, B.M., Williams, P.M.L., Willmott, P.J., 1997. Theinfluence of two species of earthworms (Aporrectodeatrapezoides and Aporrectodea rosea) on the growth of

wheat, barley and faba beans in three soil types in thegreenhouse. Soil Biol. Biochem. 29, 503–509.

Faber, J.H., Verhoef, H.A., 1991. Functional differencesbetween closely related soil arthropods with respectto decomposition and nitrogen mobilization in a pineforest. Soil Biol. Biochem. 23, 15–23.

FAO-UNESC, 1988. Soil map of the world: revised legend.FAO, Rome, Italy.

Fragoso, C., Brown, G.G., Patron, J.C., Blanchart, E.,Lavelle, P., Pashanasi, B., Senapati, B., Kumar, T.,1997. Agricultural intensification, soil biodiversity andagroecosystem function in the tropics: the role ofearthworms. Appl. Soil Ecol. 6, 17–35.

Gilot, C., 1997. Effects of a tropical geophageousearthworm, M. anomala (Megascolecidae), on soilcharacteristics and production of a yam crop in IvoryCoast. Soil Biol. Biochem. 29, 353–359.

Heemsbergen, D.A., Berg, M.P., Loreau, M., van Hal,J.R., Faber, J.H., Verhoef, H.A., 2004. Biodiversityeffects on soil processes explained by interspecificfunctional dissimilarity. Science 306, 1019–1020.

Hughes, M.S., Bull, C.M., Doube, B.M., 1994. The use ofresource patches by earthworms. Biol. Fert. Soils 18,241–244.

Jegou, D., Capowiez, Y., Cluzeau, D., 2001. Interactionsbetween earthworm species in artificial soil coresassessed through the 3D reconstruction of the burrowsystems. Geoderma 102, 123–137.

Lee, K.E., 1985. Earthworms: Their Ecology and Relation-ships with Soils and Land Use. Academic Press, Sydney.

Li, X., Fisk, M.C., Fahey, T.J., Bohlen, P.J., 2002.Influence of earthworm invasion on soil microbialbiomass activity in a northern hardwood forest. SoilBiol. Biochem. 34, 1929–1937.

Lowe, C.N., Butt, K.R., 1999. Interspecific interactionsbetween earthworms: a laboratory-based investiga-tion. Pedobiologia 43, 808–817.

Lowe, C.N., Butt, K.R., 2002. Growth of hatchlingearthworms in the presence of adults: interactions inlaboratory culture. Biol. Fert. Soils 35, 204–209.

Michel, P.H., Bloem, J., 1993. Conversion factors forestimation of cell production rates of soil bacteriafrom thymidine and leucine incorporation. Soil Biol.Biochem. 25, 943–950.

Mikola, J., Setala, H., 1998. Relating species diversity toecosystem functioning: mechanistic backgrounds andexperimental approach with a decomposer food web.Oikos 83, 180–194.

Muyzer, G., de Waal, E.C., Uitterlinden, A.G., 1993.Profiling of complex microbial populations by denatur-ing gradient gel electrophoresis analysis of polymerasechain reaction-amplified genes coding for 16S rRNA.Appl. Environ. Microbiol. 59, 695–700.

Parkin, T.B., Berry, E.C., 1999. Microbial nitrogentransformations in earthworm burrows. Soil Biol.Biochem. 31, 1765–1771.

Perel, T.S., 1977. Differences in lumbricid organizationconnected with ecological properties. In: Lohm, U.,Persson, T. (Eds.), Soil Organisms as Components ofEcosystems. Ecol. Bull., Stockholm, pp. 56-63.

ARTICLE IN PRESS

M.B. Postma-Blaauw et al.256

Postma-Blaauw, M.B., de Vries, F.T., de Goede, R.G.M.,Bloem, J., Faber, J.H., Brussaard, L., 2005. Within-trophic group interactions of bacterivorous nematodespecies and their effects on the bacterial communityand nitrogen mineralization. Oecologia 142, 428–439.

Ruz-Jerez, B.E., Ball, P.R., Tillman, R.W., 1992. Labora-tory assessment of nutrient release from a pasture soilreceiving grass or clover residues, in the presence orabsence of Lumbricus rubellus or Eisenia foetida. SoilBiol. Biochem. 24, 1529–1534.

Scheu, S., Parkinson, D., 1994. Effects of earthworms onnutrient dynamics, carbon turnover and microorgan-isms in soils from cool temperate forests of theCanadian Rocky Mountains – laboratory studies. Appl.Soil Ecol. 1, 113–125.

Sorensen, P., Jensen, E.S., 1991. Sequential diffusion ofammonium and nitrate from soil extracts to apolytetrafluoroethylene trap for N-15 determination.Anal. Chim. Acta 252, 201–203.

Swift, M.J., Izac, A.M.N., van Noordwijk, M., 2004.Biodiversity and ecosystem services in agriculturallandscapes – are we asking the right questions? AgricEcosyst. Environ. 104, 113–134.

Tiunov, A.V., Dobrovolskaya, T.G., 2002. Fungal andbacterial communities in Lumbricus terrestris burrowwalls: a laboratory experiment. Pedobiologia 46,595–605.

Van Elsas, J.D., Smalla, K., 1995. Extraction of microbialcommunity DNA from soils. In: Akkermans, A.D.L., vanElsas, J.D., de Bruijn, F.J. (Eds.), Molecular MicrobialEcology Manual. Kluwer Academic Publishers, Dor-drecht* The Netherlands, pp. 1.3.3.1-1.3.3.11.

Van Groenigen, J.W., Georgius, P.J., van Kessel, C.,Hummelink, E.W.J., Velthof, G.L., Zwart, K.B., 2005.Subsoil 15N–N2O concentrations in a sandy soil profileafter application of 15N-fertilizer. Nutr. Cycling Agroe-cosystems 72, 13–25.

Villenave, C., Charpentier, F., Lavelle, P., Feller, C.,Brussaard, L., Pashanasi, B., Barois, I., Albrecht, A.,Patron, J.C., 1999. Effects of earthworms on soilorganic matter and nutrient dynamics following earth-worm inoculation in field experimental situations. In:Lavelle, P., Brussaard, L., Hendrix, P. (Eds.), Earth-worm Management in Tropical Agroecosystems. CABIPublishing, Oxon* UK, pp. 173–198.

Walinga, I., Kithome, M., Novozamsky, I., Houba, V.J.G.,Vanderlee, J.J., 1992. Spectrophotometric determi-nation of organic-carbon in soil. Commun. Soil Sci.Plant Anal. 23, 1935–1944.

Wardle, D.A., 2002. Communities and Ecosystems: Link-ing the Aboveground and Belowground Components.Princeton University Press, Oxford.

Whalen, J.K., Parmelee, R.W., Subler, S., 2000. Quanti-fication of nitrogen excretion rates for three lumbricidearthworms using 15N. Biol. Fert. Soils 32, 347–352.

Wilcox, C.S., Dominguez, J., Parmelee, R.W., McCartney,D.A., 2002. Soil carbon and nitrogen dynamics inLumbricus terrestris. L. middens in four arable, apasture, and a forest ecosystems. Biol. Fert. Soils 36,26–34.

Wolters, V., Joergensen, R.G., 1992. Microbial carbonturnover in beech forest soils worked by Aporrectodeacaliginosa (Savigny) (Oligochaeta, Lumbricidae). SoilBiol. Biochem. 24, 171–177.