Content uploaded by Alexandru Milcu
Author content
All content in this area was uploaded by Alexandru Milcu on Nov 15, 2017
Content may be subject to copyright.
Ecology, 87(10), 2006, pp. 2548–2558
Ó2006 by the Ecological Society of America
DECOMPOSERS (LUMBRICIDAE, COLLEMBOLA) AFFECT PLANT
PERFORMANCE IN MODEL GRASSLANDS OF DIFFERENT DIVERSITY
STEPHAN PARTSCH,
1,2
ALEXANDRU MILCU,
1,3
AND STEFAN SCHEU
1
1
Institut fu
¨r Zoologie, Technische Universita
¨t Darmstadt, Schnittspahnstr. 3, D-64287, Darmstadt, Germany
Abstract. Decomposer invertebrates influence soil structure and nutrient mineralization
as well as the activity and composition of the microbial community in soil and therefore likely
affect plant performance and plant competition. We established model grassland communities
in a greenhouse to study the interrelationship between two different functional groups of
decomposer invertebrates, Lumbricidae and Collembola, and their effect on plant perfor-
mance and plant nitrogen uptake in a plant diversity gradient. Common plant species of
Central European Arrhenatherion grasslands were transplanted into microcosms with
numbers of plant species varying from one to eight and plant functional groups varying
from one to four. Separate and combined treatments with earthworms and collembolans were
set up. Microcosms contained
15
N labeled litter to track N fluxes into plant shoots. Presence of
decomposers strongly increased total plant and plant shoot biomass. Root biomass decreased
in the presence of collembolans and even more in the presence of earthworms. However, it
increased when both animal groups were present. Also, presence of decomposers increased
total N concentration and
15
N enrichment of grasses, legumes, and small herbs. Small herbs
were at a maximum in the combined treatment with earthworms and collembolans. The
impact of earthworms and collembolans on plant performance strongly varied with plant
functional group identity and plant species diversity and was modified when both decomposers
were present. Both decomposer groups generally increased aboveground plant productivity
through effects on litter decomposition and nutrient mineralization leading to an increased
plant nutrient acquisition. The non-uniform effects of earthworms and collembolans suggest
that functional diversity of soil decomposer animals matters and that the interactions between
soil animal functional groups affect the structure of plant communities.
Key words: Collembola; decomposers; functional diversity; Lumbricidae; plant–animal interaction;
plant diversity; productivity; The Jena Experiment.
INTRODUCTION
In terrestrial ecosystems, soil decomposer animals are
essential for nutrient mineralization (Bradford et al.
2002) and alter the availability of nutrients to plants
(Wardle 1999). While microorganisms dominate miner-
alization processes, their activity, community composi-
tion, and spatial distribution is strongly modified by soil
invertebrates (Scheu and Seta
¨la
¨2002, Bonkowski and
Scheu 2004). For example, feeding of earthworms and
collembolans on bacteria and fungi indirectly affect the
availability of nutrients in soil (Wardle 1999). It is well
documented that the enhanced nutrient turnover in soil
in presence of decomposer animals leads to a higher
plant nutrient acquisition and therefore stimulates plant
growth (Scheu et al. 1999, Schmidt and Curry 1999,
Kreuzer et al. 2004).
Previous studies suggested that some soil animal
species are functionally redundant and do have no
detectable influence on ecosystem functions such as N
mineralization and plant growth (e.g., Cragg and
Bardgett 2001, Liiri et al. 2002). In contrast, Cole et
al. (2004) documented that shoot biomass and total N in
soil leachates increase with richness of soil micro-
arthropods. However, for maintaining ecosystem pro-
cesses the functional characteristics of species likely are
more important than the number of species per se
(Cragg and Bardgett 2001, Cole et al. 2004, Heemsber-
gen et al. 2004), presumably due to differential effects of
animal groups on nutrient fluxes, and on structure and
dynamics of the soil microbial community (Bardgett and
Chan 1999).
In terms of biomass, earthworms are among the most
important detritivore animals in terrestrial ecosystems
(Edwards and Bohlen 1995). Especially in grasslands,
they are known to play a key role in nutrient cycling and
physical soil improvement (Spehn et al. 2000), and
therefore in plant growth (Scheu 2003). Earthworms
influence plant performance either direct, e.g., via root
feeding and translocation of seeds, or indirect via
altering microbial activity and plant nutrient availabil-
ity. Direct effects probably are rare and usually less
Manuscript received 17 January 2006; revised 31 March
2006; accepted 3 April 2006. Corresponding Editor: D. A.
Wardle.
2
E-mail: partsch@bio.tu-darmstadt.de
3
Present address: NERC Centre for Population Biology,
Division of Biology, Imperial College, Silwood Park, Ascot
SL5 7PY United Kingdom.
2548
important than indirect effects. However, the mecha-
nisms responsible for earthworm-mediated changes in
plant performance are poorly studied. Scheu et al. (1999)
and Kreuzer et al. (2004) showed that the effect of
earthworms varies with plant species and is more
pronounced in grasses than in legumes suggesting that
earthworm effects vary with plant functional groups.
Collembolans are among the most abundant soil
arthropods feeding on a range of resources (Hopkin
1997). By consuming dead plant material and fungal
hyphae, collembolans might play an important role in
enhancing decomposition processes since hyphal grazing
stimulates growth and respiration of fungi (Gange
2000). Thus, collembolans predominantly influence
plant performance and plant competition indirectly by
altering microbial activity and microbial community
structure, and therefore the competition for nutrients
between microbes and plants.
Stimulation of plant performance by decomposer
invertebrates is well known, but only few investigations
documented that different functional groups of decom-
poser invertebrates affect plant competition and there-
fore plant community structure (Scheu and Seta
¨la
¨2002,
Wardle 2002, Kreuzer et al. 2004). Studies examining the
role of decomposers for plant growth focused on single
target plant species, competition between two plant
species belonging to different plant functional groups or
specific plant communities. Studies manipulating both
plant diversity and decomposer animal species compo-
sition are lacking.
This study is the first investigating effects of different
functional groups of decomposer invertebrates on plant
communities of different species and functional group
diversity. In a microcosm experiment in the greenhouse
we investigated the effect of two decomposer animal
groups, earthworms and collembolans, on plant growth
and plant nitrogen uptake of communities differing in
plant species and functional group diversity. We
expected the two decomposer groups to strongly but
differently influence plant growth due to different
mechanisms, and to modify each others impact on plant
performance. We hypothesized that the influence of
decomposers on plant productivity varies with plant
functional group identity, since, e.g., the competitive
strength of legumes decreases with increasing nitrogen
concentration in soil. Also, we hypothesized that due to
interactions between plant species and feedbacks of
plant species on decomposers the effects of decomposers
on plant performance vary with plant species and plant
functional group diversity.
MATERIALS AND METHODS
Microcosms
Experimental containers (see Plate 1) consisted of
plastic tubes (inner diameter 10 cm, height 25 cm), sealed
with a 45-lm mesh at the bottom and a plastic barrier
(10 cm height) at the upper end, to prevent escape of
animals. A total of 256 microcosms were filled with soil
(1.4 kg fresh mass) from the Jena Biodiversity Experi-
ment field site (Jena, Thuringia, Germany; Roscher et al.
2004), including a layer of 250 mg of
15
N labeled roots of
Lolium perenne (30 atom %
15
N; fragmented ,1 mm)
placed 2 cm below the soil surface. Prior to use, the soil
(Eutric Fluvisol [FAO-UNESCO 1997]; sand content
15%, water content 13%, pH 8.1, nitrogen content 0.3%,
carbon content 4.6%, C-to-N ratio 15.7) was sieved (4
mm) and defaunated by freezing for two weeks at 228C
(Huhta et al. 1989). Microcosms were incubated in the
greenhouse and after placement into the microcosms the
soil was irrigated by adding two 50-mL portions of
deionized water every second day for eight days to leach
nutrients released as a result of the defaunation
procedure. Subsequently, microcosms were kept moist
PLATE 1. (Top) Photograph taken about one week after
transplanting the pre-germinated plants into the prepared
microcosms, thereby establishing 64 different plant combina-
tions with plant species number varying from 1 to 8 and number
of plant functional groups varying from 1 to 4. (Bottom)
Photograph taken two months later (one month before
harvest), showing a bench with some of the 256 numbered
and randomized microcosms under greenhousde conditions.
Plants are growing well and sporadically flowering. Note the
plastic frame on top of each microcosm to avoid the escape of
animals. Photo credit: S. Partsh.
October 2006 2549DIVERSITY-DEPENDENT DECOMPOSER EFFECTS
for another 14 days adding one portion of 50 mL
deionized water each third day. Plant seedlings germi-
nating (mainly Chenopodium album L.) were removed;
after 14 days, no further seedlings germinated.
Eight pre-germinated plant individuals (height 2–6
cm) of a total of 43 common species of Central
European Arrhenatherion grasslands, consisting of four
functional groups (grasses, G; legumes, L; small herbs,
Sh; and tall herbs, Th) were transplanted into each
microcosm in different combinations similar to the
design of the Jena Experiment (Roscher et al. 2004) but
with plant species diversity varying only from 1 to 8 and
plant functional group diversity from 1 to 4. Altogether,
64 plant species combinations were established includ-
ing monocultures, two species mixtures, four species
mixtures, and eight species mixtures, each with 16
replicates (see Milcu et al. [2006] for details). Sub-
sequently, 2 g of non-labeled litter material consisting
mainly of grass leaves (2.53%N, C-to-N ratio 17.3) were
placed on the soil surface to simulate field conditions.
The litter had been collected from the Jena field site. It
was dried at 608C for three days and cut into pieces of 3
cm length.
Earthworms (one individual of each of Lumbricus
terrestris (L.) and Aporrectodea caliginosa (Savigny))
and collembolans (20 individuals of each of Protaphor-
ura fimata (Gisin), Heteromurus nitidus (Templeton) and
Folsomia candida (Willem)) were added to half of the
microcosms to establish a full two factorial design with
factors earthworms (with and without) and collembo-
lans (with and without). A. caliginosa is an endogeic
geophagous earthworm species, whereas L. terrestris is
an anecic litter feeding species. Both species are among
the dominant species at the field site of the Jena
Experiment. All three collembolan species were taken
from laboratory cultures, where they were kept at
constant temperature (178C) and fed on bakers yeast.
F. candida was originally provided by IBACON GmbH
(Rossdorf, Germany). H. nitidus derived from an arable
field near Go
¨ttingen (Germany) and P. fimata from
Kranichsteiner Wald (Darmstadt, Germany). F. candida
and H. nitidus are hemiedaphic species dwelling in litter
and upper soil layers; P. fimata is a euedaphic species
living in deeper soil layers. H. nitidus is present at the
field site of the Jena Experiment.
Established microcosms were incubated in the green-
house for 11 weeks at a day : night regime of 16:8 h and
228:188C. Light intensity was between 250 and 400
lEm
2
s
1
. The water regime was successively increased
from irrigating three times a week with 25 mL (weeks 1–
2), 40 mL (weeks 3–5), and 50 mL (weeks 5–7) to
irrigating daily with 50 mL (weeks 8–9) and 80 mL
deionized water (weeks 9–11). Microcosms were
randomized every three weeks.
Sampling
Incorporation of soil surface litter was not quantified
but observed by eye during the experiment. After 11
weeks, plants were harvested. Shoots were cut at soil
surface level, separated into species, and dried at 608C
for three days. Roots were washed out of the soil using a
1 mm mesh and dried at 608C for three days. For tracing
incorporation of N from the labeled litter into plants, we
chose one target species of each of three plant functional
groups present in most of the 64 different plant
combinations, i.e., Onobrychis viciifolia (legume), Festu-
ca rubra (grass), and Plantago lanceolata (small herb).
Data on shoot and root biomass were summed up per
microcosm. From total shoot and root biomass, total
plant biomass and shoot-to-root ratio per microcosm
were calculated. Also, the mean individual shoot mass
per microcosm was calculated for each of the plant
functional groups and the three target species.
Total N concentrations and
15
N signatures of the
target species were determined by a coupled system
consisting of an elemental analyzer (NA 1500, Carlo
Erba, Milan, Italy) and a gas isotope mass spectrometer
(MAT 251, Finnigan, Thermo Electron Corporation,
Waltham, Massachusetts, USA; Reineking et al. 1993).
For
15
N atmospheric N
2
served as the primary standard
and acetanilide (C
8
H
9
NO; Merck, Darmstadt, Ger-
many) was used for internal calibration. Earthworms
were collected by hand. Collembolans were sampled
taking soil cores (diameter 5 cm) from each of the
microcosms and extracted by heat. Details on the
response of decomposer animals to variations in plant
species and functional group diversity are presented in
Milcu et al. (2006).
Statistical analyses
The effect of earthworms (EW), collembolans (COL),
number of plant species (S), number of plant functional
groups (FG), and the respective interaction effects on
each variable was determined using type I ANOVA in a
general linear model (GLM). Since only effects on
biomass of plant functional groups and target species
varied with plant diversity, only the respective Fvalues
are given in text and tables. In target species interactions
between decomposer groups and S were calculated
separately from those between decomposers and FG
due to limiting degrees of freedom. Data were tested for
normal distribution and homogeneity of variance
(Levene test) and log
10
(xþ0.1)-transformed if neces-
sary. Means presented in text and graphs were
calculated using non-transformed data (6SE). The
experimental design does not allow full separation of
the effects of S and FG, which are partially confounded.
Therefore, no interaction between S and FG was
calculated. The effects of presence and absence of L,
G, Sh, and Th always were fitted after fitting S and FG.
The Fvalues given in the text and tables for the effects of
S, FG, L, G, Sh, and Th refer to those where the
respective factor, and interaction, was fitted first
(Schmid et al. 2002). Statistical analyses were performed
using SAS 8.2 (SAS Institute 2003).
STEPHAN PARTSCH ET AL.2550 Ecology, Vol. 87, No. 10
RESULTS
Total shoot biomass
Shoot biomass per microcosm increased with increas-
ing plant species diversity, ranging from 5.89 60.28 g to
6.96 60.28 g, and, even more pronounced, with plant
functional group diversity, ranging from 5.80 60.19 g
to 7.77 60.39 g (Table 1). Also, each of the plant
functional groups strongly affected shoot biomass, but
the effects varied with plant species and plant functional
group diversity. Overall, grasses (from 6.74 60.19 g to
5.71 60.20 g) and tall herbs (from 6.51 60.19 g to 5.98
60.21 g) decreased shoot biomass, whereas legumes
(from 4.85 60.15 g to 7.85 60.16 g) and small herbs
(from 5.68 60.19 g to 6.92 60.20 g) increased it. The
generally negative effect of grasses was strongest in the
two species mixtures, whereas the positive effect of
legumes was more pronounced at higher species
diversity levels (Table 1).
Both decomposer groups strongly increased shoot
biomass by ;20%to a very similar level in each of the
treatments (Fig. 1a, Table 1). The effect of decom-
posers on shoot biomass did not significantly vary with
plant diversity. Incorporation of surface litter material
into the soil was highest in earthworm treatments and
less pronounced when only collembolans were present.
Shoot biomass of plant functional groups
Earthworms significantly increased shoot biomass of
grasses per microcosm by 0.33 g (F
1, 108
¼4.94, P¼0.03)
and the biomass of individual shoots of grasses by 0.10 g
(F
1, 108
¼5.03, P¼0.03). Collembolans also tended to
increase shoot biomass of grasses per microcosm (F
1, 108
¼3.44, P¼0.07). Overall, earthworms increased shoot
biomass of tall herbs per microcosm (þ0.26 g; F
1, 108
¼
5.83, P¼0.018), but the effect only occurred in absence
of grasses (significant EW 3G interaction; F
1, 104
¼8.92,
P¼0.004). Collembolans significantly increased shoot
biomass of tall herbs per microcosm by 0.31 g (F
1, 108
¼
7.84, P¼0.006) and the biomass of individual shoots of
tall herbs by 0.09 g (F
1, 108
¼7.63, P¼0.007). The effect
on shoot biomass of tall herbs per microcosm was most
pronounced in the combined treatment with earthworms
and collembolans (þ0.55 g). Earthworms (þ0.70 g) and
collembolans (þ0.97 g) tended to increase shoot biomass
of legumes per microcosm but this was less pronounced
in the combined treatment with earthworms and
collembolans (þ0.24 g; F
1, 108
¼3.75, P¼0.055). Also,
earthworms tended to increase the biomass of individual
shoots of legumes in absence of small herbs (marginally
significant EW 3Sh interaction; F
1, 105
¼3.33, P¼
0.071). Biomass of small herbs was not significantly
affected by the presence of decomposers.
Shoot biomass of target species
Earthworms increased the biomass of Festuca rubra
per microcosm from 2.45 60.43 g to 2.96 60.43 g
(Table 2). Also, earthworms increased the biomass of
individual shoots of Festuca rubra in the one (þ0.12 g),
two (þ0.22 g), and eight species mixtures (þ0.73 g) and
in mixtures with one (þ0.27 g) and three plant
functional groups (þ0.73 g; significant EW 3S
interaction and significant EW 3FG interaction; Table
2). The positive effect of earthworms on mixtures with
eight species (þ1.84 g) and three functional groups
(þ1.84 g), was most pronounced in absence of collem-
bolans (significant EW 3COL 3S interaction and
significant EW 3COL 3FG interaction; Table 2).
Earthworms tended to increase the biomass of Plantago
lanceolata per microcosm but the biomass of individual
shoots was not affected by decomposers. Earthworms
increased the biomass of Onobrychis viciifolia per
TABLE 1. ANOVA table for the effects of earthworms (EW), collembolans (COL), number of plant species (S), number of plant
functional groups (FG), and presence of legumes (L), grasses (G), small herbs (Sh), and tall herbs (Th) on shoot biomass, root
biomass, total plant biomass, and shoot-to-root ratio.
Parameter df
Shoot biomass Root biomass Total biomass Shoot-to-root ratio
FPFPFP F P
EW 1, 243 8.24 0.0045 0.07 0.7944 2.06 0.1529 5.20 0.0234
COL 1, 243 7.49 0.0067 0.83 0.3637 5.30 0.0222 0.40 0.5294
EW 3COL 1, 243 6.44 0.0188 6.79 0.0097 0.26 0.6138 10.15 0.0016
FG 3, 243 18.44 ,0.0001 0.48 0.6971 4.93 0.0024 4.64 0.0036
S 3, 243 9.66 ,0.0001 2.55 0.0561 0.64 0.5886 5.08 0.0020
L 1, 243 232.49 ,0.0001 0.00 0.9641 88.85 ,0.0001 46.77 ,0.0001
L3FG 2, 238 6.58 0.0017 0.77 0.4631 2.36 0.0967 2.18 0.1152
L3S 3, 238 24.35 ,0.0001 2.64 0.0504 5.90 0.0007 9.72 ,0.0001
G 1, 243 130.15 ,0.0001 6.98 0.0088 21.37 ,0.0001 48.74 ,0.0001
G3FG 2, 238 17.67 ,0.0001 2.64 0.0738 12.88 ,0.0001 2.93 0.0553
G3S 3, 238 4.24 ,0.0001 2.49 0.0609 6.07 0.0005 0.62 0.6017
Sh 1, 243 12.29 0.0005 7.03 0.0085 0.00 0.9920 14.65 0.0002
Sh 3FG 2, 238 1.47 0.2314 0.42 0.6567 2.22 0.1103 0.09 0.9152
Sh 3S 3, 238 4.16 0.0067 3.37 0.0193 2.74 0.0443 4.50 0.0043
Th 1, 243 53.94 ,0.0001 0.00 0.9563 23.17 ,0.0001 13.58 0.0003
Th 3FG 2, 238 1.56 0.2114 0.90 0.4061 1.19 0.3070 0.90 0.4082
Th 3S 3, 238 2.34 0.0742 4.19 0.0065 3.60 0.0141 2.12 0.0989
October 2006 2551DIVERSITY-DEPENDENT DECOMPOSER EFFECTS
microcosm in all (one, þ0.36 g; two, þ0.58 g; four,
þ0.12 g) but the eight plant species mixtures (0.16 g;
significant EW 3S interaction; Table 2), whereas
collembolans increased it in all (one, þ0.51 g; four,
þ0.22 g; eight, þ0.35 g) but the two plant species
mixtures (1.11 g; significant COL 3S interaction;
Table 2). The positive effect of each of the decomposer
groups on the biomass of Onobrychis viciifolia in
monocultures was less pronounced in the presence of
the respective other decomposer group (significant EW
3COL 3S interaction; Table 2). In addition, the effect
of collembolans in the two species mixtures was less
detrimental in presence of earthworms. Earthworms
also tended to interact with small herbs in affecting
biomass of individual shoots of Onobrychis viciifolia
(marginally significant EW 3Sh interaction; F
1,18
¼
4.22, P¼0.055). Biomass of individual shoots was
highest when small herbs were present but earthworms
absent. Presence of earthworms reduced this beneficial
effect of small herbs. On average, collembolans in-
creased the biomass of individual shoots of Onobrychis
viciifolia by 0.20 g (Table 2).
FIG. 1. Effects of earthworms and collembolans on (a) shoot biomass, (b) root biomass, (c) total plant biomass, (d) shoot-to-
root ratio, (e) total N content of Festuca rubra, and (f)
15
N enrichment of F. rubra. Abbreviations: ctrl, control treatment; ew,
earthworms only; col, collembolans only; ewþcol, combined treatment with earthworms and collembolans. Values are means 6SE;
for statistical analyses see Tables 1 and 3.
STEPHAN PARTSCH ET AL.2552 Ecology, Vol. 87, No. 10
Total N content and
15
N enrichment of target species
In treatments with decomposers, total N concentra-
tion and
15
N enrichment in Festuca rubra,Onobrychis
viciifolia, and Plantago lanceolata were generally in-
creased compared to the control treatment without
earthworms and collembolans. Variations in total N and
15
N concentrations were very similar in each of the three
plant species studied; the total amount of N was at a
TABLE 2. ANOVA table for effects of earthworms (EW), collembolans (COL), number of plant species (S), and number of plant
functional groups (FG) on shoot biomass per microcosm and the mean biomass of individual shoots of Festuca rubra,
Onobrychis viciifolia, and Plantago lanceolata.
Parameter
F. rubra O. viciifolia P. lanceolata
df FPdf FPdf FP
Shoot biomass per microcosm
EW 1, 15 5.21 0.0375 1, 21 0.00 0.9520 1, 38 3.34 0.0756
COL 1, 15 0.51 0.4867 1, 21 0.18 0.6740 1, 38 0.00 0.9489
EW 3COL 1, 15 1.44 0.2493 1, 21 0.20 0.6573 1, 38 2.20 0.1460
FG 3, 15 30.40 ,0.0001 2, 21 30.78 ,0.0001 2, 38 50.64 ,0.0001
EW 3FG 3, 6 4.67 0.0518 2, 15 0.26 0.7733 2, 32 2.73 0.0805
COL 3FG 3, 6 0.35 0.7899 2, 15 0.03 0.9698 2, 32 0.01 0.9927
EW 3COL 3FG 3, 6 1.83 0.2418 2, 15 0.33 0.7256 2, 32 0.50 0.6089
S 3, 15 47.26 ,0.0001 3, 21 97.53 ,0.0001 3, 38 50.24 ,0.0001
EW 3S 3, 6 1.91 0.2292 3, 12 4.88 0.0192 3, 29 2.89 0.0523
COL 3S 3, 6 0.45 0.7266 3, 12 19.01 ,0.0001 3, 29 0.77 0.5225
EW 3COL 3S 3, 6 2.06 0.2078 3, 12 32.98 ,0.0001 3, 29 1.15 0.3471
Mean biomass of individual shoots
EW 1, 15 2.00 0.1773 1, 21 0.94 0.3434 1, 38 0.17 0.6805
COL 1, 15 0.06 0.8146 1, 21 9.14 0.0065 1, 38 0.28 0.5977
EW 3COL 1, 15 2.43 0.1396 1, 21 1.17 0.2915 1, 38 0.50 0.4830
FG 3, 15 9.44 0.0010 2, 21 5.76 0.0101 2, 32 0.16 0.8552
EW 3FG 3, 6 12.37 0.0056 2, 15 1.07 0.3675 2, 32 1.18 0.3191
COL 3FG 3, 6 0.13 0.9413 2, 15 0.72 0.5031 2, 32 0.15 0.8595
EW 3COL 3FG 3, 6 26.51 0.0007 2, 15 0.99 0.3932 2, 32 0.43 0.6550
S 3, 15 8.92 0.0012 3, 21 2.31 0.1054 3, 38 1.60 0.2046
EW 3S 3, 6 8.93 0.0125 3, 12 1.27 0.3283 3, 29 2.10 0.1223
COL 3S 3, 6 0.07 0.9741 3, 12 3.17 0.0639 3, 29 0.43 0.7320
EW 3COL 3S 3, 6 20.90 0.0014 3, 12 0.27 0.8475 3, 29 0.67 0.5803
TABLE 3. ANOVA table for the effects of earthworms (EW), collembolans (COL), number of plant species (S), and number of
plant functional groups (FG) on total N and
15
N concentration of Festuca rubra,Onobrychis viciifolia, and Plantago lanceolata.
Parameter
F. rubra O. viciifolia P. lanceolata
df FPdf FPdf FP
Total N content
EW 1, 15 9.58 0.0074 1, 22 2.01 0.1699 1, 35 8.92 0.0051
COL 1, 15 0.17 0.6825 1, 22 0.48 0.4965 1, 35 1.10 0.3022
EW 3COL 1, 15 1.87 0.1919 1, 22 0.05 0.8201 1, 35 3.65 0.0642
FG 3, 15 1.65 0.2193 2, 22 1.63 0.2193 2, 35 1.57 0.2229
EW 3FG 3, 6 0.78 0.5446 2, 16 0.25 0.7819 2, 29 0.46 0.6353
COL 3FG 3, 6 2.00 0.2159 2, 16 0.48 0.6281 2, 29 2.03 0.1493
EW 3COL 3FG 3, 6 3.27 0.1012 2, 16 0.28 0.7570 2, 29 1.40 0.2636
S 3, 15 1.29 0.3155 3, 22 0.93 0.4417 3, 35 0.65 0.5874
EW 3S 3, 6 3.16 0.1073 3, 13 1.23 0.3328 3, 26 0.80 0.5071
COL 3S 3, 6 4.24 0.0626 3, 13 0.47 0.7114 3, 26 0.49 0.6942
EW 3COL 3S 3, 6 5.55 0.0364 3, 13 0.29 0.8310 3, 26 0.68 0.5732
15
N enrichment
EW 1, 15 154.44 ,0.0001 1, 22 1846.90 ,0.0001 1, 35 3.62 0.0688
COL 1, 15 522.19 ,0.0001 1, 22 8200.81 ,0.0001 1, 35 5.86 0.0231
EW 3COL 1, 15 24.25 0.0002 1, 22 6.63 0.0173 1, 35 0.65 0.4290
FG 3, 15 16.9 ,0.0001 2, 22 180.57 ,0.0001 2, 35 4.99 0.0151
EW 3FG 3, 6 5.29 0.0402 2, 16 1.36 0.2851 2, 29 2.92 0.0698
COL 3FG 3, 6 13.54 0.0044 2, 16 0.95 0.4063 2, 29 2.14 0.1353
EW 3COL 3FG 3, 6 2.53 0.1540 2, 16 2.07 0.1587 2, 29 0.98 0.3882
S 3, 15 17.16 ,0.0001 3, 22 180.15 ,0.0001 3, 35 1.34 0.2845
EW 3S 3, 6 8.41 0.0144 3, 13 0.04 0.9906 3, 26 5.39 0.0051
COL 3S 3, 6 21.40 0.0013 3, 13 0.01 0.9981 3, 26 9.26 0.0002
EW 3COL 3S 3, 6 4.06 0.0682 3, 13 0.10 0.9603 3, 26 0.48 0.7008
October 2006 2553DIVERSITY-DEPENDENT DECOMPOSER EFFECTS
maximum in the earthworm only treatment and the
enrichment in
15
N in the combined treatment with
earthworms and collembolans (Fig. 1e, f, Table 3). The
total amount of N in Festuca rubra and Plantago
lanceolata was significantly increased by earthworms
(from 7800 6500 lg/g to 9800 6500 lg/g and from
7000 6500 lg/g to 8900 6500 lg/g, respectively). The
effect of earthworms on total N concentration of
Festuca rubra varied with plant species diversity. Except
for mixtures with four plant species it was most
pronounced in absence of collembolans (significant
EW 3COL 3S interaction; Fig. 2c, Table 3).
The incorporation of
15
N into plant tissue of Festuca
rubra and Onobrychis viciifolia was significantly increased
by earthworms and collembolans, being at a maximum in
the combined treatment (significant EW 3COL inter-
action; Fig. 1f, Table 3). For
15
N enrichment in Festuca
rubra, the effect of both decomposer groups varied with
plant species and plant functional group diversity. Both
decomposer groups increased the incorporation of
15
N
into plant tissue of Festuca rubra to a maximum in
mixtures with eight species (EW, from 47 619 lg/g to 69
619 lg/g; COL, from 36 610 lg/g to 80 610 lg/g) and
three plant functional groups (EW, from 47 619 lg/g to
69 619 lg/g; COL, from 36 610 lg/g to 80 610 lg/g).
For Plantago lanceolata, only collembolans significantly
increased
15
N enrichment from 600 640 lg/g to 700 6
40 lg/g. The effects of both earthworms and collembo-
lans on the
15
N enrichment in Plantago lanceolata varied
with plant species diversity (Fig. 2a, b, Table 3). Earth-
worms increased
15
N incorporation into shoots of
Plantago lanceolata in all but the one species mixtures,
FIG. 2. Effects of (a) earthworms and (b) collembolans on
15
N enrichment of Plantago lanceolata, and (c) effects of earthworms
and collembolans on total N concentration in tissue of Festuca rubra at varying plant species diversities. The solid line shows values
with earthworms (a) or collembolans (b and c); the dashed line shows values without earthworms (a) or collembolans (b and c).
Error bars show 6SE; for statistical analyses, see Table 3. Abbreviations: þew, with earthworms; ew, without earthworms.
STEPHAN PARTSCH ET AL.2554 Ecology, Vol. 87, No. 10
whereas collembolans increased it in all but the four
species mixtures.
Root biomass
Presence of collembolans (0.43 g) and in particular
that of earthworms reduced root biomass (0.72 g).
However, root biomass increased by 0.16 g when both
animal groups were present (significant EW 3COL
interaction; Fig. 1b, Table 1). None of the decomposer
effects on root biomass was significantly affected by
plant diversity.
Total plant biomass
Total plant biomass per microcosm increased in
presence of decomposers being at a maximum in the
combined treatment with earthworms and collembolans
(Fig. 1c, Table 1). However, only the effect of
collembolans was significant (Table 1) since the com-
bined effect of earthworms and collembolans was altered
in presence of legumes (significant EW 3COL 3L
interaction; F
1, 240
¼5.53, P¼0.02; Fig. 3). The effect of
decomposers on total plant biomass did not vary
significantly with plant diversity.
Shoot-to-root ratio
Earthworms increased the shoot-to-root ratio, but the
effect was less pronounced when collembolans were also
present (Fig. 1d, Table 1). However, presence of
collembolans also tended to increase plant shoot-to-
root ratio. Since neither the effect of decomposers on
shoot nor root biomass varied with plant diversity the
shoot-to-root ratio was also unaffected by plant
diversity.
DISCUSSION
Plant diversity
Recent field studies documented that net primary
productivity and nutrient retention in ecosystems
increase as the number of plant species increases
(Hooper and Vitousek 1997, Hector et al. 1999, Spehn
et al. 2005). Major mechanisms responsible for the
increase in productivity with diversity are the comple-
mentary use of resources by plants, facilitative inter-
actions, and sampling effect (Spehn et al. 2005). In the
present greenhouse study, plant species and plant
functional group diversity also strongly influenced plant
performance, especially aboveground productivity.
These effects on plant productivity resembled those
from the main experiment of the Jena Experiment
(Roscher et al. 2005) suggesting that our laboratory
systems represented the field situation. Overall, shoot
biomass increased with plant species and plant func-
tional group diversity and in addition to plant richness
per se, further variance was explained by plant
community composition, i.e. the identity of functional
groups. The specific traits of plant functional groups
differently influenced plant productivity per microcosm
and most of the effects varied with plant species
diversity. As suggested previously (Tilman et al. 1997,
Diaz and Cabido 2001) for plant productivity, func-
tional traits of plants presumably are more important
than plant species and plant functional group diversity
per se.
Earthworms and collembolans
Soil microorganisms dominate mineralization pro-
cesses and compete with plants for nutrients (Kaye and
FIG. 3. Effects of earthworms, collembolans, and legumes on total plant biomass. Abbreviations: þew, with earthworms; þleg,
with legumes; ew, without earthworms; leg, without legumes. The solid line shows values with collembolans; the dashed line
shows values without collembolans. Error bars show 6SE; for statistical analyses see Results: Total plant biomass.
October 2006 2555DIVERSITY-DEPENDENT DECOMPOSER EFFECTS
Hart 1997). Soil invertebrates affect the soil microbial
community and functioning directly by grazing but also
indirectly by changing nutrient availability and soil
structure; both direct and indirect effects are known to
affect plant performance and ecosystem processes
(Scheu 2001, Bonkowski and Scheu 2004). Results of
the present study suggest that both earthworms and
collembolans increase plant performance mainly
through enhanced nutrient mineralization and an
accompanying increase in plant nutrient acquisition.
Incorporation of surface litter material into the soil was
highest in earthworm treatments and less pronounced
when only collembolans were present. Despite their
differential effects on litter dynamics both decomposer
groups had a similar positive influence on aboveground
plant productivity suggesting that their effects were
based on different mechanisms. Effects of earthworms
are likely to be related to N mobilization (Scheu 2003),
while collembolans are likely to affect plant growth by
altering the microbial environment of roots (Lussenhop
1992).
As hypothesized, plants from different functional
groups were differently influenced by decomposer
presence. Consistent with previous experiments (Scheu
et al. 1999, Schmidt and Curry 1999, Kreuzer et al.
2004), earthworms generally enhanced plant growth, in
particular that of grasses, less that of legumes. Earth-
worms increased grass biomass per microcosm, the
biomass of individual grass shoots, and also tall herb
biomass per microcosm, but only when grasses were
absent. Earthworms did not affect the biomass of
individual shoots of tall herbs, suggesting that in
contrast to grasses earthworms stimulated the growth
of a single or a few, but not of all tall herb species.
Presumably, plant species more independent of soil N
responded less since earthworms alter plant growth by
increasing N mineralization. Furthermore, there is
evidence that legumes are less efficient in exploiting soil
N than, e.g., grasses (Kang 1988). Increased N concen-
trations in shoots of Festuca rubra and Plantago
lanceolata in presence of earthworms support the
conclusion that the earthworm-mediated increase in
plant growth was due to increased N mobilization.
Collembolans also beneficially affect plant growth
(Scheu et al. 1999, Gange 2000, Kreuzer et al. 2004). In
contrast to earthworms, collembolans influence plant
growth via feeding on fungi. However, the mechanisms
for collembolan-mediated changes in plant growth still
are little understood. Gange (2000) and Lussenhop and
BassiriRad (2005) documented that the effect of
collembolans on plant growth depends on density.
Moderate grazing on mycorrhizal fungi stimulates the
activity of the fungi and therefore enhances plant growth
(Lussenhop 1996). In our study increased shoot bio-
mass, the accompanying increase in the shoot-to-root
ratio, and the increased
15
N uptake by the three target
species studied support the conclusion of Bardgett and
Chan (1999) and Gange (2000) that collembolans
stimulate N mineralization. Collembolans predomi-
nantly increased shoot biomass of tall herbs, but also
the mean individual shoot mass of Onobrychis viciifolia.
The strong
15
N enrichment in shoot tissue of Onobrychis
viciifolia supports the conclusion of Lussenhop (1993)
that collembolans affect N acquisition of legumes,
potentially by altering nodule occupancy. However,
despite legumes benefited from collembolans, the total
number of collembolans in presence of legumes was
significantly reduced by 23%(Milcu et al. 2006),
suggesting that they in turn suffer from the presence of
legumes.
Plants not only responded to the presence of
decomposers with increasing shoot biomass but also
by changing plant resource allocation. Compared to
roots earthworms disproportionately increased shoot
biomass, resulting in an increased shoot-to-root ratio.
The reduction in root biomass in presence of each of the
decomposer groups presumably reflects increased nu-
trient availability. Reduced root biomass in presence of
collembolans has been observed in a number of studies
(Larsen and Jakobsen 1996, Bardgett and Chan 1999,
Scheu et al. 1999, Kreuzer et al. 2004) and has been
ascribed to a collembolan-mediated increase in plant
nutrient uptake (Lussenhop and BassiriRad 2005).
Reviewing published data Scheu (2003) reported that
the effect of earthworms on root biomass is inconsistent;
it was increased in most studies but in a number of
studies it was reduced. Apart from enhanced nutrient
mineralization, the reduction in root biomass in
presence of earthworms also may have resulted from
reduced competition for nutrients between microbes and
plant roots; indeed, in our experiment microbial biomass
and respiration was significantly reduced in earthworm
treatments and this may have favored plant nutrient
acquisition (Milcu et al. 2006).
Consistent with our hypothesis, both decomposers
interacted in affecting plant performance. Compared to
their individual effect, earthworms and collembolans
combined further increased
15
N enrichment and total
plant biomass. Since both, earthworms and collembo-
lans, affected shoot biomass to a similar extent, the
increase in total biomass and the narrower shoot-to-root
ratio in the combined treatment were due to differences
in root biomass. Therefore, not only shoot but also root
biomass needs to be taken into account for under-
standing decomposer effects on plant communities. The
complementary effects of earthworms and collembolans
on total plant biomass and
15
N enrichment presumably
resulted from additivity of the effects of each of the
decomposer groups and indicate that the two groups of
decomposers are not functionally redundant, but
influence plant performance through different mecha-
nisms thereby adding to the effects of the other. In the
present study collembolans facilitated resource acquis-
ition by earthworms supporting the view of positive
feedbacks between the two decomposer groups (Milcu et
al. 2006).
STEPHAN PARTSCH ET AL.2556 Ecology, Vol. 87, No. 10
Despite both decomposer groups alone reduced root
biomass, it was increased in the combined treatment
with earthworms and collembolans. Potentially, pres-
ence of both decomposer groups stimulated the ex-
ploitation of nutrients by plants thereby increasing root
biomass. In presence of legumes the tissue N concen-
tration of A. caliginosa was independent of the presence
of collembolans, but the amount of
15
NinA. caliginosa
tissue decreased when collembolans were also present
(Milcu et al. 2006). This indicates that in presence of
collembolans earthworms incorporated more N derived
from atmospheric fixation by legumes. Thus, earth-
worms may have competed with neighbouring non-
legume plant roots for the symbiotically fixed N,
resulting in increased root competition.
Again, consistent with our hypothesis, the effect of
both decomposer groups and their interaction varied
with plant diversity. But in contrast to our expectation,
this variation was only reflected at the target species
level. Presumably, due to variable effects of decom-
posers on plant species even within functional groups
neither the effect of earthworms nor that of collembo-
lans significantly varied with plant species and func-
tional group diversity at the microcosm level. For
example, total N concentration in tissue of Festuca
rubra was increased in presence of earthworms at all
diversity levels. In four species mixtures, the additional
presence of collembolans further increased total N
concentration, while in eight species mixtures, the
additional presence of collembolans decreased it. Since
earthworms were more active in more diverse plant
communities (Milcu et al. 2006), their facilitative effects
on at least one collembolan species, presumably lead to
an optimum grazing effect by collembolans in four
species but a negative one in eight species mixtures.
In conclusion, consistent with previous studies in the
field including the Jena Experiment, results of our
laboratory study proved that functional traits of
individual plant species are more important for above-
ground plant productivity than plant species and plant
functional group diversity per se. Generally, both
earthworms and collembolans beneficially affected
aboveground plant performance and interacted in
affecting plant belowground performance. However,
the effect varied with plant functional group identity
and for single plant species also with plant diversity.
Also, as we hypothesized the plant functional compo-
sition modified decomposer effects on plant perform-
ance suggesting feedbacks between plants and
decomposers. Our results support suggestions on the
key role of earthworms and collembolans as decom-
posers for plant nutrient uptake and primary produc-
tion. Both the increased shoot biomass and the reduced
root biomass suggest a higher N mobilization in
decomposer treatments. In addition to their individual
effects, earthworms and collembolans interacted in
affecting total plant biomass, root biomass and
15
N
enrichment of shoots. Obviously, plants differentially
respond to the presence of different decomposer animal
groups. Both the non-uniform and complementary
effects of earthworms and collembolans on plant
performance indicate that functional diversity of soil
invertebrates is important for ecosystem functioning.
ACKNOWLEDGMENTS
We thank Diana Capota, Kerstin Peschel, and Nico
Eisenhauer for their help on establishing, maintaining, and
finalizing the experiment, and Reinhard Langel (Kompetenz-
zentrum Stabile Isotope, Go
¨ttingen, Germany) for analyzing
15
N samples. We gratefully acknowledge financial support by
the German Science Foundation (FOR 456; The Jena Experi-
ment).
LITERATURE CITED
Bardgett, R. D., and K. F. Chan. 1999. Experimental evidence
that soil fauna enhance nutrient mineralization and plant
nutrient uptake in montane grassland ecosystems. Soil
Biology and Biochemistry 31:1007–1014.
Bonkowski, M., and S. Scheu. 2004. Biotic interactions in the
rhizosphere: effects on plant growth and herbivore develop-
ment. Pages 71–91 in W. W. Weisser and E. Sieman, editors.
Insects and ecosystem function. Ecological Studies 173.
Springer-Verlag, Berlin-Heidelberg, Germany.
Bradford, M. A., et al. 2002. Impacts of soil faunal community
composition on model grassland ecosystems. Science 298:
615–618.
Cole, L., P. L. Staddon, D. Sleep, and R. D. Bardgett. 2004.
Soil animals influence microbial abundance, but not plant–
microbial competition for soil organic nitrogen. Functional
Ecology 18:631–640.
Cragg, R. G., and R. D. Bardgett. 2001. How changes in soil
faunal diversity and composition within a trophic group
influence decomposition processes. Soil Biology and Bio-
chemistry 33:2073–2081.
Diaz, S., and M. Cabido. 2001. Vive la difference: plant
functional diversity matters to ecosystem processes. Trends in
Ecology and Evolution 16:646–655.
Edwards, C. A., and P. Bohlen. 1995. Biology and ecology of
earthworms. Chapman and Hall, New York, New York,
USA.
FAO-UNESCO. 1997. Soil map of the world. Revised legend
with corrections and update. ISRIC, Wageningen, The
Netherlands.
Gange, A. 2000. Arbuscular mycorrhizal fungi, collembola and
plant growth. Trends in Ecology and Evolution 15:396–372.
Hector, A., et al. 1999. Plant diversity and productivity
experiments in European grasslands. Science 286:1123–1127.
Heemsbergen, D. A., M. P. Berg, M. Loreau, J. R. van Hal, J.
H. Faber, and H. A. Verhoef. 2004. Biodiversity effects on
soil processes explained by interspecific functional dissim-
ilarity. Science 306:1019–1020.
Hooper, D. U., and P. M. Vitousek. 1997. The effects of plant
composition and diversity on ecosystem processes. Science
277:1302–1305.
Hopkin, S. P. 1997. Biology of the springtails. Oxford
University Press, New York, New York, USA.
Huhta,V.,D.H.Wright,andD.C.Coleman.1989.
Characteristics of defaunated soil. I. A comparison of three
techniques applied to two different forest soils. Pedobiologia
33:415–424.
Kang, B. T. 1988. Nitrogen cycling in multiple cropping
systems. Pages 333–348 in J. R. Wilson, editor. Advances in
nitrogen cycling in agricultural ecosystems. CAB Interna-
tional, Wallingford, UK.
Kaye, J. P., and S. C. Hart. 1997. Competition for nitrogen
between plants and soil microorganisms. Trends in Ecology
and Evolution 12:139–143.
October 2006 2557DIVERSITY-DEPENDENT DECOMPOSER EFFECTS
Kreuzer, K., M. Bonkowski, R. Langel, and S. Scheu. 2004.
Decomposer animals (Lumbricidae, Collembola) and organic
matter distribution affect the performance of Lolium perenne
(Poaceae) and Trifolium repens (Fabaceae). Soil Biology and
Biochemistry 36:2005–2011.
Larsen, J., and I. Jakobsen. 1996. Effects of a mycophagous
collembola on the symbioses between Trifolium subterraneum
and three arbuscular mycorrhizal fungi. New Phytologist
133:295–302.
Liiri, M., H. Seta
¨la
¨, J. Haimi, T. Pennanen, and H. Fritze. 2002.
Relationship between soil microarthropod species diversity
and plant growth does not change when the system is
disturbed. Oikos 96:137–149.
Lussenhop, J. 1992. Mechanisms of microarthropod–microbial
interactions in soil. Advances in Ecological Research 23:1–33
Lussenhop, J. 1993. Effects of two Collembola species on
nodule occupancy by two Bradyrhizobium japonicum strains.
Soil Biology and Biochemistry 25:775–780.
Lussenhop, J. 1996. Collembola as mediators of microbial
symbiont effects upon soybean. Soil Biology and Biochem-
istry 28:363–369.
Lussenhop, J., and H. BassiriRad. 2005. Collembola effects on
plant mass and nitrogen acquisition by ash seedlings
(Fraxinus pennsylvanica). Soil Biology and Biochemistry 37:
645–650.
Milcu, A., S. Partsch, and S. Scheu. 2006. The response of
decomposers (earthworms, springtails and microorganisms)
to variation in species and functional group diversity of
plants. Oikos 112:513–524.
Reineking, A., R. Langel, and J. Schikowski. 1993.
15
N,
13
C-on-
line measurements with an elemental analyzer (Carlo Erba,
NA 1500), a modified trapping box and a gas isotope mass
spectrometer (Finnigan, MAT 251). Isotopes in Environ-
mental Health Studies 29:169–174.
Roscher, C., J. Schumacher, J. Baade, W. Wilcke, G. Gleixner,
W. W. Weisser, B. Schmid, and E. D. Schulze. 2004. The role
of biodiversity for element cycling and trophic interactions:
an experimental approach in a grassland community. Basic
and Applied Ecology 5:107–121.
Roscher, C., V. M. Temperton, M. Scherer-Lorenzen, M.
Schmitz, J. Schumacher, B. Schmid, N. Buchmann, W. W.
Weisser, and E. D. Schulze. 2005. Overyielding in exper-
imental grassland communities—irrespective of species pool
or spatial scale. Ecology Letters 8:419–429.
SAS Institute. 2003. SAS 8.2. SAS Institute, Cary, North
Carolina, USA.
Scheu, S. 2001. Plants and generalist predators as links between
the below-ground and above-ground system. Basic and
Applied Ecology 2:3–13.
Scheu, S. 2003. Effects of earthworms on plant growth: patterns
and perspectives. Pedobiologia 47:846–856.
Scheu, S., and H. Seta
¨la
¨. 2002. Multitrophic interactions in
decomposer communities. Pages 223–264 in T. Tscharntke
and B. A. Hawkins, editors. Multitrophic level interactions.
Cambridge University Press, Cambridge, UK.
Scheu, S., A. Theenhaus, and T. H. Jones. 1999. Links between
the detritivore and the herbivore system: effects of earth-
worms and Collembola on plant growth and aphid develop-
ment. Oecologia 119:541–551.
Schmid, B., A. Hector, M. Huston, P. Inchausti, I. Nijs, P. W.
Leadley, and D. Tilman. 2002. The design and analysis of
biodiversity experiments. Pages 61–75 in M. Loreau, S.
Naeem, and P. Inchausti, editors. Biodiversity and ecosystem
functioning. Oxford University Press, New York, New York,
USA.
Schmidt, O., and J. P. Curry. 1999. Effects of earthworms on
biomass production, nitrogen allocation and nitrogen trans-
fer in wheat-clover intercropping model systems. Plant and
Soil 214:187–198.
Spehn, E. M., J. Joshi, B. Schmid, J. Alphei, and C. Ko
¨rner.
2000. Plant diversity effects on soil heterotrophic activity in
experimental grassland ecosystems. Plant and Soil 224:217–
230.
Spehn, E. M., et al. 2005. Ecosystem effects of biodiversity
manipulations in European grasslands. Ecological Mono-
graphs 75:37–63.
Tilman, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, and E.
Sieman. 1997. The influence of functional diversity and
composition on ecosystem processes. Science 277:1300–1302.
Wardle, D. A. 1999. How soil food webs make plants grow.
Trends in Ecology and Evolution 14:418–420.
Wardle, D. A. 2002. Linking the aboveground and below-
ground components. Princeton University Press, Princeton,
New Jersey, USA.
STEPHAN PARTSCH ET AL.2558 Ecology, Vol. 87, No. 10