Oecologia
DOI 10.1007/s00442-013-2607-x
GLOBAL CHANGE ECOLOGY - ORIGINAL RESEARCH
The native–invasive balance: implications for nutrient cycling
in ecosystems
Jonathan E. Hickman • Isabel W. Ashton
Katherine M. Howe • Manuel T. Lerdau
•
Received: 4 September 2012 / Accepted: 28 January 2013
Ó Springer-Verlag Berlin Heidelberg 2013
Abstract We conducted single- and mixed-litter experiments in a hardwood forest in Long Island, New York,
using leaf litter from phylogenetically paired native and
invasive species. We selected long-established, abundant
invasive species with wide-ranging distributions in the
eastern United States that likely make substantial contributions to the litter pool of invaded areas. Overall, leaf
litter from invasive species differed from native litter,
though differences varied by phylogenetic grouping.
Invasive litter had lower carbon:nitrogen ratios
(30.9 ± 1.96 SE vs. 32.8 ± 1.36, P = 0.034) and invasive
Communicated by Michael Madritch.
J. E. Hickman I. W. Ashton K. M. Howe M. T. Lerdau
Department of Ecology and Evolution, Stony Brook University,
Stony Brook, NY 11794-5245, USA
Present Address:
J. E. Hickman (&)
Agriculture and Food Security Center, Earth Institute at
Columbia University, Palisades, NY 10964, USA
e-mail: [email protected]
Present Address:
I. W. Ashton
Northern Great Plains Inventory and Monitoring Network,
National Park Service, 231 East St. Joseph Street, Rapid City,
SD 57701, USA
species lost 0.03 ± 0.007 g of nitrogen and had
23.4 ± 4.9 % of their starting mass remaining at the end of
1 year compared with a loss of 0.02 ± 0.003 g nitrogen
and 31.1 ± 2.6 % mass remaining for native species.
Mixing litter from two species did not alter decomposition
rates when native species were mixed with other native
species, or when invasive species were mixed with other
invasive species. However, mixing litter of native and
invasive species resulted in significantly less mass and
nitrogen loss than was seen in unmixed invasive litter.
Mixtures of native and invasive litter lost all but
47 ± 2.2 % of initial mass, compared to 37 ± 5.8 % for
invasive litter and 50 ± 5.1 % for native litter. This nonadditive effect of mixing native and invasive litter suggests
that an additive model of metabolic characteristics may not
suffice for predicting invasion impacts in a community
context, particularly as invasion proceeds over time.
Because the more rapid decomposition of invasive litter
tends to slow to rates typical of native species when native
and invasive litters are mixed together, there may be little
impact of invasive species on nutrient cycling early in an
invasion, when native leaf litter is abundant (providing
litter deposition is the dominant control on nutrient
cycling).
Keywords Decomposition Litter Nitrogen Invasive
species Litter mixture
Present Address:
K. M. Howe
Deparment of Botany and Plant Pathology, Purdue University,
915 W. State St., West Lafayette, IN 47907, USA
Introduction
Present Address:
M. T. Lerdau
Departments of Environmental Sciences and Biology, University
of Virginia, Clark Hall, Charlottesville, VA 22904-4123, USA
One of the challenges facing both ecosystem ecology and
conservation science is to understand the effects that particular species have on ecosystem properties. Over the past
several decades, an exciting paradigm has been developed
123
Oecologia
and tested that relates the chemical, physiological, and
ecological characteristics of a species to decomposition of
its tissue and effects on element cycling (Chapin 1980;
Vitousek 1982; Hobbie 1992). More recently, studies of
invasive species have noted that many invaders tend to
have chemical characteristics associated with high photosynthetic and growth rates and tend to have correspondingly high litter decomposition rates (Ehrenfeld 2003; Liao
et al. 2008). Implicit in these studies and in the paradigm of
species-level impacts is the idea that species-level effects
will be additive and that there will not be significant
chemical and/or biological interactions among litter of
different species, so that knowing the properties of individual species and their relative abundances will suffice for
predicting ecosystem properties. However, in the last several years, multiple studies have suggested that important
interactions may exist among litter from different plant
species and that simple additive models are not sufficient
descriptions of species-level effects on ecosystem-scale
decomposition and nutrient cycling (Gartner and Cardon
2004; Gessner et al. 2010; Lecerf et al. 2011).
Some of the core principles underlying species-level
effects on decomposition and nutrient cycling have been
understood for over 30 years (Swift et al. 1979; Vitousek
1982; Wedin and Tilman 1990; Hobbie 1992). Fast-growing species with high rates of nutrient uptake have nutrientrich, high-quality litter that decomposes quickly and tends
to reinforce high rates of nutrient cycling within ecosystems, while species with lower rates of nutrient uptake
have more recalcitrant litter with higher carbon:nitrogen
(C:N) or lignin:N ratios and tend to maintain slower rates
of nutrient cycling (Hobbie 1992). Such differences are
potentially very important in invaded ecosystems because
litter from invasive species often (but not always) has
higher nutrient contents and decomposition rates than litter
from co-occurring natives, potentially leading to increased
rates of nutrient cycling (Ehrenfeld 2003; Ashton et al.
2005; Liao et al. 2008; but see Funk and Vitousek 2007).
Discerning the effects of invasive species per se can be
difficult, however, since there are often categorical differences in the physiology and/or growth form of the species
used in experimental studies, and these differences are
confounded with a species’ classification as native or
invasive. For example, many studies finding that invasive
species alter nutrient cycling rates include invasive
N2-fixing plants, which are expected to have litter with
higher N concentrations, lower C:N ratios, and faster rates
of decomposition than the non-fixing native plants to which
they are compared (Vitousek and Walker 1989; Witkowski
1991; Haubensak et al. 2004; Rice et al. 2004; Liao et al.
2008; Hickman et al. 2010). Invasions in which a novel
growth form is introduced into an ecosystem are also likely
to be associated with changes in litter decomposition rates
123
and nutrient cycling (Hobbie 1996), as has been observed
for a number of invasions (Vitousek and Walker 1989;
Wedin and Tilman 1990; Scowcroft 1997; Martin et al.
2003). Conducting comparisons using native and invasive
species that are paired by phylogeny and growth form
minimizes the likelihood of confounding differences in
origin with these other factors (Agrawal et al. 2005; Ashton
et al. 2005).
Many current models of invasive species effects on
nutrient cycling assume that the effects of mixing litter
from different native and/or invasive species is additive
(e.g., Levine et al. 2006). It is possible, however, that the
rate at which ecosystem dynamics change after plant
invasions may be complicated by interactions among litter
from multiple species, such that an invader’s effect on
ecosystem properties is not a simple function of its population density. Litter from two different species often
exhibits non-additive rates of mass and nutrient loss when
decomposing in mixtures as opposed to in isolation, and
these interactions may play an important role in determining how quickly ecosystem processes change during a
shift in plant community composition (Gartner and Cardon
2004; Lecerf et al. 2011).
Despite the potential for mixed-litter effects to alter the
ecosystem-scale impacts of biological invasions, remarkably little empirical work has examined this question
directly (e.g., Swan et al. 2008). We investigate the relationships among litter characteristics and ecosystem properties during invasion in a Mid-Atlantic forest in Long
Island, NY, USA. The forests of Long Island are extensively invaded, with invaders reaching over 50 % of cover
in some invaded areas (e.g., Howard et al. 2004; Ashton
et al. 2005), though individual species appear to be typically limited to 10–20 % of cover, and invasion by multiple species is common (Howard et al. 2004). Since
invasions in these forests commonly result in a mosaic of
native and invasive species, litter mixing between invasive
and native litters may have important consequences for
changes in ecosystem processes. Using species that deposit
substantial leaf litter in invaded areas of mesic Long Island
forests (Howard et al. 2004; Ashton et al. 2005), we conducted a series of experiments with phylogeneticallypaired native and invasive leaf litter to evaluate how well
species-specific chemical, physiological, and ecological
factors can be used to explain differences in decomposition
of native and invasive leaf litters and the potential impacts
on ecosystem processes. We expected that the invasive leaf
litter will have higher N concentrations and decompose
more quickly than native litter when decomposing singly.
We further expect that mixtures between more nutrient-rich
invasive litter and less nutrient-rich native litter will
decompose non-additively, though mechanisms for both
more and less rapid decomposition are conceivable.
Oecologia
Materials and methods
Site description
We examined decomposition of native and exotic invasive
leaf litter in the East Farm preserve along the north shore of
Long Island, Suffolk County, NY, USA (40°54.30 N,
73°08.90 W). The Preserve was cleared in the early 1900s,
used for crops and pasture, abandoned, and acquired by
The Nature Conservancy in 1970. The Preserve is a mixed
deciduous hardwood forest with loamy soils. We selected a
plot in a relatively uninvaded, native-dominated forest, with
\1 % exotic species cover (Ashton et al. 2005). An uninvaded site was chosen to provide insight into how an
uninvaded ecosystem might respond to plant invasions.
The most common native species in the plots were Acer
rubrum L. (red maple), Fagus grandifolia Ehrh. (American
beech), Quercus rubra L. (red oak), Quercus alba L. (white
oak), Prunus serotina Ehrh. (black cherry), and Viburnum
acerifolium (maple-leaf viburnum).
Experimental design
We examined the effects of mixing litter from different
species and origins on decomposition in two separate
experiments. The first, conducted in 2002–2003, was
designed to examine whether litter from invasive species
lost mass and N more quickly than litter from native species (‘‘native versus invasive experiment’’). The second,
conducted in 2003–2004, was designed to examine how
mixing of litter from native and invasive species affects
leaf litter decomposition (‘‘mixed litter experiment’’). In
autumn, 2002, we collected newly senesced leaf litter from
woody species common to Long Island for the native
versus invasive experiment, which ran from December
2002 to December 2003. In fall 2003, we collected leaf
litter for the mixed litter experiment, which ran from
December 2003 to December 2004. All the species selected
were common on Long Island and present in the East Farm
preserve. Half of the selected species were invasive species
not native to North America (hereafter referred to as
invasive), and half were native species. The invasive species all have broad distributions across the eastern United
States (USDA 2012), and are all present in the East Farm
preserve; species in this experiment were selected to be
among the most common native and invasive species at the
site (Ashton et al. 2005). Each native species selected was
paired with a corresponding invasive species by growth
form, which can influence rates of decomposition (Hobbie
1996), and by phylogentic relatedness. All combinations
represent native and invasive species within the same
genus, or, where within-genus pairing was not possible,
within the same family. Decomposition is strongly
influenced by the chemical and physical structure of leaf
litter, and litter from related species exhibits less variation
in these traits (Ehrlich and Birch 1967; Swift et al. 1979;
Becerra 1997). By selecting related taxa of native and
invasive species with the same growth form, we can help
ensure that any differences between species are due to
species origin rather than differences in phylogeny, growth
form, or functional group (e.g., N2 fixation); none of the
litter selected was from N2-fixing species. Two species
pairs: A. rubrum and A. platanoides L. (trees, genus Acer)
and Vitis novae-angliae and Ampelopsis brevipedunculata
(Maxim.) Trautvs. (vines, family Vitaceae) were used in
both experiments. The mixed litter experiment included
one additional pair [Rubus allegheniensis and R. phoenicolasius (shrubs, genus Rubus)], while the native versus
invasive experiment included three additional pairs
[R. occidentalis and R. pheonicolasius (shrubs, genus
Rubus), P. serotina and Rosa multiflora (shrubs, family
Rosaceae), and Lonicera morrowii (vine/shrub) and
V. acerifolium (shrub, family Caprifoliaceae)]. Parthenocissus quinquefolia (vine/shrub, family Vitaceae) was also
included in 2003, making the Vitaceae combination a
triplet. For all species, litter was collected from at least
three sites on Long Island, within an area of approximately
100 km2, and from multiple individuals within each site.
The litter selected was derived from plants and leaves
grown in full sun. Within each experiment, litter from all
sites was pooled for each species.
All litterbags were constructed from 1-mm2 fiberglass
mesh, with interior dimensions of 15.5 9 12 cm. Litter
collected in 2002 was dried at 38 °C, and a subsample of
dried litter from each species was weighed, dried further at
60 °C, and reweighed to calculate 38–60 °C conversion
factors. Litter collected in 2003 was dried at 60 °C. We
filled each bag with a total of 2.5–3.5 g of litter from a
single species, or with equal amounts of litter from two
different species. Filled bags were sealed, dried, and
weighed before placement in the field.
For the native versus invasive experiment in 2002–2003,
enough bags were filled with each species to allow for the
destructive harvest of three replicates at 0, 3, 15, 25, 35,
and 52 weeks for a total of 198 litter bags. An additional
set of bags was filled with 3 g of polyester material to
quantify the amount of organic and inorganic mass gained
by the bags over time. The mean mass gain by the polyester
bags collected at each harvest time was subtracted from the
litter bags to quantify the amount of organic and inorganic
matter accumulating in the bags over time (Harmon et al.
1999). For the mixed litter experiment in 2004, enough
bags were filled with every possible one- and two-species
combination to allow for the destructive harvest of three
replicates at 0, 3, 15, 25, 40, and 52 weeks, for a total of
378 litter bags.
123
Oecologia
Replicates for each harvest were randomly placed in
each of three blocks spaced 10 m apart from each other; the
site for the mixed litter experiment was located within
roughly 50 m of the 2003 experiment. Blocking was used
to account for any potential spatial variation in each site;
site differences between the 2003 and 2004 experiments
were not accounted for. Within each block, the bags were
randomly placed 0.5 m from each other in a grid, and
pinned to the soil surface, below the existing litter layer. At
each of the six harvest times in each year, bags were collected, gently cleaned of external soil and organic matter,
dried at 60 °C, and weighed to calculate mass loss. Species
were not separated. A Wiley mill was used to grind the
dried litter, which was then analyzed for C and N concentration using a CE Flash EA 1112 Elemental Analyzer
(CE Instruments, Milan, Italy). In the mixed litter experiment, harvested litter was placed overnight in a muffle
furnace at 550 °C for the determination of ash-free dry
weight (AFDW) to correct for any soil contamination of
the samples.
Statistical analysis
We conducted separate model comparisons among nested
models using maximum likelihood for each of the response
variables (mass loss and N loss). The full model included
one random factor (block) and three fixed factors [time,
species, and litter type; litter type had two levels in the
native versus invasive experiment (native and invasive)
and five levels in the mixed litter experiment (single native
species, single invasive species, mixture of two invasive
species, mixture of two native species, and mixture of one
native and one invasive species)]. The native versus invasive litter experiment also included phylogenetic group as a
factor in the full model. Reduced models excluded one or
more factors. This model comparison approach differs
from the more traditional evaluation of expected versus
observed values for mixes (as seen, e.g., in Gartner and
Cardon 2004), but is a more direct analysis for detecting
whether an ecosystem experiences perceptible changes in
mass or nutrient loss as a result of mixing. Because factors
such as site, experimental treatments, and sample processing differed in the two experiments, separate analyses
were conducted for each experiment.
We also conducted a series of four planned contrasts to
determine whether mixtures of litter from different species
decomposed more quickly or more slowly than species
decomposing alone: invasive mixtures versus invasive
single species, native mixtures versus native single species, native/invasive mixtures versus invasive single
species, and native/invasive mixtures versus native single
species. Although these four tests are not strictly orthogonal, they do represent a set of reasonable contrasts in line
123
with our a priori hypotheses regarding the effects of litter
mixtures, and the family-wide a for the tests was not
adjusted. However, P values for the unplanned comparisons are also presented.
The contrasts were conducted for the response variables
(proportion total mass remaining and proportion N
remaining, corrected using the AFDW) at week 52, since
replicates from this date exhibit the cumulative differences
in mass and N loss among the various litter types, and thus
provide the best metric for understanding how these differences may affect ecosystem processes over a longer time
scale. An additional analysis of the absolute amount of N
lost by week 52 was conducted, since higher starting N
concentrations in invasive or native litters could result in
larger N fluxes to ecosystems. A series of unplanned contrasts among the individual single and mixed litter types
were made to identify any differences between species.
Data were log- or rank-transformed as necessary to meet
the assumptions of ANOVA.
Likelihood estimates of the decomposition constants
(k) were determined for the relationship: x = e-kt, where
x proportion mass remaining and t time since litter bag
placement. The likelihood estimates were determined by
simulated annealing using the anneal function in the likelihood package developed for R by Lora Murphy
(http://www.ecostudies.org/lme_R_code_tutorials.html).
Results
Litter chemistry
Chemical characteristics, mass loss, and decomposition
constants (k) varied among species in both experiments
(Table 1). For the native versus invasive experiment
comparing decomposition of invasive and native litter in
five different phylogenetic pairs, the initial litter C:N ratio
was negatively related to species-specific decomposition
constants in a linear model when an outlier, V. acerifolium,
was excluded from the analysis (P = 0.045, adjusted
R2 = 0.34). Initial N concentrations and C:N varied significantly by phylogenetic grouping (P \ 0.0001 for both),
from 1.09 %N and a C:N of 41.2 (Aceraceae) to 1.73 %N
and a C:N of 27.1 (Roseaceae); V. acerifolium was again
excluded from the analysis in order to meet the homogeneity of variances assumption. The value of k in the native
versus invasive experiment also varied by phylogenetic
grouping (at the level of either genus or family) at
P = 0.10 (two-tailed test).
The initial C:N ratio of leaf litter was positively related
to the mass remaining after 1 year across the full range of
growth forms and families in both experiments
(P \ 0.005), though initial C:N explained more of the
Oecologia
Table 1 Mean mass loss, N loss, decomposition constants, and initial litter chemistry of individual species used in the two experiments
Species
Mass remaining (%)
N loss (mg)
Initial N (%)
Initial C:N
k
Native 2003 (native versus invasive)
Acer rubrum L.
50.3 (4.5)
3.41 (3.39)
1.14 (0.03)
40.85
0.98 (0.04)
Parthenocissus quinquefolia
34.8 (0.8)
34.2 (1.6)
1.66 (0.04)
27.23
1.19 (0.15)
Prunus serotina
29.2 (2.0)
24.2 (1.3)
1.35 (0.07)
33.37
1.38 (0.03)
Rubus occidentalis
20.4 (5.5)
34.9 (5.7)
1.60 (0.03)
28.13
1.82 (0.13)
Viburnum acerifolium
26.9 (2.1)
6.90 (1.52)
0.86 (0.05)
54.37
1.35 (0.10)
Vitis novae-angliae Fern.
24.8 (3.0)
25.8 (0.6)
1.30 (0.02)
34.43
1.25 (0.08)
Acer rubrum L.
45.6 (14.5)
-5.83 (4.91)
0.61 (0.03)
80.5
0.91 (0.19)
Rubus allegheniensis
54.8 (8.2)
0.967 (5.30)
1.55 (0.05)
31.11
0.82 (0.16)
Vitis novae-angliae Fern.
49.0 (3.8)
Invasive 2003 (native versus invasive)
-1.87 (2.71)
1.46 (0.08)
32.60
1.00 (0.13)
Native 2004 (mixed litter)
Acer platanoides L.
50.7 (1.5)
1.19 (1.41)
1.04 (0.07)
42.68
0.77 (0.07)
Ampelopsis brevipedunculata
36.9 (0.7)
17.0 (2.0)
1.47 (0.07)
29.58
0.94 (0.05)
Lonicera morrowii
17.5 (7.0)
33.7 (7.0)
1.44 (0.05)
30.93
2.12 (0.11)
Rosa multiflora
6.1 (2.7)
72.0 (4.2)
2.11 (0.04)
20.77
1.83 (0.08)
Rubus phoenicolasius
5.8 (3.0)
45.3 (4.7)
1.39 (0.03)
30.72
2.10 (0.03)
Acer platanoides L.
47.0 (1.8)
-4.73 (1.95)
1.03 (0.14)
44.87
0.83 (0.05)
Ampelopsis brevipedunculata
50.2 (0.9)
13.9 (0.4)
2.09 (0.22)
23.15
0.68 (0.04)
Rubus phoenicolasius
14.3 (3.5)
43.4 (4.2)
2.06 (0.01)
22.40
1.76 (0.16)
Invasive 2004 (mixed litter)
Standard errors are given in parentheses
variance in mass remaining for unmixed litter in the native
versus invasive experiment (R2 = 0.22) than for all mixed
litter (R2 = 0.15). Initial C:N was not related to mass lost
exclusively in mixtures of native and invasive species
(P = 0.48, R2 = 0.02). In the native versus invasive
experiment, a model that included phylogeny as a factor in
addition to C:N provided a better fit than a reduced model
excluding phylogeny (P \ 0.001).
Overall, litter from invasive species had higher N contents than native species (1.49 ± 0.095 SE vs.
1.41 ± 0.054 %N, P = 0.072 for the species used in the
native versus invasive experiment; and 1.72 ± 0.19 vs.
1.21 ± 0.15 %N, P = 0.025 for the species used in the
mixed litter experiment, one-tailed tests), as well as a lower
C:N ratio (30.9 ± 1.96 vs. 32.8 ± 1.36, P = 0.034, for the
species used in 2003; and 33.3 ± 2.51 vs. 40.85 ± 2.55,
P = 0.0032, for the species used in 2004). A significant
interaction between origin and phylogeny in the native
versus invasive experiment (P \ 0.0001 for both %N and
C:N) indicated that the direction of these differences
between native and invasive species varied depending on
phylogenic group. Planned contrasts within phylogenetic
groups found significant differences in %N between native
and invasive species within the same phylogenetic group
for the Aceraceae at P \ 0.10, and all other phylogenetic
groups at P \ 0.05, with invasive species having higher
starting N concentrations in the Caprifoliaceae, Rosaceae
(PRSE-ROMU), and Vitaceae pairs. Differences in C:N at
P \ 0.05 were found only for the Vitaceae and for the
Rosaceae PRSE-ROMU pairing with C:N ranges narrower
in the invasive than in native member of the pair, and at
P = 0.057 for the Caprifoliaceae pairing, in which the C:N
ratio was narrower in the native litter.
Decomposition of native and invasive litters
Both mass loss and N loss over time among individual
species in the native versus invasive experiment were best
described by models that included a three-way interaction
between phylogenetic group, time, and invasion status
(P \ 0.001 for mass loss and P \ 0.01 for N loss compared to models without this interaction), suggesting that
differences in mass and N loss between invasive and native
species varied depending on the phylogenetic group of
those species. After 52 weeks, differences in mass and N
loss between native and invasive species appeared to
emerge, though the difference was not significant
(P = 0.08 and P = 0.085 in one-tailed tests for mass and
N, respectively; Fig. 1). Unplanned comparisons found
significant differences in total mass loss at the end of
1 year between native and invasive litters within the
Vitaceae (P = 0.036), Rubus (P = 0.014), Caprifoliaceae
123
Oecologia
(P = 0.02), and Roseaceae (P = 0.002), and changes in N
mass for Vitaceae (P = 0.01), Rubus (P = 0.008), and
Roseaceae (P = 0.004) (Table 1; Fig. 2). Mass loss was
greater for the invasive species in the Rubus, Caprifoliaceae, and Rosaceae pairings, but the opposite was true for
the Vitaceae pairing. The native Vitaceae species also lost
more N over 52 weeks than the exotic Vitaceae, while the
invasive Rubus and Roseaceae species lost more N than
their native counterparts. There were no differences
between species in the genus Acer, which lost mass more
slowly than other families in post hoc comparisons
(P \ 0.01 for all comparisons; Fig. 2a, b).
Both mass loss and N loss over time in the litter mixture
experiment were best described by a model that included
litter type as a factor, where litter type is a factor with five
levels: three mixture types (invasive/invasive, native/
native, and native/invasive) and the two single species
types (native and invasive; Fig. 3). In an unplanned
comparison, invasive litter had a smaller proportion of
its starting mass remaining by the end of the year than
native litter when decomposing singly (37.2 ± 5.8 vs.
49.8 ± 5.1 %, P = 0.12]. While the mass lost after 1 year
by the native/native and invasive/invasive mixtures was not
different from that lost by unmixed native and invasive
species, respectively, mass remaining in the native/invasive
mixture (47.4 ± 2.2 %) was significantly more than the
unmixed invasive species [P = 0.05 (planned comparisons); P = 0.25 (unplanned comparisons)], but was not
different from the mass remaining for the unmixed native
species.
The same pattern of differences was present for N loss
from the different litter types: there was more N lost after
52 weeks from invasive than from native litters in an
unplanned comparison (18 ± 7.0 vs. -2.0 ± 2.0 mg,
P = 0.02), there were no differences between mixed and
unmixed invasive or native litters, and there was more N
lost from unmixed invasive litter than from the native/
invasive mixture [4.0 ± 2.0 mg, P = 0.015 (planned),
P = 0.033 (unplanned); Fig. 3]. Differences in mass loss
within phylogenetic pairs between invasive and native
unmixed litters were only found for the pair of Rubus
species, where the invasive species lost more mass than the
native species (P = 0.002; Fig. 4). Unplanned comparisons also showed that after 52 weeks, the mass lost for R.
phoenicolasius was greater than the mass lost from any
other unmixed litter (P \ 0.03 for all comparisons). More
N was lost by the invasive than the native species in the
Rubus and Vitaceae unmixed litter by week 52
(P \ 0.0001 and P = 0.01, respectively).
Discussion
Fig. 1 Decomposition (percentage of initial mass remaining) and
nitrogen loss (percentage of initial nitrogen remaining) of leaf litter
from native and invasive species. Open symbols native species and
filled symbols invasive. Error bars ±1SE. Planned comparisons at
week 52 for differences between native and invasive litters were
significant
123
Implicit in current models of species effects on ecosystems
is the assumption that the cumulative nutrient cycling
effects of multiple species with differences in primary
metabolism is additive—that the identity of species within
a community is secondary to the integrated canopy chemistry of the community as a whole (e.g., Levine et al. 2006).
Our results confirm the usefulness of initial C:N in litter for
the prediction of relative decomposition rates of single
species of leaf litter, and to a lesser extent for mixtures that
combined litter from two native species or two invasive
species. This relationship between C:N and decomposition
suggests that, in both monocultures and in native forest
communities of the eastern United States, models for
species impacts on ecosystems that are mediated by litter
deposition and decomposition may be built upon primary
metabolic characteristics of individual species. Our results
also confirm earlier experimental and meta-analytical evidence that leaf litter from invasive species tends to have
Oecologia
Fig. 2 Decomposition
(percentage of initial mass
remaining) and nitrogen loss
(percentage of initial nitrogen
content remaining) of litter from
native and invasive species by
genus or family. Open symbols
native species, and filled
symbols invasive species. Error
bars ±1SE
lower C:N ratios and higher rates of decomposition than
litter from native species (Ehrenfeld 2003; Liao et al.
2008).
This simple model relating physiology to decomposition
breaks down during biological invasions, when litter from
native and invasive species is mixed. Under this model, the
lower C:N ratio of the invasive species tends to reduce the
overall C:N ratio of mixtures between native and invasive
litters, and would be expected to accelerate decomposition
relative to the native species decomposing singly. However, we found that decomposition behaves non-additively
in these mixtures, with no changes in annual mass and N
loss compared to native litter, but significantly less annual
mass and N loss than is found for invasive species. While
initial C:N is significantly related to final mass for both
mixed and unmixed species, it explains less of the variance
in mass loss for the species mixtures (adjusted R2 = 0.15)
than for individual species (adjusted R2 = 0.22), and
explains none of the variance in mass loss among the nine
mixtures of native and invasive species (adjusted
R2 = 0.02, P = 0.48). These results imply that initial leaf
litter C:N concentrations are less effective in predicting
mass loss when litter from multiple species is mixed. This
complete loss in predictive power for mixtures of native
litter (with higher C:N and lower %N) and invasive litter
(with lower C:N and higher %N) may be an example of
wider patterns in which C:N differences in mixed litter lead
to non-additive decomposition dynamics (e.g., Lecerf et al.
2011).
It should be noted that these relationships could change
if litter lignin:N concentrations were evaluated, which can
be a much more powerful predictor of mass loss in leaf
litter than C:N (Melillo et al. 1982). For example, N concentrations and C:N in A. rubrum varied considerably in
the 2 years, but this variation had little effect on the estimated value of k. Regardless, the non-additive effect—at
least from the perspective of the ecosystem—of mixing
high C:N native and low C:N invasive litter indicates that
primary metabolic characteristics integrated across a
community may not be sufficient for predicting invasion
impacts in a community context. Instead, it may be necessary to consider functional differences among species
123
Oecologia
the rapid decomposition of R. phoenicolasisus seen in the
unplanned comparisons suggests that this species may have
had an important influence on the differences between
native and invasive species decomposing alone, but these
comparisons do not provide sufficient evidence to conclude
that interactions slowing R. phoenicolasisus decomposition
were driving non-additive mass loss in the native/invasive
litter mixture treatment. More generally, this experiment
does not evaluate decomposition dynamics after the first
year of decomposition. While the rate of mass loss can be
constant during the first phase of decomposition over the
first 3 years of decomposition in northeastern hardwood
forests (e.g., Aber et al. 1990), and a substantial amount of
mass was lost from most litter types during the experiment,
it is certainly conceivable that the patterns of mass and
N loss after 12 months could be different after 24 or
36 months.
Implications of mixing native and invasive litters
Fig. 3 Decomposition (percentage of initial mass remaining) and
nitrogen loss (percentage of initial nitrogen remaining) of leaf litter
from native and invasive species and litter mixtures. Open symbols
native species and filled symbols invasive species. Error bars ±1SE
such as C:N or lignin:N and their interactions when predicting how biogeochemical processes will change as plant
invasions proceed.
Interactions in decomposition of mixed litter are not
rare in published studies, and may represent the majority of
cases (Gartner and Cardon 2004; Gessner et al. 2010;
Lecerf et al. 2011). Multiple potential mechanisms for
positive or negative effects of mixing on decomposition
have been posited, and some may play a role in the
interactions observed here. Phenolic compounds or secondary metabolites transferred from one litter type to
another can slow decomposition in a mixture (Fyles and
Fyles 1993; McArthur et al. 1994; Salamanca et al. 1998;
Nilsson et al. 1999). A. rubrum generally has higher phenolic concentrations than faster-growing native trees, and
also has higher concentrations of hydrolysable tannins
(Shure and Wilson 1993). Relatively high concentrations
of polyphenolics have also been found in V. riparia (of
which V. novae-angliae is a hybrid) in comparison to other
species in the genus Vitis (Kortekamp 2006). Empirically,
123
Invasive species in the eastern United States can frequently
exceed 50 % of plant cover (e.g., Howard et al. 2004;
Ashton et al. 2005; Kourtev et al. 1998), becoming the
dominant component of a community. Early in an invasion,
invaders are likely to be less common than native members
of a community, so low C:N invasive litter will usually be
decomposing in association with higher C:N native litters.
As long as this physical association between high and low
C:N litter continues, the faster decomposition rates typically exhibited by the lower C:N invasive litter are retarded, moderating the impacts of fast-decomposing invaders
on soils. As the abundance of invaders in a community
increases—as it has in Long Island for the invasive species
chosen for this study—the chances that litter from a given
invasive species is decomposing alone or in combination
with another invader will increase, ultimately removing the
check that native litter provides on rapid decomposition
and leading to increases in nutrient cycling rates. This
moderating effect of native litter may buffer uninvaded
ecosystems against changes caused by invasive litter, and
may be an important cause of the frequently observed
‘‘invasion lag,’’ in which changes to nutrient cycling lag
substantially behind the establishment of an invader. In
addition, if litter-mediated changes in nutrient cycling and
soil communities are an important factor contributing to
invasive success, the moderating effect of native litter
may also indirectly buffer the native community against
invasion by dampening any positive feedback between
nutrient cycling rates and the introduced species’ competitive success. The strength of this buffering effect can be
expected to be inversely related to invader abundance and
strongest during the early stages of an invasion. Additional
research into understanding the mechanism behind these
Oecologia
Fig. 4 Decomposition (percentage of initial mass remaining) and
nitrogen loss (percentage of initial nitrogen content remaining) of
litter from single native and invasive species and mixtures by genus or
family. Open circles native species alone, filled circles invasive
species alone, and open triangles the mixture of the two. Error bars
±1SE
interactions and the role that species, genus, or familyspecific biology plays will be needed to better understand
the degree to which a buffering effect is likely to occur
with any given invasion.
These results should be tempered by the knowledge that
invasive plants can influence nutrient cycling via other
mechanisms (e.g., root exudation, root turnover, and
throughfall), though leaf litter represents the largest input
of nutrients to soils in most ecosystems. In addition, we
examined only one mixing ratio of litter (1:1), but the nonadditive effects of mixing may not be linear across a range
of ratios (e.g., Scowcroft 1997). Results would likely differ
in an established invaded site, where decomposition and
nutrient cycling can proceed more rapidly (Ashton et al.
2005; Meisner et al. 2012), and it is possible that, over a
longer time period, the patterns in decomposition observed
here could become more or less pronounced.
Management implications
The major finding of this study—that the generally more
rapid decomposition of lower C:N invasive litter tends to
slow to rates typical of slower decomposing, higher C:N
native species when native and invasive litters are mixed
together—has important implications for the development
of the theory and prediction of the impacts of plant invasions. This result suggests that there may be little or no
impact of invasive species on nutrient cycling during the
early stages of an invasion, when native leaf litter is
abundant. It is possible that efforts at limiting invader
impacts on ecosystems may best be achieved by working to
limit invader abundance in communities where native litter
deposition equals or exceeds invasive litter, even if
invaders are already well established in an ecosystem. In
cases where eradication of invaders is not feasible,
123
Oecologia
activities aimed at limiting invader abundance could prevent the kinds of rapid ecosystem changes that invasions
can cause (e.g., Ashton et al. 2005) and keep ecosystems
from crossing biogeochemical thresholds resulting in
changes to ecosystem processes that are likely to be more
costly and difficult to reverse.
Acknowledgments We would like to thank The Nature Conservancy for their cooperation, Wei Wang and Pengfei Zang for statistical advice, and Zoe Cardon, Howard Epstein, and Deborah
Lawrence for providing comments on earlier drafts of the manuscript.
References
Aber JD, Melillo JM, McClaugherty CA (1990) Predicting long-term
patterns of mass loss, nitrogen dynamics, and soil organic matter
formation from the initial fine litter chemistry in temperate forest
ecosystems. Can J Bot 68:2201–2208
Agrawal AA, Kotanen PM, Mitchell CE, Power AG, Godsoe W,
Klironomos J (2005) Enemy release? An experiment with
congeneric plant pairs and diverse above- and belowground
enemies. Ecology 86:2979–2989
Ashton IW, Hyatt LA, Howe KM, Gurevitch J, Lerdau MT (2005)
Invasive species accelerate decomposition and litter nitrogen
loss in a mixed deciduous forest. Ecol Appl 15:1263–1272
Becerra JX (1997) Insects on plants: macroevolutionary chemical
trends in host use. Science 276:253–256
Chapin FS (1980) The mineral nutrition of plants. Annu Rev Ecol
Syst 11:233–260
Ehrenfeld JG (2003) Effects of exotic plant invasions on soil nutrient
cycling processes. Ecosystems 6:503–523
Ehrlich PR, Birch LC (1967) Balance of nature and population
control. Am Nat 101:97–107
Funk JL, Vitousek PM (2007) Resource-use efficiency and plant
invasion in low-resource systems. Nature 446:1079–1081
Fyles JW, Fyles IH (1993) Interaction of douglas fir with red alder
and salal foliage litter during decomposition. Can J For Res,
23:358–361
Gartner TB, Cardon ZG (2004) Decomposition dynamics in mixedspecies leaf litter. Oikos 104:230–246
Gessner MO, Swan CM, Dang CK, McKie BG, Bardgett RD, Wall
DH, Hattenschwiler S (2010) Diversity meets decomposition.
Trends Ecol Evol 25:372–380
Harmon ME, Nadelhoffer K, Blair JM (1999) Measuring decomposition, nutrient turnover, and stores in plant litter. In: Robertson
GP, Coleman DCC, Bledsoe CS, Sollins P (eds) Standard soil
methods for long-term ecological research. Oxford University
Press, New York, pp 220–240
Haubensak KA, D’Antonio CM, Alexander J (2004) Effects of
nitrogen-fixing shrubs in Washington and coastal California.
Weed Technol 18:1475–1479
Hickman JE, Wu SL, Mickley LJ, Lerdau MT (2010) Kudzu (Pueraria
montana) invasion doubles emissions of nitric oxide and increases
ozone pollution. Proc Natl Acad Sci USA 107:10115–10119
Hobbie SE (1992) Effects of plant-species on nutrient cycling. Trends
Ecol Evol 7:336–339
Hobbie SE (1996) Temperature and plant species control over litter
decomposition in Alaskan tundra. Ecol Monogr 66:503–522
Howard TG, Gurevitch J, Hyatt LM, Carreiro M, Lerdau MT (2004)
Forest invasibility in communities in southeastern New York.
Biol Invasions 6:393–410
123
Kortekamp A (2006) Expression analysis of defence-related genes in
grapevine leaves after inoculation with a host and a non-host
pathogen. Plant Physiol Biochem 44:58–67
Kourtev PS, Ehrenfeld JG, Huang WZ (1998) Effectsof exotic plant
species on soil properties in hardwoodforests of New Jersey.
Water Air Soil Pollut 105:493–501
Lecerf A, Marie G, Kominoski JS, LeRoy CJ, Bernadet C, Swan CM
(2011) Incubation time, functional litter diversity, and habitat
characteristics predict litter-mixing effects on decomposition.
Ecology 92:160–169
Levine JM, Pachepsky E, Kendall BE, Yelenik SG, HilleRisLambers
J (2006) Plant–soil feedbacks and invasive spread. Ecol Lett
9:1005–1014
Liao CZ, Peng RH, Luo YQ, Zhou XH, Wu XW, Fang CM, Chen JK,
Li B (2008) Altered ecosystem carbon and nitrogen cycles by
plant invasion: a meta-analysis. New Phytol 177:706–714
Martin RE, Asner GP, Ansley RJ, Mosier AR (2003) Effects of
woody vegetation encroachment on soil nitrogen oxide emissions in a temperate savanna. Ecol Appl 13:897–910
McArthur JV, Aho JM, Rader RB, Mills GL (1994) Interspecific leaf
interactions during decomposition in aquatic and floodplain
ecosystems. J North Am Benthol Soc 13:57–67
Meisner A, de Boer W, Cornelissen JHC, van der Putten WH (2012)
Reciprocal effects of litter from exotic and congeneric native
plant species via soil nutrients. PLoS One 7:e31596
Melillo JM, Aber JD, Muratore JF (1982) Nitrogen and lignin control
of hardwood leaf litter decomposition dynamics. Ecology
63:621–626
Nilsson MC, Wardle DA, Dahlberg A (1999) Effects of plant litter
species composition and diversity on the boreal forest plant–soil
system. Oikos 86:16–26
Rice SK, Westerman B, Federici R (2004) Impacts of the exotic,
nitrogen-fixing black locust (Robinia pseudoacacia) on nitrogencycling in a pine-oak ecosystem. Plant Ecol 174:97–107
Salamanca EF, Kaneko N, Katagiri S (1998) Effects of leaf litter
mixtures on the decomposition of Quercus serrata and Pinus
densiflora using field and laboratory microcosm methods. Ecol
Eng 10:53–73
Scowcroft PG (1997) Mass and nutrient dynamics of decaying litter
from Passiflora mollissima and selected native species in a
Hawaiian montane rain forest. J Trop Ecol 13:407–426
Shure DJ, Wilson LA (1993) Patch size effects on plant phenolics in
successional openings of the southern Appalachians. Ecology
74:55–67
Swan CM, Healy B, Richardson DC (2008) The role of native riparian
tree species in decomposition of invasive tree of heaven
(Ailanthus altissima) leaf litter in an urban stream. Ecoscience
15:27–35
Swift MJ, Heal OW, Anderson JM (1979) Decomposition in
terrestrial ecosystems. Blackwell Scientific, Oxford
USDA, NRCS (2012) The plants database. National plant data
team, Greensboro, NC 27401-4901 USA. http://plants.usda.gov.
Accessed 29 August 2012
Vitousek P (1982) Nutrient cycling and nutrient use efficiency. Am
Nat 119:553–572
Vitousek PM, Walker LR (1989) Biological invasion by Myrica-Faya
in Hawaii—plant demography, nitrogen-fixation, ecosystem
effects. Ecol Monogr 59:247–265
Wedin DA, Tilman D (1990) Species effects on nitrogen cycling—a
test with perennial grasses. Oecologia 84:433–441
Witkowski ETF (1991) Effects of invasive alien acacias on nutrient
cycling in the coastal lowlands of the Cape Fynbos. J Appl Ecol
28:1–15