Ecology, 87(3), 2006, pp. 686–694
q 2006 by the Ecological Society of America
GASTROPOD HERBIVORY IN RESPONSE TO ELEVATED CO2 AND
N ADDITION IMPACTS PLANT COMMUNITY COMPOSITION
ELSA E. CLELAND,1,3 HALTON A. PETERS,1 HAROLD A. MOONEY,1
AND
CHRISTOPHER B. FIELD2
1Department of Biological Sciences, Stanford University, Stanford, California 94305 USA
Department of Global Ecology, Carnegie Institution of Washington, Stanford, California 94305 USA
2
Abstract. In this study, the influence of elevated carbon dioxide (CO 2) and nitrogen
(N) deposition on gastropod herbivory was investigated for six annual species in a California
annual grassland community. These experimentally simulated global changes increased
availability of important resources for plant growth, leading to the hypothesis that species
with the most positive growth and foliar nutrient responses would experience the greatest
increase in herbivory. Counter to the expectations, shifts in tissue N and growth rates caused
by N deposition did not predict shifts in herbivore consumption rates. N deposition increased
seedling N concentrations and growth rates but did not increase herbivore consumption
overall, or for any individual species. Elevated CO2 did not influence growth rates nor have
a statistically significant influence on seedling N concentrations. Elevated CO 2 at ambient
N levels caused a decline in the number of seedlings consumed, but the interaction between
CO2 and N addition differed among species. The results of this study indicate that shifting
patterns of herbivory will likely influence species composition as environmental conditions
change in the future; however, a simple trade-off between shifting growth rates and palatability is not evident.
Key words:
trade-off.
California grassland; elevated CO2; gastropods; herbivory; nitrogen deposition;
INTRODUCTION
Anthropogenic global changes over the coming century are expected to cause shifts in the composition of
terrestrial plant communities, as well as the functioning
of these ecosystems (Prentice et al. 2001). In order to
predict shifts in species composition, ecologists must
understand how the forces structuring plant communities are affected by global changes. Generalist herbivores have long been recognized to be important in
structuring terrestrial plant communities by forcing a
trade-off in plant allocation to growth vs. defense
(Jones 1977). Along resource gradients, plants growing
in high-resource habitats tend to have high foliar nutrient concentrations and high growth rates, and allocate little to defend their tissues, while plants in lowresource habitats grow more slowly and must allocate
resources to herbivore defense (Coley et al. 1985).
Within a community over the course of succession,
generalist herbivores tend to preferentially consume the
more palatable, fast-growing, early-successional species (Cates and Orians 1975, Mattson 1980), allowing
slow-growing, late-successional species to become
more abundant over time (Davidson 1993, Kielland and
Bryant 1998).
Manuscript received 5 April 2005; revised 16 August 2005;
accepted 17 August 2005; final version received 20 September
2005. Corresponding Editor: P. M. Groffman.
3 Present address: National Center for Ecological Analysis
and Synthesis, 735 State St., Suite 300, Santa Barbara, California 93103 USA. E-mail: [email protected]
Expected global environmental changes will alter the
availability of limiting plant resources such as carbon
and nitrogen (N), and have the potential to alter plant–
herbivore interactions via shifts in palatability among
plant species (Lincoln et al. 1993, Throop and Lerdau
2004). Experimentally elevated CO2 often lowers N
concentrations in plants, and increases nonstructural
carbohydrates and phenolics (Poorter et al. 1997, Bezemer and Jones 1998, Cortrufo et al. 1998, Wand et
al. 1999). These shifts in tissue quality tend to make
plants less palatable and can lower herbivore growth
rates (reviewed in Lincoln et al. 1993). Generalist herbivores may respond to lower plant tissue quality by
increasing the quantity of material they consume (i.e.,
compensatory consumption), or by shifting the relative
amounts of different species consumed (Lincoln et al.
1986, Lambers 1993, Bezemer and Jones 1998). Conversely, N deposition tends to increase the N content
of plants (Throop and Lerdau 2004), which may increase their palatability to herbivores, and perhaps alter
relative consumption rates among plant species.
Annual grasslands offer a model system for studying
how expected global environmental changes can alter
the forces that structure plant community composition,
because processes such as germination, establishment,
growth, competition, herbivory, and reproduction all
occur within a single growing season. In European
grasslands, gastropods have been shown to play a major
role in influencing species composition (Hanley et al.
1995b, Wilby and Brown 2001), directly by prefer-
686
HERBIVORY AND GLOBAL CHANGE
March 2006
entially removing more palatable species (Grime et al.
1968, Dirzo 1980), and indirectly by influencing competitive outcomes among plant species (Dirzo and
Harper 1980, Cottam 1986). In California annual grasslands, species composition is determined in the first
few weeks of the growing season (Bartolome 1979,
Young and Evans 1989). Gastropods are known to prefer seedlings to adult tissues (Hulme 1994, Hanley et
al. 1995a, Fenner et al. 1999), and seedlings grazed
early in the growing season are more likely to be killed
by herbivory (Fenner 1987); thus herbivory early in
the growing season may play a disproportionate role
in determining grassland species composition.
In this study, the influences of elevated carbon dioxide (CO2) and nitrogen (N) deposition on gastropod
herbivory were investigated for six annual species in
a California grassland. Because these global changes
increased the availability of important resources for
plant growth, we expected species with the most positive growth and foliar nutrient responses to experience
the greatest increase in herbivory.
METHODS
The Jasper Ridge Global Change Experiment
(JRGCE) is located on an area of sandstone-derived
soils within Stanford University’s Jasper Ridge Biological Preserve. The plant community is dominated
by naturalized European annual grasses, but also contains forbs, legumes, and perennial grasses (McNaughton 1968). The dominant gastropod herbivores are the
slugs Deroceras reticulatum and Arion intermedius; recent work has established that they are important consumers in this annual grassland community (Peters
2004). The JRGCE experimental design has been previously described elsewhere (Shaw et al. 2002, Zavaleta et al. 2003). Briefly, simulated environmental
changes were applied to undisturbed grassland plots in
a randomized block split-plot design, beginning in November 1998. The experiment described in this article
took place in the 2001–2002 growing season, which
was the fourth treatment year. Elevated CO2 was applied to whole plots (N 5 16), while N deposition was
applied to split-plots; the levels of each treatment were
chosen to simulate expected anthropogenic global
changes over the next century (Cusbasch et al. 2001).
Ambient CO2 concentrations averaged 380 ppm (parts
of CO2 per 106 parts of air); elevated CO2 concentrations of 680 ppm were achieved via free air CO2 enrichment (FACE). Nitrogen deposition was simulated
by surface applications totaling 7g N·m22·yr21 in the
form of calcium nitrate. In California most N falls as
NOx in dry deposition resulting in a pulse of N entering
the soil after the first rains following the summer dry
season; currently 0.5 g N·m22·yr21 is deposited in the
Jasper Ridge Biological Reserve (Weiss 1999). On 19
November 2 g N/m2 was added in a wet pulse, and the
remaining 5 g N/m2 was administered on 16 February
via slow-release pellets (Nutricote, Arysta LifeScience,
687
San Francisco, California, USA). The experimental
treatments were applied in a factorial design with eight
replicates of each of the four possible treatment combinations for a total of 32 plots, each with an area of
0.78 m2. Additional warming and precipitation treatments present in the JRGCE experimental design were
not utilized.
On 2 November 2001, we established two 6 3 16
cm grids in each experimental plot, which were later
planted with the focal species for this experiment. Inside each grid we assured that we would be able to
identify the focal seedlings by removing the seed bank.
To do this we cleared the litter, removed seeds near the
surface with a vacuum, disturbed the soil to 1 cm, and
encouraged germination of the remaining seed bank
with a 10-mm pulse of deionized water. On 10 November we removed the seedlings that had germinated from
the seed bank, and planted seeds of the focal species
for this experiment. Following germination, one grid
was left open to herbivory, while the second was enclosed to exclude herbivores. The 2001 growing season
began with the first substantial rains shortly after removal of the seed bank, with most germination occurring in mid-November (N. Chiariello, unpublished
data).
Six focal species were chosen to represent the variety
of growth forms present in the local species pool; half
are native and half are naturalized, and all are annual
species. Taxonomy follows Hickman (1996). Avena
barbata Link (slender wildoat) and Bromus hordeaceus
L. are the co-dominant species in the surrounding
grassland community; both are naturalized European
grasses (McNaughton 1968). Epilobium brachycarpum
C. Presl and Hemizonia congesta D.C. (hayfield tarweed) are the two most locally abundant native forbs.
Vicia sativa L. (common vetch) is the most common
legume, a naturalized European species. Finally, Lotus
purshianus (Benth.) Clements & E.G. Clements var.
purshianus is the most abundant native legume. We
planted seeds of the focal species in known locations
within the two grids using a randomized block design
in order to reduce the influence of neighbor effects
(Fenner 1987, Bergelson 1990). We planted six seeds
each of A. barbata, B. hordeaceus, E. brachycarpum,
and L. purshianus in each grid. Pilot germination trials
revealed that H. congesta and V. sativa had lower germination rates than the other species; hence we planted
12 seeds of each of these species. L. purshianus also
had low germination rates, but seed availability limited
planting to six seeds per grid. Each grid was protected
from herbivory for the first two weeks with an exclosure to ensure germination, consisting of a mesh-covered copper rectangle 2 cm high and pressed 1 cm into
the ground.
On 25 November we removed the germination exclosures, surveyed each grid, and created spatial maps
of germinated individuals. One grid in each quadrant
was subsequently enclosed with a copper barrier at the
688
ELSA E. CLELAND ET AL.
soil surface and fabric mesh reaching up 20 cm. The
other grid was left open to herbivory. In grids with
poor germination, transplants were introduced to produce even numbers within and among plots and treatments; these seedlings were germinated in field soils
and under field conditions just outside of the experiment. We resurveyed each grid on six subsequent dates:
27 November, 13 December, and 30 December 2001;
3 January, 22 January, and 12 February 2002. On each
date, we recorded whether the individual had survived,
or had suffered partial or entire herbivory. Gastropod
herbivory was evident due to abundant slime trails, and
the complete disappearance of seedlings during nighttime hours.
On four of the survey dates we harvested one individual of each species within each grid, except when
all individuals had suffered mortality previous to the
harvest. (On 3 January and 12 February no seedlings
were harvested.) Seedlings were harvested by block,
and were clipped at the soil surface. Harvested seedlings were dried at 708C for 24 hours, weighed, and
analyzed for carbon and N content on an elemental
analyzer (ECS 4010, Costech Analytical Technologies,
Valencia, California, USA). An exponential growth rate
was assumed, and the intrinsic rate of growth was estimated by regression of seedling mass (log transformed) against date, assuming a zero value for the
intercept. The slope of the regression line was used as
a metric to compare intrinsic growth rates among species and experimental treatments. This differs from a
calculation of relative growth rate (RGR), which measures the amount of growth per unit of initial biomass.
Four different seedlings were harvested at the different
time points; thus the initial seedling harvested was not
always the smallest individual. This metric makes the
best estimate of growth rate without allowing the first
seedling harvested to bias the measurement.
Statistical analyses were performed in SAS version
9.1 (SAS Institute 2004). Tissue N and growth rate were
analyzed using PROC MIXED, a general linear mixed
model procedure. The percentage of tissue N was log
transformed to improve normality. Proportions of germinated seeds were also nonnormal; normality did not
improve with transformation, and these values were
instead analyzed using the PROC GLIMMIX. This generalized linear mixed model procedure assumed a binomial error distribution appropriate for proportion
data, and a logit link function (Littell et al. 1996). The
statistical model accounted for the split-split-plot design: CO2 was the whole plot factor, N deposition the
split-plot factor, and species the split-split-plot factor
(Littell et al. 1996). Counts of seedling consumption
were severely nonnormally distributed due to zero consumption of some species under elevated CO2. Seedling
consumption was analyzed using the CATMOD procedure, an approach based on contingency table analysis using maximum likelihood estimation methods to
Ecology, Vol. 87, No. 3
evaluate the observed vs. expected frequency of individuals consumed (Stokes et al. 2001).
Additional discussion of the statistical methods,
model syntax, degrees of freedom, F tests, P values,
and parameter estimates are displayed on line in a supplementary statistical appendix (Appendix). The
threshold for statistical significance for all analyses was
chosen to be P , 0.05, and 0.05 , P , 0.10 was
considered marginally significant.
RESULTS
Germination
Of the 3072 seeds planted in known locations, 1690
individuals germinated. Most germination occurred before the first census, but 41 individuals germinated between the first and second census, and one individual
germinated between the second and third census. Most
of the late-germinating seedlings were Vicia sativa,
which had the largest seeds among the focal species.
While germination rates varied among species ( P ,
0.0001), they were not influenced by the global change
treatments (Fig. 1).
Herbivore consumption
In addition to the 1690 seedlings that germinated,
250 seedlings were transplanted to obtain more even
numbers of seedlings among grids, plots, and treatments. As seen in Table 1, the herbivore exclosures
only reduced herbivory.
Of the 909 seedlings in the grids left exposed to
herbivory, 293 were consumed over the course of the
experiment (see Table 1). An additional 34 individuals
suffered partial herbivory that did not kill the individual (17 V. sativa, 10 Avena barbata, and 7 Hemizonia
congesta). Only three individuals suffered mortality
due to other causes (e.g., frost or disease), in which
case a dead individual remained in place. The consumption statistics were calculated based only on herbivory that resulted in mortality. Seedling loss due to
mortality was recorded at multiple times, but the majority of seedlings were consumed during the first three
weeks of the experiment (Table 1). The remaining results for herbivore consumption are based only on the
seedlings consumed during the first three weeks of the
experiment. This avoids any confounding effect of the
harvests on the subsequent number of potential seedlings available to herbivores, although the trends during
the first three weeks (Fig. 2A) held true for the total
number of seedlings consumed over the three-month
experiment (Fig. 2B).
Herbivore consumption differed among species under ambient conditions (species P , 0.0001). The grass
species that are most abundant in the surrounding community had the lowest tissue N concentrations (A. barbata, Bromus hordeaceus) and were seldom consumed,
while the rare legume species had higher tissue N concentrations (Lotus purshianus and V. sativa), and were
HERBIVORY AND GLOBAL CHANGE
March 2006
689
FIG. 1. The proportion of seeds that germinated (mean 6 SE) for the six focal species under ambient and treatment
conditions. Species are abbreviated by the first letter of their genus and species names, which are listed in Methods.
often consumed. The native annual forbs did not seem
to fall predictably into this pattern; Epilobium brachycarpum and H. congesta both had the percentage of N
close to the grasses, but were consumed more often,
perhaps because of their rarity in the surrounding matrix.
Consumption declined overall under elevated CO2,
but this effect was counterbalanced in combination
with N deposition (CO2 3 N, P 5 0.05). This effect
was reversed for H. congesta (CO2 3 N 3 species, P
5 0.07), which was consumed almost twice as often
under elevated CO2 alone as under ambient conditions,
but this effect was attenuated in combination with N
deposition. The species 3 CO2 and species 3 N interactions are qualitatively apparent but could not be estimated statistically, because in the first three weeks of
TABLE 1.
the experiment, A. barbata was never consumed under
elevated CO2 nor elevated CO2 with N deposition, and
B. hordeaceus was never consumed under elevated CO2
alone (Fig. 2A).
Seedling nitrogen and growth rates
Species differed significantly in the tissue N content
of their seedlings (species, P , 0.0001). N deposition
increased seedling percentage of N for all species except V. sativa (Fig. 3A), resulting in a significant primary effect (N, P , 0.0001), as well as an interaction
with species (species 3 N, P 5 0.04). Seedling growth
rates also varied among species and increased under N
deposition for all species except V. sativa (species, P
, 0.0001; N, P 5 0.002; species 3 N, P 5 0.05, Fig.
3B). A post hoc Pearson correlation of the mean shift
Seedlings were consumed at different times throughout the experiment.
Number of seedlings consumed
Functional
group
Total
seedlings
25 Nov–
13 Dec
13 Dec–
5 Jan
6 Jan–
12 Feb
Total
consumed
Inside exclosure
A. barbata
B. hordeaceus
E. brachycarpum
H. congesta
L. purshianus
V. sativa
Total
grass
grass
forb
forb
legume
legume
179
184
180
174
49
265
1031
6
5
10
51
16
49
137
3
2
10
19
7
15
56
0
0
1
6
1
3
11
9
7
21
76
24
67
204
Open to herbivory
A. barbata
B. hordeaceus
E. brachycarpum
H. congesta
L. purshianus
V. sativa
Total
grass
grass
forb
forb
legume
legume
171
183
179
118
36
222
909
6
14
17
51
12
55
155
6
6
16
40
12
44
124
0
0
6
3
0
5
14
12
20
39
94
24
104
293
Location/species
Note: The table shows the total number of seedlings (germinated 1 transplanted) consumed
for each time period, as well as summed over the course of the experiment. Data are displayed
for the grids covered by an herbivore exclosure, as well as the open grids.
ELSA E. CLELAND ET AL.
690
Ecology, Vol. 87, No. 3
ability, because in California annual grasslands nearly
all annual grass and forb seeds germinate each autumn
after the first substantial rain event (McNaughton
1968). It has been assumed that less than half of germinating seeds survive those first few weeks (Bartolome 1979, Young and Evans 1989); in this experiment
17% of seedlings exposed to herbivory were consumed
in the first three weeks, and 32% were consumed in
the first three months, confirming the importance of
FIG. 2. The proportion of seedlings consumed (mean 6
in the open grids under ambient and treatment conditions
for each of the six focal species (A) during the period between
25 November and 13 December, and (B) over the whole
course of the experiment between 25 November and 12 February.
SE )
in seedling %N and mean shift in growth rate per species in response to N deposition revealed a tight correlation (r 5 0.97, P , 0.01, Fig. 3C), confirming the
expectation that shifts in N content and growth rates
would be coupled.
There was not a statistically significant effect of elevated CO2 on seedling %N or growth rate; however,
a post hoc comparison revealed that species with high
seedling %N under ambient conditions had the largest
declines in %N under elevated CO2 (Fig. 4B). Pearson
correlation between the mean ambient %N and the
mean shift in %N for each species was highly significant (r 5 20.96, P 5 0.002). Under ambient conditions, species with higher %N were consumed more
often; there was a significant correlation between the
mean %N and mean consumption rate for each species
(r 5 0.86, P 5 0.03; Fig. 4A). However, the observed
shifts in seedling %N and growth rates in response to
N deposition and elevated CO2 did not predict the observed shifts in herbivore consumption (Fig. 4C).
DISCUSSION
The global change treatments had little impact on
germination rates. This was surprising, as nitrate addition is known to stimulate germination in some species (Fenner 1985). In this experiment, germination
was more likely initiated by increased water avail-
FIG. 3. The effect of N deposition on (A) seedling N
content (%), (B) growth rate, and (C) the change in seedling
N content (N deposition [%] minus ambient N [%]) vs. the
change in growth rate (N deposition rate minus ambient rate),
for the six focal species grown inside the exclosures. In (B)
growth rate is the slope of the regression for seedling mass
(measured in grams) over time: log (mass per harvest interval). Species are abbreviated by the first letter of their genus
and species names, which are listed in Methods. Error bars
in parts (A) and (B) represent 6 SE.
March 2006
HERBIVORY AND GLOBAL CHANGE
FIG. 4. (A) The proportion of total seedlings of each species consumed in relation to N content (%) under ambient
conditions. (B) The shift in tissue N content caused by the
global change treatments (treatment N [%] minus ambient N
[%]) in relation to seedling N (%) under ambient conditions
[key to symbols as defined in panel (C)]. (C) The shift in
seedling N content vs. the shift in consumption (treatment
proportion consumed minus ambient proportion consumed)
under the three global change treatments. In all panels, each
point represents the treatment mean for one of the six focal
species.
early-season herbivory in structuring annual grassland
communities.
Elevated CO2 at ambient N levels caused decreased
gastropod consumption of seedlings. The effect of elevated CO2 was not explained by shifts in N content,
691
was stronger for grass seedlings, and was modulated
by N deposition for the forbs and legumes (Fig. 2A).
Thus, we found no evidence of compensatory consumption. Many studies finding compensatory consumption (e.g., Lincoln et al. 1993, Bezemer and Jones
1998) have been performed under controlled laboratory
or greenhouse conditions in the absence of predators,
and without allowing herbivores to choose which species they consume, as other authors have noted (e.g.,
Peters et al. 2000, Stiling et al. 2002). Rather, our findings support the alternative hypothesis that herbivores
will shift their preferences under elevated CO2. Controlled feeding trials using plants grown in the JRGCE
also found that gastropods displayed shifting preferences as opposed to compensatory consumption (Peters
et al. 2006). In a similar experiment in Switzerland,
Ledergerber et al. (1997) found that elevated CO 2 had
no effect on overall gastropod consumption rates, because some species were consumed more, while others
were consumed less.
Other experiments have also found lower rates of
herbivory under elevated CO2, in some cases mediated
by lower herbivore densities (Stiling et al. 2002), while
in others it was unclear whether the response was due
to shifts in herbivore populations or to individual herbivore consumption rates (Hamilton et al. 2004). Concurrent research in the JRGCE found no differences in
early-season gastropod counts among plots receiving
ambient vs. elevated CO2 or N deposition (H. A. Peters
and E. E. Cleland, unpublished data), suggesting the
lower consumption rates we observed under elevated
CO2 were not due to a decline in gastropod populations.
However, in plots receiving elevated CO2, individual
gastropods caught in traps had lower body mass early
in the season (H. A. Peters and E. E. Cleland, unpublished data). It is not clear if this is the cause or the
effect of decreased seedling consumption.
Diaz et al. (1998) found that herbivore preferences
were not shifted by elevated CO2, even though N content declined, probably because the species with the
largest N declines were not preferred under ambient
conditions. In this study, the species with the largest
declines in N content under elevated CO2 were legumes, which were the most preferred under ambient
conditions (Fig. 4A). Gastropod consumption of L. purshianus was considerably lower under elevated CO 2,
while consumption of V. sativa was unchanged. Some
legumes produce defensive cyanogenic glucosides
(Jones 1972), which discourage gastropod herbivory
(Dirzo and Harper 1982, Hulme 1994). Eucalyptus
grown under elevated CO2 has been shown to increase
N allocation toward the production of cyanogenic glycosides (Gleadow et al. 1998), compounds similar in
function to those produced in legumes. This suggests
that the decreased consumption of L. purshianus under
elevated CO2 in this experiment may have resulted from
allocation to defense compounds. Similarly, two grass
species that were consumed at low levels under ambient
692
ELSA E. CLELAND ET AL.
conditions were not consumed during the early part of
the growing season under elevated CO2, even though
their N contents changed little under elevated CO2, suggesting these species may have allocated resources toward chemical or physical defense. This possibility is
intriguing, as seedlings are generally assumed not to
allocate toward defense compounds (Fenner 1987,
Hanley et al. 1995a, Fenner et al. 1999). Previous work
in the JRGCE has documented increased lignin and
total nonstructural carbohydrates in senesced annual
grasses grown under elevated CO2 (Henry et al. 2005),
but detailed seedling tissue chemistry has not been attempted due to the difficulty of identifying newly germinated seedlings, and the extremely small size of
seedlings, which would necessitate pooling many individuals to obtain the mass necessary to analyze.
N deposition increased both seedling N content and
growth rate (Fig. 3). The tight coupling of these traits
confirms that the recently germinated seedlings were
nutrient limited, as has been previously observed
(Grime and Curtis 1976). N deposition did not cause
predictable shifts in the consumption rate of seedlings
in this experiment, despite the changes in tissue N,
indicating the per individual consumption of seedling
N likely increased. Nitrogen is a limiting nutrient to
growth for many herbivores, and N content is considered to be the best predictor of plant quality for herbivores (Mattson 1980, White 1993, but see also Speiser and Rowellrahier 1991). However, the responses of
herbivores to N fertilization have been mixed (e.g.,
Silliman and Zieman 2001, Valentine and Heck 2001,
Campo and Dirzo 2003, Throop and Lerdau 2004). Herbivores may have an optimum N consumption, and
while some herbivores have been observed to increase
consumption of N-enriched plants, they may also decrease consumption rates if N concentrations exceed
their optimum level (Mattson 1980). Alternatively, herbivore N consumption may depend on the availability
and stoichiometry of other nutrients, in particular carbon and phosphorus (Sterner and Elser 2002). In this
study N addition increased N content, but decreased
consumption of L. purshianus. As discussed in the previous paragraph, this decline in herbivory could also
have been the result of increased N allocation toward
defensive compounds.
In addition to being influenced by tissue quality, gastropods are known to demonstrate frequency-dependent
consumption, disproportionately preferring to consume
the rare species in a community (Cottam 1985). Although I sought to achieve even densities of the six
focal species within each experimental grid, the densities of these species vary greatly in the surrounding
community matrix, and the gastropod herbivores were
free to eat either the planted seedlings or those in the
matrix. Under ambient conditions the least abundant
species were the most often consumed. Herbivores may
have a disproportionate influence on species that are
already rare, and at greater risk of local extinction in
Ecology, Vol. 87, No. 3
response to changing environmental conditions (Tilman and Elhaddi 1992).
The results from this study indicate that observed
shifts in species composition in the JRGCE may be
mediated by selective herbivory, particularly under elevated CO2. In an analysis of the effects of the global
change treatments on species composition, Zavaleta et
al. (2003) found that N deposition and elevated CO2
tended to increase the abundance of grasses relative to
forbs. Decreased consumption of grass seedlings under
elevated CO2 both in this study and in controlled feeding trials (Peters et al. 2006) indicates that herbivory
may allow grasses to become more common under elevated CO2. In this study N deposition did not decrease
herbivory of grass seedlings, nor increase herbivory of
forb seedlings. However, feeding trials using material
from adult plants did find that N deposition increased
consumption of A. barbata and Geranium dissectum
(Peters et al. 2006). These conflicting results may have
arisen because in the field, gastropods can choose
among many species, while in the feeding trials the
only choices were A. barbata, V. sativa, and G. dissectum. (This common nonnative forb is seldom consumed under ambient conditions in this system, unlike
the native forbs observed in this study.) Alternatively,
the findings of this study suggest that other mechanisms
may contribute to the dominance of grasses under high
N conditions, such as decreased light availability for
forbs with short stature, due to shading from increased
litter production (Facelli and Pickett 1991, Wilby and
Brown 2001).
In this experiment, simulated global changes altered
consumption rates of seedlings by gastropod herbivores, supporting the hypothesis that shifting plant–
animal interactions will influence the response of terrestrial ecosystems to impending global changes. These
shifts in consumption rates were not explained by shifts
in tissue N content or growth rate, despite extensive
theory and numerous single-species experiments showing that shifts in tissue chemistry imposed by global
changes should lead to shifts in herbivore consumption
(Bazzaz 1990, Lambers 1993, Lincoln et al. 1993, Bezemer and Jones 1998, Throop and Lerdau 2004). Many
of these studies have concentrated on shifts in insect
herbivory; the results of this study call attention to the
need to determine the responses of other taxa in ecosystems where they may play an important role in herbivory.
ACKNOWLEDGMENTS
I am grateful to the many collaborators who support and
maintain the JRGCE. Nona Chiariello gave invaluable advice
during the development of this research. I would like to thank
Stan Harpole, Katie Suding, Peter Vitousek, and two anonymous reviewers for helpful comments on this manuscript.
This research was informed by conversations with Brian
Thomas about the importance of gastropod herbivores in the
ecosystems at Jasper Ridge. Advice from Kathleen Kiernan
at SAS technical support regarding the MIXED and GLIMMIX statistical models was greatly appreciated. E. E. Cleland
HERBIVORY AND GLOBAL CHANGE
March 2006
was supported by a U.S. Dept. of Energy Global Change
Education Program GREF Fellowship. This research was supported by a California Native Plant Society Scholarship. The
JRGCE has been supported by grants from the National Science Foundation, the Morgan Family Foundation, the David
and Lucile Packard Foundation, Jasper Ridge Biological Preserve, and the Carnegie Institution of Washington.
LITERATURE CITED
Bartolome, J. W. 1979. Germination and seedling establishment in California annual grassland. Journal of Ecology
67(1):273–281.
Bazzaz, F. A. 1990. The response of natural ecosystems to
the rising global CO2 levels. Annual Review of Ecology
and Systematics 21:167–196.
Bergelson, J. 1990. Spatial patterning in plants opposing effects of herbivory and competition. Journal of Ecology
78(4):937–948.
Bezemer, T. M., and T. H. Jones. 1998. Plant-insect herbivore
interactions in elevated atmospheric CO2: quantitative analyses and guild effects. Oikos 82(2):212–222.
Campo, J., and R. Dirzo. 2003. Leaf quality and herbivory
responses to soil nutrient addition in secondary tropical dry
forests of Yucatan, Mexico. Journal of Tropical Ecology
19:525–530.
Cates, R. G., and G. H. Orians. 1975. Successional status
and the palatability of plants to generalized herbivores.
Ecology 56:410–418.
Coley, P. D., J. P. Bryant, and F. S. Chapin. 1985. Resource
availability and plant antiherbivore defense. Science
230(4728):895–899.
Cortrufo, M. F., P. Ineson, and S. Andy. 1998. Elevated CO2
reduces the nitrogen concentration of plant tissues. Global
Change Biology 4(1):43–54.
Cottam, D. A. 1985. Frequency-dependent grazing by slugs
and grasshoppers. Journal of Ecology 73(3):925–934.
Cottam, D. A. 1986. The effects of slug-grazing on Trifoliumrepens and Dactylis-glomerata in monoculture and mixed
sward. Oikos 47(3):275–279.
Cusbasch, U., G. A. Meehl, G. J. Boer, R. J. Stouffer, M.
Dix, A. Noda, C. A. Senior, S. Raper, and K. S. Yap. 2001.
Projections of future climate change. In J. R. Houghton et
al., editors. Climate change 2001: the scientific basis. Contribution of Working Group I to the third assessment report
of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK.
Davidson, D. W. 1993. The effects of herbivory and granivory on terrestrial plant succession. Oikos 68:23–35.
Diaz, S., L. H. Fraser, J. P. Grime, and V. Falczuk. 1998. The
impact of elevated CO2 on plant-herbivore interactions: experimental evidence of moderating effects at the community level. Oecologia 117(1–2):177–186.
Dirzo, R. 1980. Experimental studies on slug-plant interactions. 1. The acceptability of 30 plant-species to the slug
Agriolimax caruanae. Journal of Ecology 68(3):981–998.
Dirzo, R., and J. L. Harper. 1980. Experimental studies on
slug-plant interactions. 2. The effect of grazing by slugs
on high-density monocultures of Capsella-bursa-pastoris
and Poa-annua. Journal of Ecology 68(3):999–1011.
Dirzo, R., and J. L. Harper. 1982. Experimental studies on
slug plant interactions 4. The performance of cyanogenic
and acyanogenic morphs of Trifolium repens in the field.
Journal of Ecology 70(1):119–138.
Facelli, J. M., and S. T. A. Pickett. 1991. Plant litter: its
dynamics and effects on plant community structure. Botanical Review 57(1):1–32.
Fenner, M. 1985. Seed ecology. Chapman and Hall, London,
UK.
Fenner, M. 1987. Seedlings. New Phytologist 106(1):35–47.
693
Fenner, M., M. E. Hanley, and R. Lawrence. 1999. Comparison of seedling and adult palatability in annual and
perennial plants. Functional Ecology 13(4):546–551.
Gleadow, R. M., W. J. Foley, and I. E. Woodrow. 1998. Enhanced CO2 alters the relationship between photosynthesis
and defense in cyanogenic Eucalyptus cladocalyx f. Muell.
Plant, Cell and Environment 21:12–22.
Grime, J. P., and A. V. Curtis. 1976. The interaction of
drought and mineral nutrient stress in calcareous grassland.
Journal of Ecology 64(3):975–988.
Grime, J. P., S. F. MacPherson-Stewart, and R. S. Dearman.
1968. An investigation of leaf palatability using snail Cepaea nemoralis l. Journal of Ecology 56(2):405–415.
Hamilton, J. G., A. R. Zangerl, M. R. Berenbaum, J. Pippen,
M. Aldea, and E. H. DeLucia. 2004. Insect herbivory in
an intact forest understory under experimental CO2 enrichment. Oecologia 138(4):566–573.
Hanley, M. E., M. Fenner, and P. J. Edwards. 1995a. The
effect of seedling age on the likelihood of herbivory by
the slug Deroceras reticulatum. Functional Ecology 9(5):
754–759.
Hanley, M. E., M. Fenner, and P. J. Edwards. 1995b. An
experimental field study of the effects of mollusk grazing
on seedling recruitment and survival in grassland. Journal
of Ecology 83(4):621–627.
Henry, H. A. L., E. E. Cleland, C. B. Field, and P. M. Vitousek. 2005. Interactive effects of elevated CO2, N deposition and climate change on plant litter quality in a
California annual grassland. Oecologia 142(3):465–473.
Hickman, J. C. 1996. The Jepson manual: higher plants of
California. University of California Press, Berkeley, California, USA.
Hulme, P. E. 1994. Seedling herbivory in grassland: relative
impact of vertebrate and invertebrate herbivores. Journal
of Ecology 82(4):873–880.
Jones, D. A. 1972. Cyanogenic glucosides and their function.
Pages 103–124 in B. Harborne, editor. Phytochemical ecology. Academic Press, London, UK.
Jones, D. A. 1977. Plant chemistry and herbivory, or why
the world is green. Pages 284–324 in M. J. Crawley, editor.
Plant ecology. Blackwell Science, Oxford, UK.
Kielland, K., and J. P. Bryant. 1998. Moose herbivory in
taiga: effects on biogeochemistry and vegetation dynamics
in primary succession. Oikos 82(2):377–383.
Lambers, H. 1993. Rising CO2, secondary plant-metabolism,
plant-herbivore interactions and litter decomposition—theoretical considerations. Vegetatio 104:263–271.
Ledergerber, S., G. H. Thommen, and B. Baur. 1997. Grazing
damage to plants and gastropod and grasshopper densities
in a CO2-enrichment experiment on calcareous grassland.
Acta Oecologica–International Journal of Ecology 18(3):
255–261.
Lincoln, D. E., D. Couvet, and N. Sionit. 1986. Response of
an insect herbivore to host plants grown in carbon dioxide
enriched atmospheres. Oecologia (Berlin) 69(4):556–560.
Lincoln, D. E., E. D. Fajer, and R. H. Johnson. 1993. Plantinsect herbivore interactions in elevated CO2 environments.
Trends in Ecology and Evolution 8:64–68.
Littell, R. C., W. W. Stroup, and R. J. Freund. 1996. SAS
for mixed models. SAS Institute, Cary, North Carolina,
USA.
Mattson, W. J. 1980. Herbivory in relation to plant nitrogencontent. Annual Review of Ecology and Systematics 11:
119–161.
McNaughton, S. J. 1968. Structure and function in California
grasslands. Ecology 49:962–972.
Peters, H. A. 2004. Consumer control of animal-structured
plant communities in current and future environments. Dissertation. Stanford University, Stanford, California, USA.
694
ELSA E. CLELAND ET AL.
Peters, H. A., B. Baur, F. Bazzaz, and C. Korner. 2000. Consumption rates and food preferences of slugs in a calcareous
grassland under current and future CO2 conditions. Oecologia 125(1):72–81.
Peters, H. A., E. E. Cleland, H. A. Mooney, and C. B. Field.
2006. Herbivore control of annual grassland composition
in current and future environments. Ecology Letters 9:86–
94.
Poorter, H., Y. VanBerkel, R. Baxter, J. DenHertog, P. Dijkstra, R. M. Gifford, K. L. Griffin, C. Roumet, J. Roy, and
S. C. Wong. 1997. The effect of elevated CO2 on the chemical composition and construction costs of leaves of 27 C3 species. Plant, Cell and Environment 20(4):472–482.
Prentice, I. C., G. D. Farquhar, M. J. R. Fasham, M. L. Goulden, M. Heimann, V. J. Jaramillo, H. S. Kheshgi, C. Le
Quere, R. J. Scholes, and D. W. R. Wallace. 2001. The
carbon cycle and atmospheric carbon dioxide. Pages 198–
237 in J. R. Houghton et al., editors. Climate Change 2001:
the scientific basis. Contribution of Working Group I to the
third assessment report of the Intergovernmental Panel on
Climate Change. Cambridge University Press, Cambridge,
UK.
SAS Institute. 2004. SAS version 9.1. SAS Institute, Cary,
North Carolina, USA.
Shaw, M., E. Zavaleta, N. Chiariello, E. Cleland, H. Mooney,
and C. Field. 2002. Grassland responses to global environmental changes suppressed by elevated CO2. Science
298(5600):1987–1990.
Silliman, B. R., and J. C. Zieman. 2001. Top-down control
of Spartina alterniflora production by periwinkle grazing
in a Virginia salt marsh. Ecology 82:2830–2845.
Speiser, B., and M. Rowellrahier. 1991. Effects of food availability, nutritional value and alkaloids on food choice in
the generalist herbivore Arianta-arbustorum gastropoda
helicidae. Oikos 62(3):306–318.
Sterner, R. W., and J. J. Elser. 2002. Ecological stoichiometry: the biology of elements from molecules to the biosphere. Princeton University Press, Princeton, New Jersey,
USA.
Stiling, P., M. Cattell, D. C. Moon, A. Rossi, B. A. Hungate,
G. Hymus, and B. Drake. 2002. Elevated atmospheric CO2
Ecology, Vol. 87, No. 3
lowers herbivore abundance, but increases leaf abscission
rates. Global Change Biology 8(7):658–667.
Stokes, M. E., C. S. Davis, and G. G. Koch. 2001. Categorical
data analysis using the SAS system. SAS Institute, Cary,
North Carolina, USA.
Throop, H. L., and M. T. Lerdau. 2004. Effects of nitrogen
deposition on insect herbivory: implications for community
and ecosystem processes. Ecosystems 7(2):109–133.
Tilman, D., and A. Elhaddi. 1992. Drought and biodiversity
in grasslands. Oecologia 89(2):257–264.
Valentine, J. F., and K. L. Heck. 2001. The role of leaf nitrogen content in determining turtlegrass (Thalassia testudinum) grazing by a generalized herbivore in the northeastern Gulf of Mexico. Journal of Experimental Marine
Biology and Ecology 258(1):65–86.
Wand, S. J. E., G. F. Midgley, M. H. Jones, and P. S. Curtis.
1999. Responses of wild C4 and C3 grass (Poaceae) species to elevated atmospheric CO2 concentration: a metaanalytic test of current theories and perceptions. Global
Change Biology 5(6):723–741.
Weiss, S. B. 1999. Cars, cows, and checkerspot butterflies:
nitrogen deposition and management of nutrient poor grasslands for a threatened species. Conservation Biology 13(6):
1476–1486.
White, T. 1993. The inadequate environment: nitrogen and
the abundance of animals. Springer-Verlag, New York, New
York, USA.
Wilby, A., and V. K. Brown. 2001. Herbivory, litter and soil
disturbance as determinants of vegetation dynamics during
early old-field succession under set-aside. Oecologia
127(2):259–265.
Young, J. A., and R. A. Evans. 1989. Seed production and
germination dynamics in California annual grassland. Pages 39–45 in L. F. Huenneke and H. A. Mooney, editors.
Grassland structure and function: California annual grassland. Kluwer Academic, Dordrecht, The Netherlands.
Zavaleta, E., M. Shaw, N. Chiariello, B. Thomas, E. Cleland,
C. Field, and H. Mooney. 2003. Grassland responses to
three years of elevated temperature, CO2, precipitation, and
N deposition. Ecological Monographs 73:585–604.
APPENDIX
Supplementary information regarding statistical analyses and outputs (Ecological Archives E087-039-A1).