Most theoretical models of species coexistence assume that habitat patches... dynamic habitat P S

PROJECT SUMMARY
Overview
Most theoretical models of species coexistence assume that habitat patches are spatially fixed in
size and constant in quality. How communities are organized in systems with dynamic habitat
patches – those that are neither spatially fixed in size nor temporally constant in quality – is a
crucial yet unresolved ecological question. The proposed research uses realistic field
manipulations to understand the creation of dynamic habitat patches and how associated
communities interact with them. The research takes advantage of a model ecological system: the
carnivorous pitcher plant Sarracenia purpurea, its associated aquatic food web, its prey, and its
herbivores. In this system, the tubular leaves of the plant fill with rainwater and function as
dynamic habitat patches for a self-contained aquatic food web. Changes in patch location,
quantity, and quality may be mediated by reciprocal interactions and feedback loops among
patches (the plant and its growing leaves), its primary prey (ants), and its herbivores (moth larvae
that feed on the plants’ leaves). A set of field experiments are proposed in which patch size,
density, and spatial pattern, and levels of herbivory are manipulated. These experiments address
the following questions:
1. Do the feedback loops between the moth, the plant, and the ant result in local aggregations
of plants and spatial association between ant nests and plants?
2. Does spatial clustering of moth larva attacks leads to “traveling waves” of suitable and
unsuitable patches?
3. How does the structure of the aquatic food web in a pitcher respond to changes in patch
structure (leaf size and number of plants)?
Intellectual Merit
The research proposed here is a novel experimental study of species interactions in dynamic
habitats, which change in size, quality, and spatial arrangement. The Sarracenia system is a
well-used model for studies of ecological stoichiometry, population dynamics in changing
environments, community assembly and metacommunity dynamics at local, regional, and
continental scales, and food web structure. The proposed research broadly expands the
theoretical context of patch dynamics and the scope of Sarracenia research by exploring
mechanisms responsible for creating and maintaining dynamic habitats, and testing their role in
structuring multi-trophic communities.
Broader Impacts
There are three broader impacts of the proposed research. First, undergraduates, graduate
students, teachers, research assistants, and post-docs will be educated and mentored through
continuous involvement in theory and methods of basic ecological research and in fundamentals
of natural history. Second, the proposed research will yield a general framework for
understanding how multi-trophic communities assemble and disassemble in changing
environments. Third, an established model ecological system – the Sarracenia system - will be
propelled to a new level, where it can be used to explore mechanisms driving food web dynamics
at the landscape scale.
A-1
I. Objective
The goal of this research is to experimentally elucidate how interactions among prey,
predators, and herbivores control the spatial pattern of dynamic habitat patches, and to
understand how these interactions indirectly affect multi-trophic communities inhabiting
these patches.
Community ecology seeks to explain the distribution and abundance of interacting species, but general yet realistic models of coexistence remain elusive. For analytical convenience,
models of species coexistence in patchy landscapes (e.g., Hanski 1985, Mouquet & Loreau 2002)
use an important simplification – stable patches of fixed sizes with constant attributes. In the research proposed here, we will move beyond this simplification and study species interactions in
dynamic habitats, which change in size, quality, and spatial arrangement.
We will continue to use the
Sarracenia model system (Fig. 1) for the
proposed research. Sarracenia purpurea
is a long-lived perennial carnivorous
plant that grows in bogs, poor fens, and
seepage swamps throughout the eastern
U.S. and Canada. Sarracenia purpurea
serves three important ecological
functions. First, it is a predator; its tubeshaped leaves fill with water and capture
arthropod prey (mostly ants and flies).
Second, Sarracenia is prey for notcuid
moth larvae that feed on leaves,
inflorescences, and rhizomes. Finally,
Sarracenia is a dynamic habitat patch; a
Figure 1. The Sarracenia model system. The plant produces
specialized, co-evolved micro-ecosystem
6-10
pitcher-shaped leaves (center) each year; these leaves
develops in each intact Sarracenia leaf.
are pitfall traps with which this carnivorous plant captures
The captured prey are the inputs for this
prey. Sphagnum-nesting ants, such as Myrmica lobifrons
detritus-based ecosystem.
(inset red box in center), gather nectar (a carbon source)
from the plant’s extrafloral nectaries. A small percentage of
Past NSF funding to the PIs,
foraging ants drown in the rainwater-filled pitchers, in
along with research by several other
which midge and sarcophagid fly larvae shred the prey cargroups, has led to the emergence of this
casses. This detritus is the basis for the well-studied pitchersystem as a model for: studies of ecoplant food web (top left) of bacteria, protozoa and rotifers,
mites, mosquito and sarcophagid larvae. A plant consists of
logical stoichiometry (Ellison & Gotelli
a set of pitchers, each of which houses a unique food web
2002, Wakefield et al. 2005); population
whose structure depends on pitcher morphology and timing
dynamics in changing environments
of pitcher production. Larvae of the noctuid moth Exyra fax
(Bradshaw and Holzapfel 2001, Gotelli
(red box at top right) feed on and drain pitchers, reduce
& Ellison 2002c, 2005); community asplant growth, and remove food web habitat (pitchers with
red “×” in center), but provide nest sites for the ant Dolichosembly and metacommunity dynamics at
derus pustulatus (nest and ant at bottom right). The interaclocal, regional, and continental scales
tions among moths, plants, and ants, which are the focus of
(Ellison et al. 2003, Buckley et al. 2003,
this proposal, result in a shifting mosaic of plants in bogs
Trzcinski et al. 2003, Krawchuk & Tay(lower left), which themselves are isolated habitats on the
lor 2003, Miller & Kneitel 2005); and
New England landscape.
×
××
C-1
food web structure (Addicott 1974, Bradshaw & Creelman 1984, Kneitel & Miller 2002, Ellison
et al. 2003, Trzcinski et al. 2005).
Research on the Sarracenia system to date has focused primarily on structure and dynamics of Sarracenia’s aquatic food web, and on the interplay between the food web and the plant.
We now have a rich picture of the dynamics of this food web, but we have not yet addressed the
fact that its habitat – the plant population itself – is not simply a static landscape of green “testtubes” in which food web dynamics play out. Rather, it is a spatially structured, dynamic system.
The organization of this dynamic system appears to be controlled by interactions between Sarracenia, the ants that are its primary prey, and herbivores that kill leaves or entire plants. Our proposed research into these interactions will move this model system to a new level – where it can
be used to explore mechanisms driving food web dynamics at the landscape scale.
II. Theoretical Background and Motivation
Many theoretical models predict that two or more species competing for a single (or similar) resource can stably coexist in a landscape if there is sufficient dispersal, immigration, and emigration among habitat patches (Skellam 1951, Hutchinson 1951, Horn & MacArthur 1972). Recent
expansion of food web theory into a broader landscape context (Polis et al. 1997, 2005) explicitly acknowledges that fluxes of material resources, organisms, and energy can control trophic
dynamics and can both respond to and drive ecosystem processes (Holt 2005). Both metacommunity models (Mouquet & Loreau 2002) and landscape analyses of food webs incorporate a set
of assumptions about the structure of the “landscape” and the meaning of spatial heterogeneity: a
landscape is viewed as a mosaic of fixed patches connected by immigration and emigration of
organisms, and exchange of resources and energy (Horn & MacArthur 1972, Forman & Godron
1986). Patches differ in their fixed attributes, such as vegetation, soil, water, or geology (Cadenasso et al. 2005).
Horn & MacArthur (1972) recognized that patches of fixed sizes with constant attributes
were but an analytical convenience. If habitats appear, change, or disappear locally, a set of separate equations are required to link their dynamics with those of their dependent communities
(Horn & MacArthur 1972). Spatial and landscape models of population dynamics of single species (e.g., Clark 1991a, 1991b) or several competing species (e.g., Ellison & Bedford 1995, Olff
et al. 2000, Kennedy & Storer 2000, Ricketts 2001, Zavala & Zea 2004, Fuhlendorf & Engle
2004) have taken up Horn & MacArthur’s challenge and extended patch dynamic models to the
shifting mosaics of what we refer to as dynamic habitats. These models all illustrate that coexistence and persistence of species depend on the timescale at which habitats change and the spatial
linkages among them. However, experimental confirmation of processes that create dynamic
habitats and the roles of dynamic habitats in structuring ecological communities are lacking. Further, these models have mostly focused on communities consisting of competing species within a
single trophic level. Extensions of these models to encompass multi-trophic communities are required to comprehend the dynamics of complete food webs on landscapes.
The Sarracenia system (Fig. 1) offers both a unique opportunity and a generalizable
model with which to examine mechanisms responsible for creating and maintaining dynamic
habitats, and exploring their role in structuring multi-trophic communities. In a series of experiments focused on three trophic levels – ants (prey); plants (intermediate predator and habitat
patches for the aquatic food web); moths (herbivore and “top predator” in this three-level system) – we address the following questions and associated hypotheses:
C-2
1. WHAT SETS THE INITIAL CONDITIONS? We hypothesize that temporal feedbacks between
mobile prey and sessile predators result in spatial segregation of different prey species,
spatial association of prey and predators, and local aggregations of habitat patches.
2. WHAT CREATES AND MAINTAINS DYNAMIC HABITATS? We hypothesize that the spatial
clustering of herbivores leads to “traveling waves” of suitable and unsuitable patches.
3. HOW DO THESE PATCH DYNAMICS CONTROL STRUCTURE OF THE ASSOCIATED AQUATIC
FOOD WEB? We hypothesize that the structure of the aquatic food web in any given patch
(i.e., pitcher) is a function of number of patches, creation of new groups of patches (i.e.,
plants), and distance among groups of patches.
III. The Sarracenia system
A. Background
In the last decade, our research (see §IX. Results from Prior NSF Support, below), together
with work by several other research groups, has firmly established the Sarracenia system (Fig. 1)
as a model for exploring a wide range of ecological questions (reviews in (Ellison & Gotelli
2001, Ellison et al. 2003, Srivastava et al. 2004, Miller & Kneitel 2005). Until the mid-1990s,
research on this system focused on dynamics of aquatic food webs in single pitchers (e.g., Addicott 1974, Bradshaw & Creelman 1984, Heard 1994b, Błędzki & Ellison 1998), but more recent
work has emphasized the importance of scale-dependence on the composition and structure of
the Sarracenia food web (Harvey & Miller 1996, Miner & Taylor 2002, Ellison et al. 2003,
Kneitel & Miller 2003, Trzcinski et al. 2003, Krawchuk & Taylor 2003, Srivastava et al. 2004,
Miller & Kneitel 2005). We now can map dynamics of the Sarracenia system directly onto a
three-level hierarchy (Table 1) that theorists (Leibold et al. 2004, Amarasekare et al. 2004) have
Table 1. The hierarchy of scales at which metacommunity dynamics are analyzed (after Amarasekare et al. 2004
and Leibold et al. 2004), our proposed mapping of the landscape context of the Sarracenia system onto this hierarchy, and how this mapping can actually be observed and experimentally manipulated. Bold italic type indicates processes focused on in this proposal.
Metacommuunity
Patch
Occupied by a single
individual
Locality
A community consisting of a collection of
identical patches
Region
A metacommunity consisting of a collection
of localities linked by
dispersal
Definition in a:
Food web
Sarracenia system
A single pitcher with its food web whose
A habitat occupied by a existence is contingent on draining of the
single food web
pitcher by herbivorous moths (Exyra), and
whose structure depends on prey inputs.
A single plant, consisting of a collection of
n pitchers. Plant growth and new pitcher
production is controlled by nutrients minA collection of identieralized by food webs and shared among
cal habitats with simipitchers. These processes are modulated
lar food webs
by spatial and temporal distribution of
prey (ants) and removal of pitchers by
moth larvae (Exyra).
A collection of plants whose food webs are
linked by energy/nutrient flow from the
A collection of dispasurrounding bog into the pitchers; migrarate food webs linked
tion of bacteria, protozoa, rotifers and,
by allochthonous enmites; oviposition dynamics of adult Dipergy flow
tera; foraging behavior of ant colonies,
and dispersal and oviposition dynamics of
herbivorous moths.
C-3
defined as the theater in which metacommunity dynamics play out. Relevant nested scales include individual pitchers (“patches”), multiple pitchers within single plants (“localities”), clumps
of plants and groups of clumps in single bogs (“regions”), and multiple bogs in the landscape.
Within individual patches (pitchers), the presence and abundance of the top predators
Wyeomyia smithii and Fletcherimyia fletcheri exert top-down effects on bacteria, protozoa, and
rotifers (Addicott 1974, Błędzki & Ellison 1998, Kneitel & Miller 2002, Trzcinski et al. 2005).
Prey availability has bottom-up effects on bacteria, protozoa, rotifers, and mite population dynamics and species richness of protozoa (Kneitel & Miller 2002, Miller & Kneitel 2005, Trzcinski et al. 2005). Oviposition rates of both predators and of the midge Metriocnemus knabi, which
shreds captured prey, vary with pitcher size, and adult emergence is synchronized with the availability of newly-opened pitchers (Mogi & Mokry 1980, Istock et al. 1983, Heard 1994a, 1994c,
Krawchuk & Taylor 1999). Prey contribute ~10% of the plant’s nutrient budget (Chapin & Pastor 1995, Newell & Nastase 1998, Dixon et al. 2005); in the northern portion of its range, S. purpurea receives 20-50% of its N budget from atmospheric deposition (Błędzki & Ellison 1998).
Excess N input shifts leaf morphology from pitchers to non-carnivorous phyllodes (Ellison &
Gotelli 2002), which do not host food webs, but contribute ~ 25% more carbon per unit leaf area
than do pitchers (Ellison & Gotelli 2002). However, atmospheric deposition may eventually lead
to population decline and extinction as root and rhizome reserves are exhausted and seedlings are
killed by nitrogen “burn” (Gotelli & Ellison 2002c, Güsewell 2005, Gotelli & Ellison 2005).
Within localities (whole plants, which consist of a set of water-filled leaves), dispersal
among patches maintains populations of protozoa and rotifers that suffer high levels of predation
in individual pitchers (Błędzki & Ellison 1998, Kneitel & Miller 2003). Larvae of Fletcherimyia
can track prey by moving among pitchers on a
single plant (Trzcinski et al. 2003). Pitcherplant flowers also provide roosts and mating
sites for adult Fletcherimyia (Krawchuk and
Taylor 1999), which pollinate S. purpurea
(Burr 1979). Pitcher plants produce only one
flower per plant, and at our study sites ~5-30%
of the plants flower annually.
Within a region (an entire bog), plant
distribution and bog physiognomy affect oviposition dynamics of Wyeomyia, Metriocnemus, and Fletcherimyia (Heard 1994a, 1994c,
Miner & Taylor 2002). Plants are rarely randomly arrayed in a bog, but rather tend to form
clumps (Fig. 2; Ellison & Parker 2002, KrawFigure 2. Spatial arrangement of Sarracenia purchuk & Taylor 2003), in part because of the
purea adults (large circles) and seedlings (small cirextremely limited seed dispersal of Sarracenia
cles) in a 500 × 500 cm mapped plot at Hawley Bog,
(Ellison & Parker 2002). Krawchuk & Taylor
in 2001. Spatial clustering is significant up at scales
(2003) showed that the limited movement disup to 200cm (P < 0.05, Monte Carlo simulations on
tances of adult Wyeomyia and Metriocnemus
Ripley’s K); adult plants are clustered with adult
plants, and seedlings are clustered with seedlings
(10s of m) and Fletcherimyia (ca. 100m) re(Moran’s I = 3.4, P = 0.006). Figure from Ellison
sulted in positive effects of leaves, plants, or
and Parker (2002). Tick labels are cm.
clumps of plants on density of larvae of these
C-4
three species, and generally negative effects of
isolation of clumps of plants (Fletcherimyia
only) or peatlands (Wyeomyia and
Fletcherimyia) on larval density.
B. Preliminary data
ANTS are the primary prey of pitcher plants;
they account for the majority of individual prey
items and total prey biomass (Newell & Nastase
1998, Ellison et al. 2002, and unpublished). In
northern bogs, the most common ants are
Myrmica lobifrons (sensu Francoeur 1997),
Formica subaenescens (sensu Francoeur 1973),
Figure 3. Spatial distribution of ants and pitcher
and Dolichoderus pustulatus (Ellison et al. 2002,
plants at Hawley Bog. Circle size (largest = 13
Gotelli & Ellison 2002b). The first two and
individuals) is proportional to the number of Myrspecies nest in Sphagnum moss, whereas D. pusmica lobifrons ants (pink) caught in regularlyspaced pitfall traps and number of plants (green) in
tulatus builds carton nests, often in old or motha 0.25 × 0.25 m quadrat centered on each pitfall
damaged S. purpurea pitchers (Fig. 1). Plants
trap. Counts of ants are no more aggregated than
show distinct spatial clustering (Fig. 2; Ellison &
expected by chance (SADIE index of aggregation
Parker 2002), but ant occurrences appear to be
[Perry 1998] Ia = 1.004; P =0.4).
randomly arrayed in space (Fig. 3). The ant data
illustrated in Fig. 3 are counts of individuals collected in pitfall traps; detailed mapping of nests,
as proposed below, is required to test the hypotheses that ant nests are spatially associated with
plants, if different species of ants are spatially segregated, and how spatial distributions of nests
of different ant species are related to plant population structure.
MOTH larvae are among the few organisms that feed on the unpalatable leaves of Sarra10
A
Meters (South to North)
0
-10
-20
-30
Healthy
E only
P only
P and E
-40
-50
-50
-40
-30
-20
-10
0
10
Meters (West to East)
Figure 4. Distribution of Sarracenia and moth larvae at the
Tom Swamp study site in 2003. ○ – no herbivory; ● – plants
damaged by Exyra only; ▼– plants damaged by Papaipema
only; ▲– plants damaged by both Exyra and Papaipema.
Herbivore attacks are significantly clustered (P = 0.002). Figure and data from Atwater et al. (2006).
cenia purpurea (Rymal & Folkerts
1982). Papiapema appassionata is a
root-borer that kills entire plants. It is
rare throughout its range (it is listed as
“threatened” in Massachusetts), and
will not be examined in detail in this
proposal. Exyra fax eats pitcher leaves,
and occurs in > 30% of the plants at
our study sites in Massachusetts (Atwater et al. 2006). Each larva chews a
drainage hole at the base of the pitcher.
The damaged leaf cannot capture prey
or host a food web, but it does not kill
the plant outright (Fish 1976, Rymal &
Folkerts 1982, Folkerts & Folkerts
1996). After draining the pitcher, the
larva seals the top of the leaf either
with a web of silken threads or by girgirdling and closing the leaf just under
the pitcher lip. The larva then feeds on
C-5
the interior surface of the pitcher chamber (Jones 1921, Rymal & Folkerts 1982, Folkerts &
Folkerts 1996). Larvae can move among pitchers; we observed a single larva feeding on three
pitchers/plant in a single season. In our region, Exyra is univoltine. Larvae overwinter as third
instars, pupate in early spring, and eclose in late June.
REU student Dan Atwater studied the distribution of Sarracenia and the moth larvae at
Tom Swamp, a ~50 ha bog at the Harvard Forest LTER site (Atwater et al. 2006). He found that
moths fed preferentially on larger plants, and that larvae are spatially aggregated (Fig. 4). Given
the apparent selectivity of adults for ovipositing on larger plants and their low mobility – most
fly < 10 m from their eclosion site – we hypothesize that the temporal distribution of larvae
would lead to a “traveling wave” of attack. Once an individual within a cluster of adult plants is
attacked, larvae would inhabit all suitable plants. Plants may then be reduced in quality (size,
number of leaves) to the point at which the patch is no longer suitable, and ovipositing adults
seek other nearby patches. After Exyra has migrated out of a patch, the plants in that patch could
recover to suitable size until they were attacked again. Our proposed research tests this hypothesis by tracking spatial and temporal distribution of Exyra across the bog, and examines the impact of this patterning on interactions among ants, plants, and the plants’ food webs.
IV. Hypotheses and mechanisms
We will test the overall hypothesis that interactions mobile prey, sessile predators, and mobile herbivores yield a shifting mosaic of prey (the allocthonous input to the Sarracenia
food web), pitchers (predators of ants and the habitat for the food web), and plants (the ultimate producer of new pitchers), and that the structure of this mosaic is controlled in large
part by the herbivore. The field observations and experiments that we will conduct will test a
set of explicit mechanisms underlying the generation of this mosaic and its consequences for the
dynamics of Sarracenia and its aquatic food web.
1. Because pitcher plants provide an energy source (carbon via extrafloral nectaries) for foraging ants:
1.1. the distribution of ant (prey) nests and individual pitcher plants (predator) will be spatially correlated;
1.2. the strength of the spatial association of plants and ant nests will increase with plant
size and density;
1.3. the three common bog ant species compete for resources provided by Sarracenia, and
thus their nests will be spatially segregated within bogs.
2. Because pitcher plants prey on ants, and prey capture rate increases with plant size and
prey encounter rate:
2.1. growth rate and reproductive effort of individual plants will be positively associated
with the size of, and negatively associated with distance to, the nearest ant colony;
2.2. seed production and seedling distribution will be clumped within the bog and positively associated in space with ant nests;
2.3. “saturation” (§VI.B, Table 2, below) of the aquatic food web in Sarracenia pitchers
will be positively associated with both prey capture rate and spatial distribution of ant
nests.
3. Herbivory by Exyra, which results in loss of entire pitchers and alters the spatial arrangement of available habitat (pitchers and plants) within bogs, will disrupt the spatial association among ants and Sarracenia and the dynamics of the aquatic food web:
C-6
3.1.
3.2.
3.3.
3.4.
limited dispersal rates of adult Exyra lead to “traveling waves” of pitcher mortality
across bogs;
growth rate and reproductive effort of individual plants will be negatively associated
with loss of pitchers;
ant colony distribution will be hyperdispersed (negatively associated in space) relative
to Sarracenia attacked by Exyra;
saturation of the aquatic food web will be hyperdispersed relative to patches of plants
attacked by Exyra.
V. Proposed Research
We will use a combination of spatial mapping, field experiments and observations, and simulation modeling to understand the relationships between
Sarracenia, ant colonies, and moth attacks. All experimental
manipulations will be conducted at Tom Swamp, a ~50 ha bog at
the Harvard Forest LTER site (red square in Fig. 5). This site
supports a total Sarracenia population of >10,000 individuals,
approximately 30% of which are annually attacked by Exyra
(Atwater et al. 2006). We will also extensively sample spatial
patterns at 26 bogs in VT, MA, and CT (Fig. 5) and at 24
additional sites in NY, NH, and ME (Buckley et al. 2003).
A. Spatial and temporal associations of ants, plants, and moths
All three sets of hypotheses in §IV require detailed information
on the spatial location of plants, ant nests, and moth larvae, and
how these spatial patterns vary through time. We are especially
interested in predicting regional patterns that may drive local
interactions by scaling up from small-scale spatial processes.
We will first map the distribution of these species in 5
replicated 10 × 10 m plots at Tom Swamp. Plants with and without moths will be permanently
tagged, measured (Ellison & Gotelli 2002) and mapped ± 1 cm using a compass, meter tapes,
and a Sonin® electronic range-finder (Ellison & Parker 2002, Atwater et al. 2006). Entrances to
nests of Myrmica lobifrons, Formica subaenescens, and Dolichoderus pustulatus will be located
and mapped by following workers from bait stations placed at 1-m intervals throughout each
plot. Plant death, seedling recruitment, moth oviposition patterns, and temporal shifts in ant nests
will be documented by repeated sampling of these plots each summer. We will also measure pore
water NH4 and vegetation structure, as the distribution of these ants covary with these variables
(Ellison & Gotelli 2002, Ellison et al. 2002). Degree of spatial aggregation, regularity, or hyperdispersion within and among taxa will be assessed using nearest-neighbor methods (Dixon
1994); temporal changes in these patterns will be determined using SADIE - spatial analysis by
distance indices (Perry et al. 1999, Winder et al. 2001).
This sampling regime will provide detailed information on spatiotemporal patterns of
ants, plants, and moths at a single site; will patterns be similar among bogs? In Year 1 and Year
5 of this funding period, we will conduct a snapshot survey of the spatial pattern of plants, ant
nests, and moth attacks within each of 50 bogs throughout New England and New York (Fig. 5).
At each bog, we will map ants, plants, and moths within a fixed 10 × 10 m plot, measure pore
Figure 5. Bogs to be sampled
in VT, MA, and CT. Tom
Swamp site indicated in red.
These sites will be augmented
with 24 additional sites in NY,
NH, and ME (Buckley et al.
2003).
C-7
water NH4, vegetation structure, and record latitude, longitude, and bog elevation, all of which
will enter our statistical models as covariates (Gotelli and Ellison 2002b).
B. Manipulating spatial interactions
We propose a novel manipulation of spatial pattern at the “locality scale” (Table 1, above),
for which we can measure shifts in the spatial pattern of associated species. Our experiments
focus on experimental manipulations of the spatial arrangement of Sarracenia, because moths
and ants cannot be directly manipulated in this system. Plants that we have already raised from
seed in the greenhouse and growing in plastic pots filled with Sphagnum will be transplanted to
the field, where they will grow, overwinter, and persist in defined spatial configurations. To
avoid genetic contamination of the Tom Swamp population of pitcher plants, we will only use
plants grown from seed collected at Tom Swamp for all our manipulations. Thus, we will experimentally establish patches of plants that differ in size and density and then quantify the responses of ants and moths to these manipulations.
Number of Plants per Clump
EXPERIMENT 1 – LOCALITY-SCALE MANIPULATIONS OF PLANT SIZE AND DENSITY
Our first experiment addresses the hypotheses that plant size and density alter spatial patterns of
ant colonies (foragers and prey) and Exyra (herbivores), and examines prey capture rates by, and
food web structure in, individual leaves of plants of different sizes growing in different densities.
We will establish replicate plots with 6 categories of plant density (1 to 6 plants in a tight
clump), orthogonally crossed with 6 categories of plant size (range 2 – 20 cm rosette diameter).
Each plot will consist of 126 plants: 1 – 6 plants per clump × 6 plant size classes (Fig. 6).
Clumps of plants will be placed in an evenly-spaced 6 × 6 m grid, with random assignments of
different size × density treatments to each grid location, and a 1-m separation between adjacent
clumps. In this experimental regression design (Gotelli & Ellison 2004, Cottingham et al. 2005),
there are two continuous predictor variables, plant size and density, and a sample size of N = 36
plant size × density combinations per plot. We will establish 5 replicate plots, for a total of 5 ×
126 = 630 plants. Treatment plants will be allowed to persist in the field for the duration of the
proposed research (5 field seasons).
To address temporal changes in the
system (Winder et al. 2001) we will
6
measure spatial patterns of ants,
5
plants, and moth attacks at the start of
the experiment, and at the start and
4
end of each field season. Dead or
3
dying plants will not be replaced
during the course of the experiment
2
because they may serve as ant nest
1
habitat and may attract (or repel)
moths. The distribution of all
1
2
3
4
5
6
Plant Size Category
naturally occurring plants, ant coloFigure 6. Treatment combinations in a 6 × 6 experimental renies, and Exyra larvae on
gression design. Six combinations of plant size (small to large
unmanipulated plants within the 36
circles) are crossed with 6 combinations of plant number per
m2 plot also will be mapped at the
clump (dark blue to light blue). In the field, these plants will be
start of the experiment.
placed in a regular spatial array, but the position of each particular treatment within the grid will be randomized.
C-8
At the end of each field season, we will measure the following variables within each plot:
1) Presence and density of Exyra on natural and experimental plants;
2) Occurrence and distribution of all ant nest entrances;
3) Prey and invertebrate abundance and food web structure and saturation (Table 2) within
the leaves of pitcher plants within
Table 2. How to quantify structure and saturation of the
each experimental clump, and
Sarracenia food web. As shown below, each taxon is given
within the leaves of the naturallya binary value; the saturation of the food web can be deoccurring plant nearest to each
scribed as the sum of the binary values for each taxon
clump.
present. Thus, a food web with a midge (Metriocnemus),
a rotifer (Habrotrocha) and a sarcophagid (Fletcheri4) Occurrence and distribution of any
myia) would be described as 1 + 10 + 10000 = 10011,
new Sarracenia seedlings within
for which the decimal equivalent = 19. There are 32 posthe plot;
sible food webs that can be assembled with these 5 taxa;
5) Occurrence and distribution of
the decimal value for each food web ranges from 0 – 31,
flowering plants (both exwith increasing numbers indicating more complete, or
saturated, food webs. Food web saturation increases with
perimental and natural);
prey availability, and is significantly lower in test-tubes
6) Leaf growth of all plants (both
or in plants lined with test-tubes than in intact plants
experimental and natural).
(Ellison et al. 2005).Thus, dynamics of the plant itself is
strongly involved in regulating food web structure.
The analysis will be based on
nearest-neighbor distances and Ripley’s K
Taxon
Binary value
for ant colonies and spatial association of
Metriocnemus
1
Habrotrocha
10
Exyra larvae with respect to size × density
Sarraceniopus
100
combinations of Sarracenia (Ripley 1976,
Wyeomyia
1000
Harkness and Isham 1983, Dixon 1994).
Fletcherimyia
10000
Null expectations of response variables
will be derived from computer simulation, with random placement of discs (ant nest entrances)
or points (Exyra larvae) within a mapped plot (Winder et al. 2001). The occurrence of Exyra attacks and flowering plants will be analyzed by simulating the random occurrence of Exyra and
plant flowering among the established plants (given their fixed spatial distribution), and then
measuring nearest neighbor distances and other measures of spatial aggregation and segregation.
Logistic regression will be used to assess the effects of plant size and clumping on the probability of flowering and the probability of Exyra attack. Regression slope parameters from these
analyses will provide inputs for the long-term simulation model (described below).
EXPERIMENT 2 – REGIONAL-SCALE MANIPULATION OF PLANT SIZE AND DENSITY
The first experiment will reveal effects of leaf size (“patch”), and plant size and density in each
“locality” (Table 1). But do these responses “scale up” to the “region”? To answer this question,
we will use a whole-plot experiment in which we will establish 5 plots of 6 × 6 m, each of which
will receive a different treatment: The plots will be carefully chosen at the start of the experiment
to minimize difference in background vegetation, microtopography, and Sarracenia density.
Background variation in established plants will be measured as a covariate, but not controlled for
experimentally.1 In the manipulated plots, we will transplant 36 clumps of the same treatment in
an evenly spaced array within each plot:
1) Control: no manipulations
1
We have chosen not to remove background plants in each plot. Removals would destroy bog vegetation unnecessarily and would introduce artificial disturbances that would probably affect ant distributions.
C-9
2) Small clumps (1 plant), small plant size (< 5 cm leaf length)
3) Small clumps (1 plant) large plant size (≥ 20 cm leaf length)
4) Large clumps (6 plants), small plant size (≤ 5 cm leaf length)
5) Large clumps (6 plants), large plant size (≥ 20 cm leaf length).
This experiment parallels at the regional (whole-plot) scale the manipulations at the locality
scale in EXPERIMENT 1, and we will measure the same set of variables. However, in this experiment we will make initial measurements in Years 1 and 2, make the transplants in Year 3, and
monitor post manipulation in Years 3, 4, and 5. Because this is an unreplicated BACI design
(Gotelli & Ellison 2004), we will use simple randomization tests to establish trends from the premanipulation data and then detect shifts in the response variables after the clumps are added
(Stewart-Oaten et al. 1992, Underwood 1994, Stewart-Oaten & Bence 2001).
In combination, EXPERIMENTS 1 and 2 will allow us to identify processes that set initial conditions for creation and maintenance of dynamic habitats, and how these processes
control the structure of a food web in a dynamic habitat. Specifically, these experiments
will reveal effects of plant size and density on prey capture, plant growth, and structure of
the Sarracenia food web, and on the spatial distribution of ant nests and moth larvae at the
locality and regional scales.
EXPERIMENT 3 – IMPACT OF EXYRA ON SARRACENIA AND ANTS
Exyra larvae cut holes at the base of Sarracenia pitchers (Fig. 7). These drained leaves contain
little or no rainwater, and lack the typical food web associated with S. purpurea. Loss of even a
single leaf may have important impacts on the long-term growth and survivorship of the plant.
Moreover, dead or dying leaves are used as habitat by the ant
Dolichoderus pustulatus, so there are potential direct and
indirect effects of Exyra attacks in this system. In our study sites
in MA, up to 30% of the plants are attacked by Exyra (Atwater
et al. 2006). However, densities can be low at other sites, and it
is difficult statistically to tease apart direct effects of Exyra
attacks from confounding effects of spatial variation in plant
microhabitat. Our proposed experiments allow us to more rigFigure 7. Example of Exyra
orously quantify the impacts of Exyra on ants and plants.
damage. Photo by Dan Atwater.
We will simulate Exyra attacks by cutting a hole in a
Sarracenia leaf that mimics the effects of larval feeding. To increase the realism of the experiment, we will rub the inside of the wound with scrapings from real Exyra attacks, to facilitate the
transfer of fungi and bacteria that are often associated with natural bore holes. Four experimental
treatments will be applied:
1) Control (no plant manipulations)
2) Single Exyra attack (only the first leaf of the season will be cut)
3) Multiple Exyra attacks (all leaves will be cut)
4) Single leaf removal (the first leaf of the season will be entirely removed before it opens).2
Treatments will be applied to both adult (≥ 10 cm longest leaf) and juvenile (< 5 cm longest leaf)
plants. Selected plants will be at located least one meter from their nearest experimental
2
There is one remaining orthogonal treatment in this design, which would be to removal all leaves of a plant. However, plants with no leaves will not survive more than one year, so we are foregoing this treatment.
C-10
neighbor. Ten plants will be assigned randomly to each treatment, for a total sample size of 80
plants. Treatments will be applied at the beginning of the field season in Year 2, and repeated in
Year 4. Plant growth, prey capture, and ant-nest distributions will be measured at the end the
field season. This is an orthogonal two-way design (treatment × plant size), and we will use a
priori contrasts to test specific comparisons among the different treatments.
EXPERIMENT 3 focuses specifically on processes that maintain spatial patterning of
dynamic habitats and on temporal shifts in these patterns. Specifically, this experiment
tests for direct effects of Exyra herbivory and subsequent leaf loss on plant growth, morphology, and survivorship, and will reveal secondary effects of moths on ant nesting behavior and density.
C. Modeling temporal changes in spatial pattern
We will develop a spatially explicit simulation model to understand the consequences of species
interactions between Sarracenia, ants, and Exyra (Figure 8; next page). The simulation will randomly place circles (ant nests) in a bounded two-dimensional space, just as in our field sampling
and experimental schemes. A simple null model will place these ant occurrences randomly
within the plot. The interaction model will choose a random location, but the probability of occurrence will be proportional to a nearest-neighbor distance function that takes into account the
existing spatial position of plants and nest entrances. The function itself will be based on the experimental regression results (EXPERIMENT 1 – §V.B) or the patterns of covariation among species (SAMPLING DATA – §V.A.). Sarracenia recruitment is rare at this small spatial scale, but we
will incorporate a simple random recruitment function to introduce new Sarracenia into the
model plot. The model will have a discrete time step of one year for plants, which will recruit,
grow, or die annually. Thus, the habitat patch – the plant – has its own set of equations in this
model, as required for dealing with dynamic habitats. Ant nest movements and Exyra attacks will
have a semi-annual time step, with movement, growth, or death of ant colonies recorded twice
yearly, to reflect within-year variation in these parameters that we will measure from our field
data (EXPERIMENTS 1-3 – §V.B). The model will reveal how species interactions alter spatial
patterns in this system, and how spatial patterns themselves change stochastically through
time, in both stationary and non-stationary environments.
VI. Feasibility and Timetable
The proposed 5-year project (4/1/2006 – 3/28/2011) spans 5 field seasons (F: April-October) and
5 off-seasons (November-March). Experimental work will begin immediately in Year 1.
Task
Spatial mapping – Tom Swamp
Spatial mapping – 50 bogs in New England
Experiment 1
Experiment 2 – pre-treatment monitoring
Experiment 2 – post-treatment monitoring
Experiment 3
Simulation model development
Manuscript preparation and submission
2006-07
F
O
2007-08
F
O
2008-09
F
O
2009-10
F
O
2010-11
F
O
C-11
Figure 8. Flow chart of the simulation model of interactions prey (ants), predators/habitat patches (Sarracenia),
the aquatic food web that depends on the patches, and herbivores (moths). Parameters for major model pathways will be derived from experiments and surveys that are illustrated. The model will have an annual time step
for Sarracenia dynamics and a semi-annual time step for ants, moths, and the food web.
VII. Responsibilities of the Investigators
PI Ellison and research assistant Jessica Butler will be responsible for growing plants for use in
the field experiments, and for coordinating the field manipulations at Tom Swamp. PI Gotelli
will be responsible for computer modeling work and for coordinating the extensive field sampling in New York and New England. As with our previous awards, both PIs, our students, and
our research assistant will share responsibilities for data analysis and manuscript production.
All data will be permanently archived in the Harvard Forest Archives, and in the Harvard
Forest LTER on-line data catalog (http://harvardforest.fas.harvard.edu/data.html). We budget
~10% of total direct project costs per year for data management costs at the Harvard Forest.
C-12
VIII. Significance of the Proposed Research and Broader Impacts
Ecologists recognize the importance of patchy habitats for species coexistence in models of
metacommunities and models of food web structure, but most models and empirical systems
continue to treat patches as static elements of the landscape. In nature, community and food web
processes interact with biotic habitats and change their spatial distribution at the landscape scale.
The overall objective of the proposed research is to understand how species interactions change
the spatial distribution of dynamic habitat patches at the landscape scale and affect food web
structure within and among those patches. Using the Sarracenia model system, we will experimentally study how the spatial distribution of Sarracenia and the structure of its food web interact with two other taxa: ants, the prey base for the plant and its food web; and herbivorous moth
larvae, which drain the pitcher-shaped leaves and remove habitat for the food web. Our results
will be integrated into a simulation model that will predict the changes in spatial patterns that
occur because of the direct and indirect effect of species interactions among the ants, moths,
plants, and the aquatic food web hosted by the plant.
There are three broader impacts of the proposed research. First, we will sustain our commitment to educate and mentor undergraduates, graduate students, teachers, research assistants,
and post-docs through continuous involvement in basic ecological research. This commitment is
epitomized by the fact that REU students have contributed substantively to the preliminary data
undergirding this proposal (§III.B. and Fig. 4), and we currently engage a local K-6 teacher in
this research through an RET supplement. Second, we will develop a general framework for understanding how multi-trophic communities assemble and disassemble in changing environments. Third, we will further develop our understanding of the natural history and the architecture of an emerging model system in ecology: the Sarracenia system.
IX. Results from prior NSF support
We have had two previous collaborative awards for our Sarracenia research – DEB 98-05722 /
DEB 98-08504 (Inquiline communities in changeable pitchers: do nutrients link community assembly to dynamic habitats?) and DEB 02-35128 / 02-34710 (Effects of nutrient stress on a coevolved food web).
A. Goals and key achievements of previous proposals
This on-going collaborative project has had three overarching goals:
(1) Analyze and predict assembly and persistence of multi-trophic communities in a habitat
that changes on the same time scale as that governing the assembly of the community.
Our experimental system is the “Sarracenia model system” (Ellison and Gotelli 2001, Ellison et al. 2003, Buckley et al. 2003).
(2) Construct a nitrogen budget for the Sarracenia system and assess how the dynamics of
this system are changing in response to persistent and increasing levels of anthropogenic nitrogen deposition. Sources of N include NH4-N and NO3-N from atmospheric
deposition (Ellison and Gotelli 2002), mineralization of captured prey by the arthropod
food web (Ellison et al. 2002, Wakefield et al. 2005), direct excretion by rotifers (Błędzki
and Ellison 1998, 2002), and organic N absorbed directly by pitchers from prey and by
roots from the peat (Karagatzides et al. 2005). We have demonstrated experimentally that
plant leaf allocation patterns, growth, and reproduction (Ellison and Gotelli 2002, Ellison
et al. 2004), and effects of plant structure on the food web and vice-versa all are driven by
nutrient availability (Ellison et al. 2005). We have used stable isotopes – both natural
C-13
abundance and 15N- and 13C-enriched – to trace organic N, prey N, and N from atmospheric deposition through the plant and the food web (Karagatzides et al. 2005, Butler and
Ellison 2006). The whole system is changing rapidly due to persistent nitrogen deposition
over much of the range of S. purpurea (Ellison and Gotelli 2002, Gotelli and Ellison
2002c, Buckley et al. 2003, Ellison et al. 2004, Wakefield et al. 2005).
(3) Develop models that link the assembly and persistence of the food web to the demography of S. purpurea. A unique and central aspect of our work on the Sarracenia system is
the explicit incorporation of habitat (i.e., plant) dynamics on food web assembly, structure,
and dynamics. We have developed individual-based, stochastic models of the growth of
individual plants (Ellison and Gotelli in prep. b), stochastic matrix models of S. purpurea
population dynamics (Gotelli and Ellison 2002c, 2005), and examined the relationship of
these models to the food web (Ellison and Gotelli in prep. a).
We have addressed these goals through a series of 16 manipulative experiments:
•
Two 2-year field experiments examined the assembly of the Sarracenia food web in leaves of different
sizes and ages (Ellison et al. 2003, Ellison and Gotelli in prep. a).
•
Two 1-year field experiments manipulated nitrogen availability and food web composition and habitat to
examine how the food web and the plant interact in processing N (Ellison et al. 2005).
•
Four 3-year greenhouse and field press experiments quantified the effects of nutrient additions on plant
growth (Ellison and Gotelli 2002, Wakefield et al. 2005);
•
A 1-year greenhouse experiment determined the rate at which inorganic nitrogen from atmospheric deposition (as 15N-NH4NO3) cycles through the Sarracenia food web in the presence or absence of a prey source
(Butler et al. 2005b)
•
Two 2-year experiments measured uptake, translocation, remobilization, and turnover of 15N-NH4NO3 in S.
purpurea (Butler and Ellison 2006).
•
A 72-hour pulse-chase experiment determined that S. purpurea and its congener S. flava can directly assimilate organic N (as 15N, 13C-glycine) and showed that organic N is taken up more slowly than 15NNH4NO3 in S. purpurea (Karagatzides et al. 2005).
•
Based on cohorts of permanently marked plants, we constructed demographic models for plant population
growth (Gotelli and Ellison 2002c, 2005), and developed individual-based models that will allow us to link
food web dynamics, leaf morphology, and plant population dynamics (Ellison and Gotelli in prep. b).
•
Plant reproductive dynamics were characterized separately, in a series of two additional experiments that
we had not initially proposed (Ellison 2001, Ellison and Parker 2002).
•
Two field experiments measured prey capture rates as a function of habitat characteristics and plant morphology (Ellison and Gotelli unpublished, Hart et al. in prep.). These field studies led to three studies focused on structure of prey assemblages and pitcher-plant herbivores:
1. We discovered that the dominant prey of S. purpurea in Massachusetts and Vermont is the ant
Myrmica lobifrons, which was not known previously from New England. With additional support
from the Massachusetts Natural Heritage and Endangered Species Program, we explored ant diversity and assembly rules in pitcher-plant bogs throughout New England (Ellison et al. 2002,
Gotelli and Ellison 2002a, 2002b).
2. We observed that red spotted newt (Notophthalmus viridescens) larvae were regularly captured by
S. purpurea, and we examined the relationship between plant morphology and newt capture, and
the potential contribution of newts to the plant’s N budget (Butler et al. 2005).
3. We examined spatial distribution of two obligate pitcher-plant herbivores, larvae of the noctuid
moths Exyra fax and Papaipema appassionata, and impacts of herbivory on plant growth (Atwater
et al. 2006).
C-14
Our results and models have provided a more complete picture of linkages between
pitcher-plant food webs and their host plants, at individual leaf, whole-plant, within-bog and
among-bog scales. Our studies fill a lacuna in prior studies of the Sarracenia food web (reviewed
by Ellison et al. 2003), which, with few exceptions (Judd 1959, Plummer and Kethley 1964, Fish
and Hall 1978, Bradshaw 1983, Joel and Gepstein 1985, Heard 1994b), emphasized population
dynamics and interspecific interactions only among the invertebrates and ignored the plant.
B. Human resource development
The first award was an RUI award (PI Ellison was then at Mount Holyoke College), and during
that award period we focused intensively on supporting undergraduates. The second award, both
through the core budget and through two REU supplements, also supported undergraduates. In
total, we have supported 20 undergraduates: 9 Mount Holyoke students, 7 University of Vermont
students, and REU students from University of Kansas, Haverford College, New College of Florida, and Pueblo (Colorado) Community College. Six of the Mount Holyoke students are pursuing
graduate degrees in ecology, and three work in the private sector. Two of the Vermont students
are pursuing graduate degrees in ecology, two are in the Peace Corps, and three work in the private sector. Some of the 2003 and 2004 REU students are still in college, one will be starting
graduate school in fall 2005, and two are at work as research technicians; the 2005 REU students
are still enrolled in college. To date, we have written four papers with undergraduates (Ellison
and Parker 2002, Hart et al. 2005, Butler et al. 2005, Atwater et al. 2006). With an RET supplement, we also currently work with a K-6 teacher who is introducing bog ecology to her students
in a Winchendon, MA elementary school. These awards also supported Amy Wakefield’s master’s thesis (Wakefield et al. 2005) in PI Gotelli’s lab; partially supported two Ph.D. students
(Sarah Wittman and Catherine Farrell-Gray) in Gotelli’s lab; supported a visiting post-doctoral
fellow (Jim Karagatzides) in Ellison’s lab; and supported research associates Leszek Błędzki
(1998-2002) and Jessica Butler (2002-present) in Ellison’s lab. Both research associates have
published papers from this research (Błędzki and Ellison 1998, 2002, 2003, Butler et al. 2005,
Karagatzides et al. 2005, Atwater et al. 2006, Butler and Ellison 2006). In March 2005, Jim
Karagatzides received a 2-year NSERC postdoctoral fellowship to continue his research in Ellison’s lab on cycling of organic N in the Sarracenia microecosystem.
C. Publications resulting from DEB 98-05722 / 98-08504 and DEB 02-35128 / 02-34710
To date, 30 papers (marked with an * in the Literature Cited section of the proposal) and 15
presentations at national and international meetings have resulted from these awards, including
primary research articles in Proceedings of the National Academy of Sciences (Ellison and
Gotelli 2002), Ecology (Gotelli and Ellison 2002b, 2002c, Dixon et al. 2005, Wakefield et al.
2005), Ecological Applications (Gotelli and Ellison 2005), Ecology Letters (Buckley et al. 2003),
and Oikos (Gotelli and Ellison 2002a); and review papers in Trends in Ecology and Evolution
(Ellison and Gotelli 2001), Advances in Ecological Research (Ellison et al. 2003), and Ecology
Letters (Ellison 2004). The PNAS paper was highlighted in the News and Comments section of
Trends in Ecology and Evolution (17: 305 [2002]). We also organized a well-attended symposium on carnivorous plants as model ecological systems for ESA’s annual meeting in 2000,
which resulted in the two review papers. Ellison will give a keynote address on this work in a
symposium on carnivorous plants at the International Botanical Congress in Vienna in July 2005.
During this funding period, we published a new textbook for ecologists, A Primer of Ecological
Statistics (Gotelli and Ellison 2004). Many examples of data analysis and experimental design
from our Sarracenia research are featured in this book.
C-15
Literature Cited
Addicott, J. F. 1974. Predation and prey community structure: an experiment study of the effect
of mosquito larvae on the protozoan communities of pitcher plants. Ecology 55:475-492.
Amarasekare, P., M. F. Hoopes, N. Mouquet, and M. Holyoak. 2004. Mechanisms of coexistence
in competitive metacommuniites. American Naturalist 164:310-326.
*Atwater, D. Z., J. L. Butler, and A. M. Ellison. 2006. Spatial distribution and impacts of moth
larvae on northern pitcher plants. Northeastern Naturalist 13:000-000 (in press).
*Błędzki, L. A., and A. M. Ellison. 1998. Population growth and production of Habrotrocha
rosa Donner (Rotifera: Bdelloidea) and its contribution to the nutrient supply of its host,
the northern pitcher plant, Sarracenia purpurea L. (Sarraceniaceae). Hydrobiologia
385:193-200.
* Błędzki, L. A., and A. M. Ellison. 2002. Nutrient regeneration by rotifers in New England
(USA) bogs. Verhandlung Internationale Vereinigung Limnologie 28:1328-1331.
* Błędzki, L. A., and A. M. Ellison. 2003. Diversity of rotifers from northeastern USA bogs with
new species records for North America and New England. Hydrobiologia 497:53-62.
Bradshaw, W. E. 1983. Interaction between the mosquito Wyeomyia smithii, the midge
Metriocenemus knabi, and their carnivorous host Sarracenia purpurea. Pages 161-189 in
J. H. Frank, and L. P. Lounibos, editors. Phytotelmata: terrestrial plants as host for
aquatic insect communities. Plexus Publishing Inc., New Jersey.
Bradshaw, W. E., and R. A. Creelman. 1984. Mutualism between the carnivorous purple pitcher
plant Sarracenia purpurea and its inhabitants. American Midland Naturalist 112:294304.
Bradshaw, W. E., and C. M. Holzapfel. 2001. Genetic shift in photoperiodic response correlated
with global warming. Proceedings of the National Academy of Sciences, USA 98:1450914511.
*Buckley, H. L., T. E. Miller, A. M. Ellison, and N. J. Gotelli. 2003. Reverse latitudinal trends in
species richness of pitcher-plant food webs. Ecology Letters 6:825-829.
*Butler, J. L., and A. M. Ellison. 2006. Nitrogen cycling dynamics in the northern pitcher plant,
Sarracenia purpurea. To be submitted to Journal of Ecology.
*Butler, J. L., D. Z. Atwater, and A. M. Ellison. 2005. Red-spotted newts: an unusual nutrient
source for northern pitcher plants. Northeastern Naturalist 12:1-10.
*Butler, J. L., A. M. Ellison, and N. J. Gotelli. 2005. Cycling of inorganic nitrogen through the
pitcher plant food web. Abstract. 16th annual Harvard Forest Ecology Symposium,
Petersham, Massachusetts.
Cadenasso, M. L., S. T. A. Pickett, and K. C. Weathers. 2005. Effect of landscape boundaries on
the flux of nutrients, detritus, and organisms. Pages 154-168 in G. A. Polis, M. E. Power,
and G. R. Huxel, editors. Food webs at the landscape level. University of Chicago Press,
Chicago, IL.
Chapin, C. T., and J. Pastor. 1995. Nutrient limitation in the northern pitcher plant Sarracenia
purpurea. Canadian Journal of Botany 73:728-734.
Clark, J. S. 1991a. Disturbance and population structure on the shifting mosaic landscape.
Ecology 72:1119-1137.
Clark, J. S. 1991b. Disurbance and tree life history on the shifting mosaic landscape. Ecology
72:1102-1118.
D-1
Cottingham, K. L., J. T. Lennon, and B. L. Brown. 2005. Knowing when to draw the line:
designing more informative ecological experiments. Frontiers in Ecology and the
Environment 3:145-152.
Dixon, P. 1994. Testing spatial segregation using a nearest-neighbor contingency tables. Ecology
75:1940-1948.
*Dixon, P. M., A. M. Ellison, and N. J. Gotelli. 2005. Improving the precision of estimates of the
frequency of rare events. Ecology 86:1114-1123.
*Ellison, A. M. 2001. Interspecific and intraspecific variation in seed size and germination
requirements of Sarracenia (Sarraceniaceae). American Journal of Botany 88:429-437.
*Ellison, A. M. 2002. Food for thought: a review of recent research on pitcher-plant bogs in New
England. Conservation Perspectives.
http://www.MassSCB.org/epublications/summer2002/.
*Ellison, A. M. 2004. Bayesian inference for ecologists. Ecology Letters 7:509-520.
Ellison, A. M., and B. L. Bedford. 1995. Response of a wetland vascular plant community to
disturbance: a simulation study. Ecological Applications 5:109-123.
*Ellison, A. M. and E. J. Farnsworth. 2005. The cost of carnivory for Darlingtonia californica
(Sarraceniaceae): evidence from relationships among leaf traits. American Journal of
Botany 92:000-000 (in press).
*Ellison, A. M., and N. J. Gotelli. 2001. Evolutionary ecology of carnivorous plants. Trends in
Ecology and Evolution 16:623-629.
*Ellison, A. M., and N. J. Gotelli. 2002. Nitrogen availability alters the expression of carnivory
in the northern pitcher plant Sarracenia purpurea. Proceedings of the National Academy
of Sciences, USA 99:4409-4412.
*Ellison, A. M. and N. J. Gotelli. in prep. a. Community assembly in a dynamic habitat. To be
submitted to The American Naturalist.
*Ellison, A. M. and N. J. Gotelli. in prep. b. Demography of the northern pitcher plant
Sarracenia purpurea: population dynamics from individual trait-based models. To be
submitted to Journal of Ecology.
*Ellison, A. M. and J. N. Parker. 2002. Seed dispersal and seedling establishment of Sarracenia
purpurea (Sarraceniaceae). American Journal of Botany 89:1024-1026.
*Ellison, A. M., E. J. Farnsworth and N. J. Gotelli. 2002. Ant diversity in pitcher-plant bogs of
Massachusetts. Northeastern Naturalist 9:267-284.
*Ellison, A. M., H. L. Buckley, T. E. Miller, and N. J. Gotelli. 2004. Morphological variation in
Sarracenia purpurea (Sarraceniaceae): geographic, environmental, and taxonomic
correlates. American Journal of Botany 91:1930-1935.
*Ellison, A. M., N. J. Gotelli, L. A. Błędzki, and J. L. Butler. 2005. Regulation of food-web
structure in the Sarracenia microecosystem. Poster presented at the 2005 Annual Meeting
of the Ecological Society of America.
*Ellison, A. M., N. J. Gotelli, J. S. Brewer, D. L. Cochran-Stafira, J. Kneitel, T. E. Miller, A. C.
Worley, and R. Zamora. 2003. The evolutionary ecology of carnivorous plants. Advances
in Ecological Research 33:1-74.
*Farrell-Gray, C. C. and N. J. Gotelli. 2005. Allometric exponents support a 3/4 power scaling
law. Ecology 86:0000-0000 (in press).
Fish, D. 1976. Insect-plant relationships of the insectivorous pitcher plant Sarracenia minor. The
Florida Entomologist 59:199-203.
D-2
Fish, D., and D. W. Hall. 1978. Succession and stratification of aquatic insects inhabiting the
leaves of the insectivorous pitcher plant Saracenia purpurea. American Midland
Naturalist 99:172-183.
Folkerts, D. R., and G. W. Folkerts. 1996. Aids for field identification of pitcher plant moths of
the genus Exyra (Lepidoptera: Noctuidae). Entomological News 107:128-136.
Forman R. T. T., and M. Godron. 1986. Landscape ecology. John Wiley and Sons, New York,
NY.
Francoeur, A. 1973. Révision taxonomique des espéces néarctiques du groupe Fusca, genre
Formica (Formicidae, Hymenoptera). Memoires de la Société Entomologique du Québec
3:316.
Francoeur, A. 1997. Ants (Hymenoptera: Formicidae) of the Yukon. Pages 901-910 in H. V.
Danks, and J. A. Downes, editors. Insects of the Yukon. Biological Survey of Canada
(Terrestrial Arthropods), Ottawa, Ontario, Canada.
Fuhlendorf, S. D., and D. M. Engle. 2004. Application of the fire-grazing interaction to restore a
shifting mosaic on tallgrass prairie. Journal of Applied Ecology 41:604-614.
*Gotelli, N. J., and A. M. Ellison. 2002a. Assembly rules for New England ant assemblages.
Oikos 99:591-599.
*Gotelli, N. J., and A. M. Ellison. 2002b. Biogeography at a regional scale: determinants of ant
species density in bogs and forests of New England. Ecology 83:1604-1609.
*Gotelli, N. J., and A. M. Ellison. 2002c. Nitrogen deposition and extinction risk in the northen
pitcher plant Sarracenia purpurea. Ecology 83:2758-2765.
*Gotelli N. J., and A. M. Ellison. 2004. A primer of ecological statistics. Sinauer Associates,
Sunderland, Massachusetts, USA.
*Gotelli, N. J., and A. M. Ellison. 2005. Forecasting extinction risk with non-stationary matrix
models. Ecological Applications 15:000-000 (in press).
Güsewell, S. 2005. High nitrogen:phosphorus ratios reduce nutrient retention and second-year
growth of wetland sedges. New Phytologist 166:537-550.
Hanski, I. 1985. Single-species spatial dynamics may contribute to long-term rarity and
commonness. Ecology 66:335-343.
Harkness, R. D., and V. Isham. 1983. A bivariate spatial point pattern of ants' nests. Applied
Statistics 32:293-303.
*Hart, C. M., C. Ordoyne, and A. M. Ellison. in prep. Competition for prey between spiders and
pitcher plants. To be submitted to Ecology.
Harvey, E., and T. E. Miller. 1996. Variance in composition of inquiline communities in leaves
of Sarracenia purpurea L on multiple spatial scales. Oecologia 108:562-566.
Heard, S. B. 1994a. Imperfect oviposition decisions by the pitcher plant mosquito (Wyeomyia
smithii). Evolutionary Ecology 8:493-502.
Heard, S. B. 1994b. Pitcher plant midges and mosquitoes: a processing chain commensalism.
Ecology 75:1647-1660.
Heard, S. B. 1994c. Wind exposure and distribution of pitcher-plant mosquito (Diptera:
Culicidae). Environmental Entomology 23:1250-1253.
Holt, R. D. 2005. Implications of system openness for local community structure and ecosystem
function. Pages 96-114 in G. A. Polis, M. E. Power, and G. R. Huxel, editors. Food webs
at the landscape level. University of Chicago Press, Chicago, IL.
D-3
Horn, H. S., and R. H. MacArthur. 1972. Competition among fugitive species in a harlequin
environment. Ecology 53:749-752.
Hutchinson, G. E. 1951. Copepodology for the ornithologist. Ecology 32:571-577.
Istock, C. A., K. Tanner, and H. Zimmer. 1983. Habitat selection by the pitcher-plant mosquito,
Wyeomyia smithii: behavioral and genetic aspects. Pages 191-204 in J. H. Frank, and L.
P. Lounibos, editors. Phytotelmata: terrestrial plants as hosts for aquatic insect
communities. Plexus Publishing, Inc., Medford, N.J.
Joel, D. M., and S. Gepstein. 1985. Chloroplasts in the epidermis of Sarracenia (the American
pitcher plant) and their possible role in carnivory - an immunocytochemical approach.
Physiologia Plantarum 63:71-75.
Jones, F. M. 1921. Pitcher plants and their moths. Natural History 21:296-316.
Judd, W. W. 1959. Studies of the Byron Bog in southwestern Ontario. X. Inquilines and victims
of the pitcher plant, Saracenia purpurea L. Canadian Entomologist 91:171-180.
*Karagatzides, J. D., J. L. Butler, and A. M. Ellison. 2005. Sarracenia can directly acquire
organic nitrogen and short-circuit the inorganic nitrogen cycle. Functional Ecology in
review.
Kennedy, G. G., and N. P. Storer. 2000. Life systems of polyphagous arthropod pests in
temporally unstable cropping systems. Annual Review of Entomology 45:467-493.
Kneitel, J. M., and T. E. Miller. 2002. Resource and top-predator regulation in the pitcher plant
(Sarracenia purpurea) inquiline community. Ecology 83:680-688.
Kneitel, J. M., and T. E. Miller. 2003. Dispersal rates affect species composition in
metacommunities of Sarracenia purpurea inquilines. American Naturalist 162:165-171.
Krawchuk, M. A., and P. D. Taylor. 1999. Roosting behaviour by Fletcherimyia fletcheri
(Diptera: Sarcophagidae) in Sarracenia purpurea (Sarraceniaceae). Canadian
Entomologist 131:829-830.
Krawchuk, M. A., and P. D. Taylor. 2003. Changing importance of habitat structure across
multiple spatial scales for three species of insects. Oikos 103:153-161.
Leibold, M. A., M. Holyoak, N. Mouquet, P. Amarasekare, J. M. Chase, M. F. Hoopes, R. D.
Holt, J. B. Shurin, R. Law, D. Tilman, M. Loreau, and A. Gonzalez. 2004. The
metacommunity concept: a framework for multi-scale community ecology. Ecology
Letters 7:601-613.
Miller, T. E., and J. M. Kneitel. 2005. Inquiline communities in pitcher plants as a prototypical
metacommunity. Page in press in M. Holyoak, R. Holt, and M. Leibold, editors.
Metacommunities. University of Chicago Press, Chicago, IL.
Miner, J. A., and P. D. Taylor. 2002. Effect of peatland size and exposure on two species of
Diptera inhabiting the pitcher plant Sarracenia purpurea L. Écoscience 9:347-354.
Mogi, M., and J. Mokry. 1980. Distribution of Wyeomyia smithii (Diptera, Culicidae) eggs in
pitcher plants in Newfoundland, Canada. Tropical Medicine 22:1-12.
*Morales, S.E., P. J. Mouser, N. Ward, S. P. Hudman, N. J. Gotelli, D. S. Ross, and T. A. Lewis.
2005. Comparison of bacterial communities in New England Sphagnum bogs using
terminal restriction fragment length polymorphism (T-RFLP). Microbial Ecology (in
press).
Mouquet, N., and M. Loreau. 2002. Coexistence in metacommunities: the regional similarity
hypothesis. American Naturalist 159:420-426.
D-4
Newell, S. J., and A. J. Nastase. 1998. Efficiency of insect capture by Sarracenia purpurea
(Sarraceniaceae), the northern pitcher plant. American Journal of Botany 85:88-91.
Olff, H., B. Hoorens, R. G. M. de Goede, W. H. Van der Putten, and J. M. Gleichman. 2000.
Small-scale shifting mosaics of two dominant grassland species: the possible role of soilborne pathogens. Oecologia 125:45-54.
Perry, J. N. 1998. Measures of spatial pattern for counts. Ecology 79:1008-1017.
Perry, J. N., L. Winder, J. M. Holland, and R. D. Alston. 1999. Red-blue plots for detecting
clusters in count data. Ecology Letters 2:106-113.
Plummer, G., and J. B. Kethley. 1964. Foliar absorption of amino acids, peptides, and other
nutrients by the pitcher plant, Sarracenia flava. Botanical Gazette 125:245-260.
Polis G. A., M. E. Power, and G. R. Huxel. 2005. Food webs at the landscape level. University
of Chicago Press, Chicago, IL.
Polis, G. A., W. B. Anderson, and R. D. Holt. 1997. Toward an integration of landscape and
foodweb ecology: the dynamics of spatially subsidized food webs. Annual Review of
Ecology and Systematics 28:289-316.
Ricketts, T. H. 2001. The matrix matters: effective isolation in fragmented landscapes. American
Naturalist 158:87-99.
Ripley, B. D. 1976. The second-order analysis of stationary point processes. Journal of Applied
Probability 13:255-266.
Rymal, D. E., and G. W. Folkerts. 1982. Insects associated with pitcher plants (Sarracenia:
Sarraceniaceae), and their relationship to pitcher plant conservation: a review. Journal of
the Alabama Academy of Science 53:131-151.
Skellam, J. G. 1951. Random dispersal in theoretical populations. Biometrika 38:196-218.
Srivastava, D. S., J. Kolasa, J. Bengtsson, A. Gonzalez, S. P. Lawler, T. E. Miller, P. Munguia,
T. Romanuk, D. C. Schneider, and M. K. Trzcinski. 2004. Are natural microcosms useful
model systems for ecology? Trends in Ecology and Evolution 19:379-384.
Stewart-Oaten, A., and J. R. Bence. 2001. Temporal and spatial variation in environmentental
impact assessment. Ecological Monographs 71:305-339.
Stewart-Oaten, A., J. R. Bence, and C. W. Osenberg. 1992. Assessing effects of unreplicated
perturbations: no simple solutions. Ecology 73:1396-1404.
Trzcinski, M. K., S. J. Walde, and P. D. Taylor. 2003. Colonisation of pitcher plant leaves a
several spatial scales. Ecological Entomology 28:482-489.
Trzcinski, M. K., S. J. Walde, and P. D. Taylor. 2005. Stability of pitcher-plant microfaunal
populations depends on food web structure. Oikos 110:146-154.
Underwood, A. J. 1994. On beyond BACI: sampling designs that might reliably detect
environmental disturbances. Ecological Applications 4:3-15.
*Wakefield, A. E., N. J. Gotelli, S. E. Wittman, and A. M. Ellison. 2005. The effect of prey
addition on nutrient stoichiometry, nutrient limitation, and morphology of the carnivorous
plant Sarracenia purpurea (Sarraceniaceae). Ecology 86:1737-1743.
Winder, L., C. J. Alexander, J. M. Holland, C. Woolley, and J. N. Perry. 2001. Modeling the
dynamic spatio-temporal response of predators to transient prey patches in the field.
Ecology Letters 4:568-576.
Zavala, M. A., and E. Zea. 2004. Mechanisms maintaining biodiversity in Mediterranean pineoak forests: insights from a spatial simulation model. Plant Ecology 171:197-207.
D-5