ecosystem size, but not disturbance, determines food …

Ecology, 89(11), 2008, pp. 3001–3007� 2008 by the Ecological Society of America

ECOSYSTEM SIZE, BUT NOT DISTURBANCE, DETERMINES FOOD-CHAINLENGTH ON ISLANDS OF THE BAHAMAS

GAKU TAKIMOTO,1,3,4 DAVID A. SPILLER,2 AND DAVID M. POST3

1Department of Biology, Faculty of Science, Toho University, Funabashi, Chiba 274-8510 Japan2Section of Evolution and Ecology, One Shields Avenue, University of California, Davis, California 95616-8755 USA

3Department of Ecology and Evolutionary Biology, Yale University, New Haven, Connecticut 06511 USA

Abstract. Ecologists have long struggled to explain variation in food-chain length amongnatural ecosystems. Food-chain length is predicted to be shorter in ecosystems subjected togreater disturbance because longer chains are theoretically less resilient to perturbation.Moreover, food-chain length is expected to be longer in larger ecosystems because increasingecosystem size increases species richness and stabilizes predator–prey interactions, or increasestotal resource availability. Here we test the roles of disturbance and ecosystem size indetermining the food-chain length of terrestrial food webs on Bahamian islands. We foundthat disturbance affected the identity of top predators, but did not change food-chain lengthbecause alternative top predators occupied similar trophic positions. On the other hand, a 106-fold increase in ecosystem size elevated food-chain length by one trophic level. We suggest thatthe effect of disturbance on food-chain length is weak when alternate top predators are trophicomnivores and have similar trophic positions. This and previous work in lakes suggest thatecosystem size may be a strong determinant of food-chain length in both aquatic andterrestrial ecosystems.

Key words: Anolis lizards; Bahamian islands; disturbance; ecosystem size; food-chain length; food web;omnivory; orb spiders; predation; stable isotope; trophic position.

INTRODUCTION

Food-chain length is an important characteristic of

biological communities. Variation in food-chain length

influences primary and secondary productivity (Carpen-

ter and Kitchell 1993), nutrient cycling and community

stability (DeAngelis et al. 1989), and the bioaccumula-

tion of contaminants in top predators including many

consumed by humans (Cabana and Rasmussen 1994,

Kidd et al. 1998). Ecologists have proposed several

hypotheses to explain natural variation in food-chain

length (Pimm 1982, Lawton 1989, Post 2002a), but no

clear answer has been reached. Among potential

determinants in previous hypotheses, disturbance and

ecosystem size have attracted considerable recent

attention (Post 2002a). In the dynamic constraints

hypothesis, more frequent or more intense disturbance

is predicted to shorten food-chain length because simple

theoretical models suggest that longer chains are less

resilient, and thus unlikely to persist in disturbed

habitats (Pimm and Lawton 1977). Except for a few

experiments that lend indirect support (Jenkins et al.

1992, Lawler and Morin 1993), there is little direct

evidence for an effect of disturbance on food-chain

length in natural systems (Post 2002a). The ecosystem

size hypothesis (Post et al. 2000) predicts that food-chain

length will be longer in larger ecosystems because larger

space can retain more individuals and thus more species

(Cohen and Newman 1991), foster metapopulation

persistence of higher trophic levels (Holt 1996), and

stabilize predator–prey interactions (Spencer and War-

ren 1996, Holt 2002). The productive space hypothesis

also predicts that increasing ecosystem size will increase

food-chain length because larger ecosystems provide

greater total primary production (local primary produc-

tion 3 ecosystem size) that can support longer food

chains (Schoener 1989). Several studies have found an

effect of ecosystem size on food-chain length in aquatic

food webs (Spencer and Warren 1996, Post et al. 2000,

Thompson and Townsend 2005), but there exists only

indirect evidence from terrestrial food webs that higher-

order predators tended to occur on larger ecosystems

(Schoener and Schoener 1983a, Schoener 1989).

Trophic omnivory, defined as feeding on more than

one trophic level, can have profound influences on food-

chain length. Growing empirical evidence suggests that

both the presence/absence of predators on the top of

food webs and shifts in the top predator’s trophic

position are fundamental causes for natural variation in

food-chain length (Post et al. 2000, Post and Takimoto

2007). Trophic position is a continuous index averaging

the number of trophic levels from basal species to a

consumer through all trophic pathways. When top

predators are trophic omnivores, changing their primary

diets from lower trophic levels to higher trophic levels

will increase their trophic position, leading to an

increase in food-chain length (Diehl and Feissel 2001,

Manuscript received 4 December 2007; revised 15 May 2008;accepted 28 May 2008. Corresponding Editor: B. P. Kotler.

4 E-mail: [email protected]

3001

Rep

orts

Post and Takimoto 2007). Moreover, adding higher-

order predators to the top of food webs does notnecessarily cause an increase in food-chain length, if they

are trophic omnivores feeding on multiple lower trophiclevels.

Here we test the effects of disturbance and ecosystemsize on food-chain length of terrestrial food webs found

on Bahamian islands. Storm surges that inundate smallislands, typically caused by hurricanes and winterstorms, have profound influences on food-web structure

and dynamics because they extirpate spider and lizardpopulations (Spiller et al. 1998), and alter plant–animal

interactions (Spiller and Agrawal 2003). Ecosystem size,in this case island area, can strongly affect species

membership of insular food webs (Schoener andSchoener 1983b) and modify the strength of local species

interactions (Schoener and Spiller 1987). These obser-vations suggest that both disturbance and ecosystem size

may jointly affect food-chain length on islands in theBahamas.

The top of the food webs found on small- to medium-sized islands are occupied by orb spiders and Anolis

lizards (see Plate 1), both omnivores that prey uponherbivorous and predatory arthropods (Schoener 1989).

Anolis lizards also prey upon orb spiders. Here we definefood-chain length as the trophic position of the apical

predator (i.e., top predator with the maximum trophicposition; Post and Takimoto 2007). The trophic

positions of spiders and lizards were estimated by stableisotope techniques, which provide an energy-flow-weighted measure of trophic position (Post 2002b). We

also surveyed visually the identity of the top predatorson each island, and evaluated the effects of disturbance

and ecosystem size on the incidence of Anolis lizards.

METHODS

Study region and environmental gradients

Our study islands were located around Great Exuma,

Bahamas (238300 N, 758460 W; Fig. 1). The effect ofdisturbance was compared between islands exposed to

and protected from storm surges. Protected islands werein the waters surrounded by the three largest islands:Great Exuma, Little Exuma, and Stocking. Exposed

islands were in the open waters to the southwest ofGreat Exuma (Spiller et al. 1998, Spiller and Agrawal

2003). The effect of ecosystem size was examined acrossislands of various size ranging from 5.0 3 102 to 3.0 3

108 m2. This range is comparable to the range ofecosystem size along which a significant increase in food-

chain length was detected in North American lakes(Vander Zanden et al. 1999, Post et al. 2000). There was

no significant difference in ecosystem size betweenexposed and protected islands (two-sided, two-sample

Kolmogorov-Smirnov test, D¼ 0.35, P ¼ 0.26).

Stable isotope sampling

Lizards and spiders for isotope samples were collected

on 19 exposed, 14 protected, and the three largest

islands. We collected spiders and lizards from each

island during one-week sampling sessions in December

2003, May and October 2004, July and October 2005,

and October 2006. Some islands were visited more than

once. On each island, we sampled terrestrial plants,

spiders, and lizards in an area devoid of C4 plants, which

were usually rare on the study islands. For small islands,

the sampling area was typically the entire island. For

islands larger than 10 000 m2, we sampled areas .20 m

inland from the shore. All of the sampled lizards were

the most common species in the Bahamas, Anolis sagrei.

Sampled orb spider species included the common

species: Argiope argentata, Gasteracantha cancriformis,

Metepeira datona, and Nephila clavipes. We collected the

whole bodies of spiders (3–15 individuals per island,

median¼ 5) and tail tips (;2 cm) of adult lizards (2–10

individuals per island, median ¼ 5). Lizards were

released on site immediately after they were handled.

We found no significant difference in d15N or d13Cbetween the tail and whole body of lizards (two-tailed

FIG. 1. Top: Map of Exuma, Bahamas showing the threelargest study islands (1, Great Exuma; 2, Little Exuma; 3,Stocking). Bottom: Close-up of the boxed area of the top panel,showing the 33 smaller islands used in the food-chain-lengthanalysis. Islands are numbered consecutively from the largest tothe smallest.

GAKU TAKIMOTO ET AL.3002 Ecology, Vol. 89, No. 11R

epor

ts

paired t test: t9¼ 0.30 and P¼ 0.77 for d15N; t9¼ 1.2 and

P ¼ 0.28 for d13C).Because lizards and spiders on our study islands can

receive a significant amount of carbon and nitrogen

from both terrestrial and marine sources, we sampled

terrestrial plant leaves for the terrestrial isotopic

baseline, and marine-derived detritus (particulate or-

ganic matter, POM) for the marine isotopic baseline, of

our study food webs. Isotopic baselines are the isotopic

signatures of producers (or primary consumers) to

account for spatial and temporal variation in isotopic

signatures at the base of food chains when calculating

the trophic positions of secondary consumers (Post

2002b). For the terrestrial isotopic baseline, we haphaz-

ardly sampled plant leaves within a 20 m radius from the

area where spiders and lizards were collected. We

decided that 20 m was an appropriate scale for capturing

spatial variation in the isotope signature of terrestrial

plants because orb spiders are sedentary and lizard

territories average ;10 m2 (Schoener and Schoener

1982). From each of five plants on small islands and 10

plants on large islands, we collected and combined five

leaves into a single isotope sample. Major plant species

included Borrichia arborescens, Coccoloba uvifera, Coc-

coloba diversifolia, Conocarpus erectus, and Pithecel-

lobium keyense. Marine POM was sampled at two

locations in the low-littoral zone of each island by

filtrating up to 6 L of seawater through 1-lm mesh glass

fiber filters. Because marine POM collected in this way

provides a short temporal snapshot of the marine

isotopic baseline, we averaged the d15N and d13C of

marine POM across all sampling sessions to obtain

d15Nmar and d13Cmar. Macro green algae (probably

Neomeris or Batophora) in the low-littoral zone of each

island could also be a source of marine-derived material

to spiders and lizards, but using green algae as the

marine isotopic baseline did not change our conclusions

(see Appendix).

Sample preparation and stable isotope measurements

All samples were dried at 558C for 48 h. Plant leaves,

spider bodies, and lizard tails were ground to fine

powder. The POM filters were cut into halves; one-half

was used to estimate d15N and the other was used to

estimate d13C after fumigation with hydrochloric acid to

eliminate carbonates. Isotope analysis was performed

with a Delta Plus Advantage stable isotope mass

spectrometer (Thermo Electron, Bremen, Germany) at

Earth Systems Center for Stable Isotope Studies

(ESCSIS) at Yale Institute for Biospheric Studies. The

standard deviation of replicates of ESCSIS animal

standard was 0.14ø for d13C and 0.22ø for d15N,

and of ESCSIS plant standard 0.06ø for d13C and

0.15ø for d15N. All stable-isotope values were reported

in the d notation: d13C or d15N ¼ ([Xsam/Xstd] � 1) 3

1000, where X is 13C/12C or 15N/14N. d13C values of

spider bodies and lizard tails were corrected for lipids

following Post et al. (2007) using the relationship

PLATE 1. Top predators on small to medium-sized Bahamian islands. (Left) Argiope argentata (orb spider). (Right) Anolissagrei (lizard). Photo credits: G. Takimoto (left) and D. A. Spiller (right).

November 2008 3003DETERMINANTS OF FOOD-CHAIN LENGTHR

eports

between C:N ratio and lipid-caused d13C change

established for our study organisms (d13Ccor¼ d13Cuncor

þ 0.99 3 [C/N ratio] � 2.5).

Estimates of food-chain length and statistical analysis

We estimated food-chain length as the trophic

position of the apical predator (Post et al. 2000, Post

and Takimoto 2007). Trophic position is calculated as kþ (d15Norg � d15Nbase)/DN, where k is the trophic

position of the organism used to estimate d15Nbase (k ¼1 in our case), d15Norg is measured directly, and DN is theaverage trophic fractionation in d15N per trophic level.

For all but the largest islands, we estimated the trophic

position of each spider and lizard using a two-end-

member mixing model for isotopic-baseline correction

(Post 2002b). For the three largest islands, spiders and

lizards were collected from the center of the island where

marine inputs were minimal; therefore, we used terres-trial plants as the single isotopic baseline for calculating

trophic position. We employed as the nominal values the

standard trophic fractionation of 3.4ø for d15N and 0øfor d13C, and also confirmed that deviation from these

values did not alter our conclusion (Appendix). All

statistical analyses were performed in R version 2.4.1 (RDevelopment Core Team 2006).

RESULTS

We found a strong significant positive relationship

between ecosystem size and food-chain length among

protected and exposed islands without the largest islands(ANCOVA, F1,31¼ 20.1, P , 10�4, food-chain length¼0.270 log(area)þ 0.831, R2¼ 0.39; plotting residuals vs.

fitted values showed no trend), but no effect of

disturbance on the intercept (F1,30 ¼ 0.80, P ¼ 0.38) or

the slope (the ecosystem size 3 disturbance interaction,

F1,29 ¼ 0.82, P ¼ 0.37) of the relationship between

ecosystem size and food-chain length. Because distur-

bance had no effect, we included data from the three

largest islands to further investigate the relationship

between food-chain length and ecosystem size. The

relationship for all islands combined was best represent-

ed by the asymptotic model (2.99 [1 � exp(�0.26 log

area)]; Fig. 2), which explained 55% of the variation in

food-chain length (R2 ¼ 0.55). Food-chain length

increased by about one trophic level from the smallest

to the largest islands.

Both disturbance and ecosystem size influenced the

identity of the apical predators. The incidence of Anolis

lizards increased with ecosystem size (Fig. 3; GLM with

binomial errors and logit link, v21 ¼ 34.8, P , 10�8), and

was lower on exposed than protected islands (v21 ¼ 11.1,

P ¼ 0.001; logit incidence ¼ �17.5 þ 4.6 log area for

exposed islands, and logit incidence ¼�14.2 þ 4.6 log

area for protected islands). The interaction between

disturbance and ecosystems size was not significant (v21¼

0.18, P¼ 0.67). Orb spiders were found on most islands

without Anolis lizards.

Changes in the identity of the apical predators did not

translate into the shifts of food-chain length with

ecosystem size. ANCOVA with apical predator (lizard

vs. spider) as a main effect and ecosystem size as a

covariate found no evidence that difference in the apical

predators changed food-chain length (apical predator

effect, F1,30 ¼ 2.34, P ¼ 0.14; ecosystem size 3 apical

predator interaction, F1,29 ¼ 0.0050, P ¼ 0.94).

DISCUSSION

Disturbance and ecosystem size have been hypothe-

sized as major determinants of food-chain length (Pimm

1982, Lawton 1989, Post 2002a), but no study to our

knowledge has tested their simultaneous effects on food-

FIG. 2. Relationship between food-chain length and eco-system size. Lizards and spiders occupied similar high trophicpositions on each of the largest islands where they coexisted.

FIG. 3. The response of Anolis lizard incidence to distur-bance and ecosystem size. The data are from visual surveys on37 exposed and 16 protected islands including those used forestimating food-chain length.


epor

ts

chain length in natural ecosystems. While disturbance

was shown to strongly affect food-web structure and

dynamics on small islands of the Bahamas (Spiller and

Agrawal 2003), we found no evidence that disturbance

influenced food-chain length. On the other hand, the

effect of ecosystem size on food-chain length was highly

significant. Our results support the hypothesis that

ecosystem size, but not disturbance, affects food-chain

length on these islands.

Why did we not find an effect of disturbance on food-

chain length? We did find that disturbance affected the

identity of the apical predators; orb spiders were the

apical predators on small exposed islands and Anolis

lizards were the apical predator on small protected

islands. However, this difference in the identity of the

apical predators did not change food-chain length;

lizards on small protected islands occupied a low trophic

position similar to that of orb spiders on small exposed

islands (Fig. 2). Previous observational and experimen-

tal studies on Bahamian islands suggest that the

coexistence of orb spiders and Anolis lizards is difficult

on small islands (Schoener and Toft 1983, Schoener and

Spiller 1996, Spiller and Schoener 1998), and in this

study we rarely observed orb spiders on small protected

islands that had Anolis lizards (D. A. Spiller and G.

Takimoto, personal observation). The pattern on small

islands is caused by very intense lizard predation on orb

spiders. Field experiments revealed that lizard predation

on orb spiders was devastating on small islands

(Schoener and Spiller 1996), but weaker on a large

island (Spiller and Schoener 1998). This suggests that

strong predation by lizards extirpated or drastically

reduced orb spiders and other predatory arthropods on

small protected islands, shifting the diet of lizards

toward organisms at lower trophic levels such as

herbivorous arthropods, and effectively lowering the

trophic position of lizards. Therefore, the presence of

lizard populations on small protected islands did not

cause an increase in food-chain length. On the other

hand, an increase in ecosystem size promotes coexistence

of lizards and predatory arthropods. This allows lizards

to feed on predatory arthropods on larger islands,

increasing the trophic position of lizards, and thus

promoting longer food chains.

We note that our exposed and protected islands were

spatially clustered, which might have confounded the

effect of disturbance. Apart from the levels of distur-

bance, the protected and exposed islands may differ in at

least two other respects (Spiller and Agrawal 2003).

First, the exposed islands are generally more distant

from the largest island, Great Exuma (Fig. 1). This

might result in some predators being absent from

exposed islands and shorten food-chain length, poten-

tially exaggerating the expected negative effect of

disturbance. Yet this seems to have had little influence

given that we found no disturbance effect. Second,

exposed islands might receive higher rainfall when the

predominant eastern trade winds blow across the

warmer land mass of Great Exuma, causing rainfall on

these islands but not on the protected ones. This could

counter the expected disturbance effect by enriching and

therefore lengthening food chains on exposed islands in

accordance with the productive space hypothesis. The

second possibility might partially explain the lack of a

disturbance effect in our system.

This study provides the first quantitative evidence for

the effect of ecosystem size on food-chain length in a

terrestrial ecosystem. Our finding is consistent with

related evidence that food-chain length was truncated in

fragmented landscapes (Komonen et al. 2000, Kruess

and Tscharntke 2000), although fragmentation may not

always reduce food-chain length (Mikkelson 1993).

Beyond terrestrial systems, our result parallels the

finding from North American lakes in which increases

in food-chain length strongly correlated with increases in

ecosystem size (Vander Zanden et al. 1999, Post et al.

2000). On the other hand, a recent meta-analysis of

aquatic food-chain length across the globe did not

generally find evidence for the relationship between

ecosystem size and food-chain length (Vander Zanden

and Fetzer 2007). Arguably, biogeographic and coevo-

lutionary histories play important roles in determining

food-chain length (Kitching 2000, Post 2002a). Our

results and that of others suggest that ecosystem size is

an important determinant of food-chain length for food

webs sharing the same regional species pool.

The ecosystem size hypothesis (Post et al. 2000)

suggests that an increase in ecosystem size enhances

species richness (Cohen and Newman 1991), stabilizes

predator–prey interactions (Spencer and Warren 1996),

and fosters metapopulation persistence of higher trophic

levels (Holt 1996); all of these favor a longer food chain.

Our results suggest that, unlike the original suggestion

of disturbance by Pimm and Lawton (1977), ecosystem

size is a rather important driver that determines the

effects of dynamic constraints on food-chain length.

Large islands provide spiders and other carnivorous

arthropods with spatial refuges from lizard predation

where lizards are temporally or permanently absent. It is

also possible that the species–area relationship (Connor

and McCoy 1979, Rosenzweig 1995) dictates higher

species richness on larger islands; this may provide

alternative prey for lizards and carnivorous arthropods,

potentially stabilizing intraguild predation (Daugherty

et al. 2007, Holt and Huxel 2007, Namba et al. 2008).

These suggest that ecosystem size may act to alleviate

dynamic constraints in our system. Although the

productive space hypothesis, emphasizing the effect of

resource availability, predicts longer chains in larger

ecosystems (Schoener 1989), our study was not designed

to test between the ecosystem size vs. productive space

hypotheses because we did not measure primary

production (such a test would require independent

gradients of ecosystem size and local productivity as in

Post et al. [2000]).


eports

Changes in food-chain length on our islands were

driven primarily by shifts in the trophic position of top

predators rather than differences in the identity of top

predators, suggesting that structural changes at the

middle of food webs are important. Shifts in the top

predator’s trophic position can occur either by changing

the proportions in their diet from different trophic levels

(without changing the number of species in food webs),

or by inserting new intermediate predators (Post et al.

2000, Post and Takimoto 2007). Both mechanisms may

be at play on our islands. Anolis lizards feed on a wide

trophic spectrum (Schoener 1968), possibly changing the

degree of trophic omnivory as ecosystem size changes.

However, at least on the largest islands where lizards

and orb spiders coexisted, lizards and spiders showing

similar trophic positions (Fig. 2) suggest that the

incidence of spiders in lizard diets was not large enough

to make their trophic position higher than that for

spiders. On the other hand, larger islands tend to

contain a greater number of intermediate predators such

as parasitoids (Schoener et al. 1995) and other carniv-

orous arthropods (M. Murakami and T. Hirao, un-

published data), as is expected from the species–area

relationship. These would insert new intermediate

trophic levels, increasing the top predator’s trophic

position. Altogether, the combination of opportunistic

feeding by lizards and spiders and higher prey diversity

following from the species–area relationship appears to

play a large role in increasing their trophic position with

increasing ecosystem size. Unlike our results, both the

trophic position and alternation of top predators are

similarly important in the North American lakes (Post et

al. 2000). For the Bahamian islands, Schoener (1989)

suggested that in addition to spiders and Anolis lizards,

higher-order predators, such as birds, snakes, and larger

lizards capable of preying upon Anolis lizards, might

elevate food-chain length on larger islands. Whether the

addition of these higher-order predators can actually

increase food-chain length on the Bahamian islands will

depend on their trophic position.

In conclusion, we suggest that a simple relationship

between disturbance and food-chain length may be

obscured by complex responses of natural food webs

with trophic omnivory to perturbations (Menge and

Sutherland 1987, Power et al. 1996, Post 2002a). On the

other hand, a shift in the diet of omnivorous top

predators can be a major structural mechanism causing

a positive effect of ecosystem size on food-chain length.

Our results support the growing evidence for a positive

effect of ecosystem size on food-chain length in aquatic

ecosystems (Spencer and Warren 1996, Vander Zanden

et al. 1999, Post et al. 2000, Thompson and Townsend

2005) and now in a terrestrial ecosystem.

ACKNOWLEDGMENTS

We thank M. Hasegawa, M. Hori, R. Holt, M. Kagami, J.Piovia-Scott, M. Spencer, A. Wright, L. Yang, and ananonymous reviewer for helpful comments on the manuscript.

Assistance in the field or in the laboratory by G. Olack, H.Kurokawa, and M. Spiller was invaluable. This work wassupported by Japan Society for the Promotion of Science(KAKENHI 03J09516 and 19870022) and the U.S. NationalScience Foundation (DEB-0316679, DEB-9904121, and DEB-0516431).

LITERATURE CITED

Cabana, G., and J. B. Rasmussen. 1994. Modeling food chainstructure and contaminant bioaccumulation using stablenitrogen isotopes. Nature 372:255–257.

Carpenter, S. R., and J. F. Kitchell, editors. 1993. The trophiccascade in lakes. Cambridge University Press, Cambridge,UK.

Cohen, J. E., and C. M. Newman. 1991. Community area andfood-chain length: theoretical predictions. American Natu-ralist 138:1542–1554.

Connor, E. F., and E. D. McCoy. 1979. The statistics andbiology of the species–area relationship. American Naturalist113:791–833.

Daugherty, M. P., J. P. Harmon, and C. J. Briggs. 2007.Trophic supplements to intraguild predation. Oikos 116:662–677.

DeAngelis, D. L., S. M. Bartell, and A. L. Brenkert. 1989.Effects of nutrient recycling and food-chain length onresilience. American Naturalist 134:778–805.

Diehl, S., and M. Feissel. 2001. Intraguild prey suffer fromenrichment of their resources: a microcosm experiment withciliates. Ecology 82:2977–2983.

Holt, R. D. 1996. Food webs in space: an island biogeographicperspective. Pages 313–323 in G. A. Polis and K. O.Winemiller, editors. Food webs: integration of patterns anddynamics. Chapman and Hall, New York, New York, USA.

Holt, R. D. 2002. Food webs in space: on the interplay ofdynamic instability and spatial processes. Ecological Re-search 17:261–273.

Holt, R. D., and G. R. Huxel. 2007. Alternative prey and thedynamics of intraguild predation: theoretical perspectives.Ecology 88:2706–2712.

Jenkins, B., R. L. Kitching, and S. L. Pimm. 1992. Productivity,disturbance and food web structure at a local spatial scale inexperimental container habitats. Oikos 65:249–255.

Kidd, K. A., D. W. Schindler, R. H. Hesslein, and D. C. G.Muir. 1998. Effects of trophic position and lipid onorganochlorine concentrations in fishes from subarctic lakesin Yukon Territory. Canadian Journal of Fisheries andAquatic Sciences 55:869–881.

Kitching, R. L. 2000. Food webs and container habitats.Cambridge University Press, Cambridge, UK.

Komonen, A., R. Penttila, M. Lindgren, and I. Hanski. 2000.Forest fragmentation truncates a food chain based on an old-growth forest bracket fungus. Oikos 90:119–126.

Kruess, A., and T. Tscharntke. 2000. Species richness andparasitism in a fragmented landscape: experiments and fieldstudies with insects on Vicia sepium. Oecologia 122:129–137.

Lawler, S. P., and P. J. Morin. 1993. Food web architecture andpopulation dynamics in laboratory microcosms of protists.American Naturalist 141:675–686.

Lawton, J. H. 1989. Food webs. Pages 43–78 in J. M. Cherrett,editor. Ecological concepts. Blackwell Scientific, Oxford,UK.

Menge, B. A., and J. P. Sutherland. 1987. Communityregulation: variation in disturbance, competition, and preda-tion in relation to environmental stress and recruitment.American Naturalist 130:730–757.

Mikkelson, G. M. 1993. How do food webs fall apart? A studyof changes in trophic structure during relaxation on habitatfragments. Oikos 67:539–547.

Namba, T., K. Tanabe, and N. Maeda. 2008. Omnivory andstability of food webs. Ecological Complexity 5:73–85.


epor

ts

Pimm, S. L. 1982. Food webs. Chapman and Hall, London,UK.

Pimm, S. L., and J. H. Lawton. 1977. The number of trophiclevels in ecological communities. Nature 275:542–544.

Post, D. M. 2002a. The long and short of food-chain length.Trends in Ecology and Evolution 17:269–277.

Post, D. M. 2002b. Using stable isotopes to estimate trophicposition: models, methods, and assumptions. Ecology 83:703–718.

Post, D. M., C. A. Layman, D. A. Arrington, G. Takimoto, J.Quattrochi, and C. G. Montana. 2007. Getting to the fat ofthe matter: models, methods, and assumptions for dealingwith lipids in stable isotope analysis. Oecologia 152:179–189.

Post, D. M., M. L. Pace, and N. G. Hairston. 2000. Ecosystemsize determines food-chain length in lakes. Nature 405:1047–1049.

Post, D. M., and G. Takimoto. 2007. Proximate structuralmechanisms for variation in food-chain length. Oikos 116:775–782.

Power, M. E., M. S. Parker, and J. T. Wootton. 1996.Disturbance and food chain length in rivers. Pages 286–297in G. A. Polis and K. O. Winemiller, editors. Food webs:integration of pattern and dynamics. Chapman and Hall,New York, New York, USA.

R Development Core Team. 2006. R: a language andenvironment for statistical computing. R Foundation forStatistical Computing, Vienna, Austria.

Rosenzweig, M. L. 1995. Species diversity in space and time.Cambridge University Press, Cambridge, UK.

Schoener, T. W. 1968. The Anolis lizards of Bimini: resourcepartitioning in a complex fauna. Ecology 49:704–726.

Schoener, T. W. 1989. Food webs from the small to the large.Ecology 70:1559–1589.

Schoener, T. W., and A. Schoener. 1982. Intraspecific variationin home-range size in some Anolis lizards. Ecology 63:809–823.

Schoener, T. W., and A. Schoener. 1983a. Distribution ofvertebrates on some very small islands. I. Occurrence

sequences of individual species. Journal of Animal Ecology52:209–235.

Schoener, T. W., and A. Schoener. 1983b. Distribution ofvertebrates on some very small islands. II. Patterns in speciesnumber. Journal of Animal Ecology 52:237–262.

Schoener, T. W., and D. A. Spiller. 1987. Effect of lizards onspider populations: manipulative reconstruction of a naturalexperiment. Science 236:949–952.

Schoener, T. W., and D. A. Spiller. 1996. Devastation of preydiversity by experimentally introduced predators in the field.Nature 381:691–694.

Schoener, T. W., D. A. Spiller, and L. W. Morrison. 1995.Variation in the hymenopteran parasitoid fraction onBahamian islands. Acta Oecologica 16:103–121.

Schoener, T. W., and C. A. Toft. 1983. Spider populations:extraordinarily high densities on islands without top preda-tors. Science 219:1353–1355.

Spencer, M., and P. H. Warren. 1996. The effects of habitat sizeand productivity on food web structure in small aquaticmicrocosms. Oikos 75:419–430.

Spiller, D. A., and A. A. Agrawal. 2003. Intense disturbanceenhances plant susceptibility to herbivory: natural andexperimental evidence. Ecology 84:890–897.

Spiller, D. A., J. B. Losos, and T. W. Schoener. 1998. Impact ofa catastrophic hurricane on island populations. Science 281:695–697.

Spiller, D. A., and T. W. Schoener. 1998. Lizards reduce spiderspecies richness by excluding rare species. Ecology 79:503–516.

Thompson, R. M., and C. R. Townsend. 2005. Energyavailability, spatial heterogeneity and ecosystem size predictfood-web structure in streams. Oikos 108:137–148.

Vander Zanden, M. J., and W. W. Fetzer. 2007. Global patternsof aquatic food chain length. Oikos 116:1378–1388.

Vander Zanden, M. J., B. J. Shuter, N. Lester, and J. B.Rasmussen. 1999. Patterns of food chain length in lakes: astable isotope study. American Naturalist 154:406–416.

APPENDIX

A table showing the sensitivity of the relationship between food-chain length and ecosystem size to different assumptions abouttrophic fractionation and isotopic baseline (Ecological Archives E089-174-A1).


eports

ecosystem size, but not disturbance, determines food …

Documents