Historical Overfishing and the Recent Collapse of Coastal Ecosystems
Ecological extinction caused by overfishing precedes all other pervasive human disturbance to coastal ecosystems, including pollution, degradation of water quality, and anthropogenic climate change. Historical abundances of large consumer species were fantastically large in comparison with recent observations. Paleoecological, archaeological, and historical data show that time lags of decades to centuries occurred between the onset of overfishing and consequent changes in ecological communities, because unfished species of similar trophic level assumed the ecological roles of overfished species until they too were overfished or died of epidemic diseases related to overcrowding. Retrospective data not only help to clarify underlying causes and rates of ecological change, but they also demonstrate achievable goals for restoration and management of coastal ecosystems that could not even be contemplated based on the limited perspective of recent observations alone.
Historical Overﬁshing and the Recent
Collapse of Coastal Ecosystems
Jeremy B. C. Jackson,
* Michael X. Kirby,
Wolfgang H. Berger,
Karen A. Bjorndal,
Louis W. Botsford,
Bruce J. Bourque,
Roger H. Bradbury,
James A. Estes,
Terence P. Hughes,
Carina B. Lange,
Hunter S. Lenihan,
John M. Pandolﬁ,
Charles H. Peterson,
Robert S. Steneck,
Mia J. Tegner,
† Robert R. Warner
Ecological extinction caused by overﬁshing precedes all other pervasive
human disturbance to coastal ecosystems, including pollution, degrada-
tion of water quality, and anthropogenic climate change. Historical abun-
dances of large consumer species were fantastically large in comparison
with recent observations. Paleoecological, archaeological, and historical
data show that time lags of decades to centuries occurred between the
onset of overﬁshing and consequent changes in ecological communities,
because unﬁshed species of similar trophic level assumed the ecological
roles of overﬁshed species until they too were overﬁshed or died of
epidemic diseases related to overcrowding. Retrospective data not only
help to clarify underlying causes and rates of ecological change, but they
also demonstrate achievable goals for restoration and management of
coastal ecosystems that could not even be contemplated based on the
limited perspective of recent observations alone.
Few modern ecological studies take into ac-
count the former natural abundances of large
marine vertebrates. There are dozens of places
in the Caribbean named after large sea turtles
whose adult populations now number in the
tens of thousands rather than the tens of mil-
lions of a few centuries ago (1, 2). Whales,
manatees, dugongs, sea cows, monk seals, croc-
odiles, codfish, jewfish, swordfish, sharks, and
rays are other large marine vertebrates that are
now functionally or entirely extinct in most
coastal ecosystems (3–10). Place names for
oysters, pearls, and conches conjure up other
ecological ghosts of marine invertebrates that
were once so abundant as to pose hazards to
navigation (11), but are witnessed now only by
massive garbage heaps of empty shells.
Such ghosts represent a far more profound
problem for ecological understanding and
management than currently realized. Evi-
dence from retrospective records strongly
suggests that major structural and functional
changes due to overfishing (12) occurred
worldwide in coastal marine ecosystems over
many centuries. Severe overfishing drives
species to ecological extinction because over-
fished populations no longer interact signifi-
cantly with other species in the community
(5). Overfishing and ecological extinction
predate and precondition modern ecological
investigations and the collapse of marine eco-
systems in recent times, raising the possibil-
ity that many more marine ecosystems may
be vulnerable to collapse in the near future.
Importance of Historical Data
Most ecological research is based on local field
studies lasting only a few years and conducted
sometime after the 1950s without longer term
historical perspective (1, 8, 13). Such observa-
tions fail to encompass the life-spans of many
ecologically important species (13, 14) and crit-
ically important environmental disturbances
such as extreme cyclones or ENSO (El Nin˜o–
Southern Oscillation) events (8), as well as
longer term cycles or shifts in oceanographic
regimes and productivity (15–17). To help ad-
dress this problem, we describe ecosystem
structure predating modern ecological studies
using well-dated time series based on biological
(18, 19), biogeochemical (20, 21), physical
(22), and historical (23) proxies that are infor-
mative over a variety of spatial scales and bio-
geographic realms (24). Although proxies vary
in precision and clarity of the signals they mea-
sure, the use of multiple proxies that give the
same ecological signal greatly increases confi-
dence in results. Precision in age dating varies
from centuries to a single year, season, or event
in the exceptional case of varved sediments, ice
cores, and written historical records (25). Pre-
cision decreases with the amount of biological
or physical disturbance to the sediment ana-
We exploited data from many disciplines
that span the period over which anthropogen-
ic changes may have occurred. Because our
hypothesis is that humans have been disturb-
ing marine ecosystems since they first
learned how to fish, our time periods need to
begin well before the human occupation or
European colonization of a coastal region.
Broadly, our data fall into four categories and
1) Paleoecological records from marine
sediments from about 125,000 years ago to
the present, coinciding with the rise of mod-
ern Homo sapiens.
2) Archaeological records from human
coastal settlements occupied after about
10,000 years before the present (yr B.P.)
when worldwide sea level approached
present levels. These document human ex-
ploitation of coastal resources for food and
materials by past populations that range from
small-scale aboriginal societies to towns, cit-
ies, and empires.
3) Historical records from documents,
journals, and charts from the 15th century to
the present that document the period from the
first European trade-based colonial expansion
and exploitation in the Americas and the
South Pacific (23).
4) Ecological records from the scientific
literature over the past century to the present
covering the period of globalized exploitation
of marine resources. These also help to cali-
brate the older records.
Scripps Institution of Oceanography, University of
California, San Diego, La Jolla, CA 92093–0244, USA.
Center for Tropical Paleoecology and Archeology,
Smithsonian Tropical Research Institute, Box 2072,
Balboa, Republic of Panama.
National Center for
Ecological Analysis and Synthesis, 735 State Street,
Suite 300, Santa Barbara, CA 93101, USA.
Carr Center for Sea Turtle Research and Department
of Zoology, University of Florida, Gainesville, FL
Department of Wildlife, Fish, and Con-
servation Biology, University of California, Davis, CA
Department of Anthropology, 155 Pet-
tengill Hall, Bates College, Lewiston, ME 04240, USA.
Centre for Resource and Environmental Studies, Aus-
tralian National University, Canberra, ACT 0200, Aus-
Department of Anthropology, University of
Oregon, Eugene, OR 97403, USA.
Survey, A-316 Earth and Marine Sciences Building,
University of California, Santa Cruz, CA 95064, USA.
Center for Coral Reef Biodiversity, Department of
Marine Biology, James Cook University, Townsville,
QLD 4811, Australia.
Department of Geophysical
Sciences, University of Chicago, 5734 South Ellis Av-
enue, Chicago, IL 60637, USA.
Institute of Marine
Sciences, University of North Carolina at Chapel Hill,
3431 Arendell Street, Morehead City, NC 28557, USA.
Department of Paleobiology, National Museum of
Natural History, Smithsonian Institution, Washington,
DC 20560–0121, USA.
School of Marine Sciences,
University of Maine, Darling Marine Center, Orono,
ME 04573, USA.
Department of Ecology, Evolution,
and Marine Biology, University of California, Santa
Barbara, CA 93106, USA.
*To whom correspondence should be addressed. E-
www.sciencemag.org SCIENCE VOL 293 27 JULY 2001 629
E COLOGY T HROUGH T IME
Time Periods, Geography, and Analysis
We recognize three different but overlapping
periods of human impact on marine ecosys-
tems: aboriginal, colonial, and global. Ab-
original use refers to subsistence exploitation
of near-shore, coastal ecosystems by human
cultures with relatively simple watercraft and
extractive technologies that varied widely in
magnitude and geographic extent. Colonial
use comprises systematic exploitation and de-
pletion of coastal and shelf seas by foreign
mercantile powers incorporating distant re-
sources into a developing market economy.
Global use involves more intense and geo-
graphically pervasive exploitation of coastal,
shelf, and oceanic fisheries integrated into
global patterns of resource consumption, with
more frequent exhaustion and substitution of
fisheries. In Africa, Europe, and Asia, these
cultural stages are strongly confounded in
time and space, so that their differential sig-
nificance is difficult to establish. However, in
the Americas, New Zealand, and Australia
the different stages are well separated in time,
and the aboriginal and colonial periods began
at different times in the different regions.
Thus, we can distinguish between cultural
stages, as well as between human impacts and
natural changes due to changing climate.
The addition of a deep historical dimen-
sion to analyze and interpret ecological prob-
lems requires that we sacrifice some of the
apparent precision and analytical elegance
prized by ecologists (1, 13, 14). Paleoeco-
logical, archaeological, and historical data
were collected for many purposes, vary wide-
ly in methods of collection and quality, and
are less amenable to many types of statistical
analysis than well-controlled experiments.
But none of these problems outweighs the
benefits of a historical approach. Clearly, we
cannot generate realistic null hypotheses
about the composition and dynamics of eco-
systems from our understanding of the
present alone, since all ecosystems have al-
most certainly changed due to both human
and natural environmental factors (8, 16, 27,
28). Here, we briefly review long-term hu-
man impacts in several key marine ecosys-
tems. These reconstructions provide insight
into the nature and extent of degraded eco-
systems that point to new strategies for mit-
igation and restoration that are unlikely to
emerge from modern monitoring programs.
Kelp forests characterize shallow, rocky hab-
itats from warm temperate to subarctic re-
gions worldwide and provide complex envi-
ronments for many commercially important
fishes and invertebrates (29). Northern Hemi-
sphere kelp forests have experienced wide-
spread reductions in the number of trophic
levels and deforestation due to population
explosions of herbivores following the re-
moval of apex predators by fishing (Fig. 1, A
and B). Phase shifts between forested and
deforested states (the latter known as “sea
urchin barrens”) result from intense grazing
due to increased abundance and altered for-
aging patterns of sea urchins made possible in
turn by human removal of their predators and
competitors (7, 8, 30–32).
The kelp forest ecosystem of the Northern
Pacific arose during the last 20 million years
with the evolution of kelps, strongylocent-
rotid sea urchins, sea otters, and the extinct
Steller’s sea cow (6 ). Sea cows were widely
distributed across the northern Pacific Rim
through the Late Pleistocene. They may have
been eliminated from most of their range by
aboriginal hunting at the end of the Pleisto-
cene and in the early Holocene, because they
survived thousands of years longer in the
western Aleutian Islands that were not peo-
pled until about 4000 yr B.P. (6). By the time
of European contact in 1741, sea cows per-
Fig. 1. Simpliﬁed coastal food webs showing changes in some of the important top-down
interactions due to overﬁshing; before (left side) and after (right side) ﬁshing. (A and B) Kelp forests
for Alaska and southern California (left box), and Gulf of Maine (right box). (C and D) Tropical coral
reefs and seagrass meadows. (E and F) Temperate estuaries. The representation of food webs after
ﬁshing is necessarily more arbitrary than those before ﬁshing because of rapidly changing recent
events. For example, sea urchins are once again rare in the Gulf of Maine, as they were before the
overﬁshing of cod, due to the recent ﬁshing of sea urchins that has also permitted the recovery of
kelp. Bold font represents abundant; normal font represents rare; “crossed-out” represents extinct.
Thick arrows represent strong interactions; thin arrows represent weak interactions.
27 JULY 2001 VOL 293 SCIENCE www.sciencemag.org630
E COLOGY T HROUGH T IME
sisted only in the Commander Islands, the
only islands of the Aleutians unoccupied by
aboriginal people. European fur traders killed
the last sea cow 27 years later in 1768. We
have no idea to what extent abundant sea
cows grazed kelp forests, although their ap-
parent inability to dive deeply probably lim-
ited their grazing to the surface canopy of
kelps and to seaweeds lining the shore (6 ).
Northern Pacific kelp forests presum-
ably flourished before human settlement
because predation by sea otters on sea ur-
chins prevented the urchins from overgraz-
ing kelp (30). Aboriginal Aleuts greatly
diminished sea otters beginning around
2500 yr B.P., with a concomitant increase
in the size of sea urchins (31). Fur traders
subsequently hunted otters to the brink of
extinction in the 1800s with the attendant
collapse of kelp forests grazed away by sea
urchins released from sea otter predation.
Legal protection of sea otters in the 20th
century partially reversed this scenario.
However, kelp forests are again being de-
pleted in areas of Alaska because of in-
creased predation on sea otters by killer
whales (33). The whales shifted their diet to
sea otters from seals and sea lions, which
are in drastic decline.
A similar sequence of events occurred in
kelp forests of the Gulf of Maine (7, 34). Sea
otters were never present, but Atlantic cod
and other large ground fish are voracious
predators of sea urchins. These fishes kept
sea urchin populations small enough to allow
persistence of kelp forests despite intensive
aboriginal and early European hook-and-line
fishing for at least 5000 years. New mecha-
nized fishing technology in the 1920s set off
a rapid decline in numbers and body size of
coastal cod in the Gulf of Maine (7 ) (Fig. 2A
and Table 1) that has extended offshore to
Georges Bank (35). Formerly dominant pred-
atory fish are now ecologically extinct and
have been partially replaced by smaller and
commercially less important species. Lob-
sters, crabs, and sea urchins rose in abun-
dance accordingly (7). Kelp forests disap-
peared with the rise in sea urchins due to
removal of predatory fish, and then reap-
peared when sea urchins were in turn reduced
to low abundance by fishing.
The more diverse food web of southern
California kelp forests historically included
spiny lobsters and large sheephead labrid fish
in addition to sea otters as predators of sea
urchins, as well as numerous species of aba-
lone that compete with sea urchins for kelps
(Fig. 1, A and B) (36 ). Aboriginal exploita-
tion began about 10,000 yr B.P. and may
have had local effects on kelp communities
(37). The fur trade effectively eliminated sea
otters by the early 1800s (38), but kelp forests
did not begin to disappear on a large scale
until the intense exploitation and ecological
extinction of sheephead, spiny lobsters, and
abalone starting in the 1950s (8, 36) (Table 1
and Fig. 1, A and B). Subsequent fishing of
the largest sea urchin species in the 1970s and
1980s resulted in the return of well-devel-
oped kelp forests in many areas that, as in the
Gulf of Maine, effectively lack trophic levels
higher than that of primary producers (36,
Coral reefs are the most structurally complex
and taxonomically diverse marine ecosys-
tems, providing habitat for tens of thousands
of associated fishes and invertebrates (40).
Aboriginal fishing in coral reef environments
began at least 35,000 to 40,000 years ago in
the western Pacific (41) but appears to have
had limited ecological impact. Recently, cor-
al reefs have experienced dramatic phase
shifts in dominant species due to intensified
human disturbance beginning centuries ago
(1) (Fig. 1, C and D). The effects are most
pronounced in the Caribbean (42) but are also
apparent on the Great Barrier Reef in Austra-
lia despite extensive protection over the past
three decades (43).
Large species of branching Acropora cor-
als dominated shallow reefs in the tropical
western Atlantic for at least half a million
years (44–46) until the 1980s when they
declined dramatically (42, 47) (Fig. 2B and
Table 1). Patterns of community membership
and dominance of coral species were also
highly predictable (44), so that there is a clear
baseline of pristine coral community compo-
sition before human impact.
Western Atlantic reef corals suffered sud-
den, catastrophic mortality in the 1980s due
to overgrowth by macroalgae that exploded
in abundance after mass mortality of the su-
perabundant sea urchin Diadema antillarum
that was the last remaining grazer of macroal-
gae (42, 47). Early fisheries reports suggest
that large herbivorous fishes were already
rare before the 20th century (48). However,
macroalgae were held in check until the last
major herbivore, Diadema, was lost from the
system through disease (42, 47).
Corals on the Great Barrier Reef have
experienced recurrent mass mortality since
1960 due to spectacular outbreaks of the
crown-of-thorns starfish Acanthaster planci
that feeds on coral (49). The causes of out-
breaks are controversial, but they are almost
certainly new phenomena. There are no early
records of Acanthaster in undisturbed fossil
deposits, in aboriginal folklore, or in accounts
of European explorers and fishers. Now, in
recent decades, the frequency and intensity of
outbreaks have exceeded the capability of
longer lived species to recover as outbreaks
have become more chronic than episodic
Fig. 2. Retrospective data showing baselines
before ecosystem collapse. (A) Time series of
mean body length of Atlantic cod from kelp
forests in the coastal Gulf of Maine. The earlier
ﬁve data points are derived from archaeological
records, whereas the last three points are from
ﬁsheries data (113). Vertical bars represent the
standard error. Horizontal bars represent the
time range of data for a single interval of
observations. (B) Paleoecological and ecological
data showing the percentage of Caribbean lo-
calities with Acropora palmata (Œ)orA. cervi-
cornis (■) as the dominant shallow-water coral
in the Late Pleistocene, Holocene, before 1983,
and after 1983 (114). Percentages of localities
are signiﬁcantly different over the four time
periods for A. palmata (
⫽ 34.0, P ⬍0.0001,
df ⫽ 3) and A. cervicornis (
⫽ 22.4, P
⬍0.0001, df ⫽ 3). Vertical and horizontal bars
are as in (A). (C) Paleoecological and ﬁsheries
data from Chesapeake Bay showing the ratio
in abundance of planktonic to benthic dia-
toms (dotted line) (77) and landings of the
oyster Crassostrea virginica (solid line) (80).
The planktonic to benthic diatom ratio is a
proxy for eutrophication that shows the rel-
ative amount of planktonic to benthic prima-
ry production (77). For over 1200 years this
ratio remained fairly constant at about 1:1,
but then increased threefold coincidentally
with increased runoff of sediments and nutrients due to European agriculture after 1750. The
ratio remained at about 3:1 between 1830 and 1930, after which it increased dramatically to
about 8:1. Oyster landings show an initial increase in the early 19th century, peak in 1884, and
subsequent collapse as deep channel reefs were destroyed by mechanical dredging (80). These
data strongly imply that oysters were able to limit the potential for eutrophication induced by
increased inputs of nutrients between 1750 and 1930 until oyster populations collapsed as a
result of overﬁshing.
www.sciencemag.org SCIENCE VOL 293 27 JULY 2001 631
E COLOGY T HROUGH T IME
Table 1. Retrospective records from coastal ecosystems that offer baselines that contrast with recent observations. Data source: P, paleoecologial; A, archaeological; H, historical; F, ﬁsheries; E, ecological.
Inferred causes: 1, ﬁshing; 2, mechanical habitat destruction by ﬁshing; 3, inputs. Abbreviations: BSi, biologically bound silica; DOP, degree of pyritization of iron; dec., decrease; inc., increase. References after
115 are located on Science Online (www.sciencemag.org/cgi/content/full/293/5530/629/DC1).
observation or estimate
observation or estimate
Sea Otter Paciﬁc Ocean H, E Area estimates 260 ⬎100,000 individuals 30,000 individuals ⬎3.3-fold dec. 1 116
Alaska H Herd size 259 ⬍5,000 sea cows 0 Extinction 1 117
Atlantic cod Gulf of Maine A Cod vertebrae 3550 Mean body length of 1.0 m Mean body length of
3-fold dec. 1 113
White abalone California E Number per area 30 ⬎2,000 per ha 1.0 ⫾ 0.4 per ha ⬎2,000-fold dec. 1 118
Coral Caribbean Sea P, E % sites with A. palmata
125,000 80% of Pleistocene sites 15% of post-1982 sites 5.3-fold dec. 1 114
Coral Caribbean Sea P, E % sites with A. cervicornis
125,000 63% of Pleistocene sites 0% of post-1982 sites 100% loss 1 114
Coral Bahamas P, E Standardized abundance of A.
125,000 12 1 12-fold dec. 1 119
Coral Belize P Relative abundance 3,130 A. cervicornis dominant A. cervicornis absent 100% loss 1 45
Coral Netherlands Antilles E Coral cover at 10 m 27 54% coral cover 31% coral cover 1.7-fold dec. 1 120
Coral Jamaica E Coral cover at 10 m 23 73% coral cover 4% coral cover 18-fold dec. 1 42
Monk seal Caribbean Sea H Historical reports ⬎300 Abundant 0 Extinction 1 4, 68
Coral Moreton Bay P, E Acropora dominance in fossil
8,000 Dominated reefs throughout
Only one small Acropora
Decrease 3 121
Tropical and subtropical seagrass beds
Green turtle Caribbean Sea E Biomass estimates ⬎300 ⬎16.1 ⫻ 10
50-kg turtles ⬎1.1 ⫻ 10
15-fold dec. 1 2, 122
Green turtle Caribbean Sea H Hunting, biomass estimates ⬎300 ⬎3.3 ⫻ 10
adult turtles ⬎1.1 ⫻ 10
30-fold dec. 1 1, 122
Seagrass beds Tampa Bay H Area 121 30,970 ha 10,759 ha 3-fold dec. 1, 2, 3 123, 124
Dugong Eastern Australia H Herd size ⬎100 ⬎1.0 ⫻ 10
⬎74-fold dec. 1 125, 126
Dugong Moreton Bay H Herd size 107 ⬎104,000 estimated
500 estimated dugongs ⬎208-fold dec. 1 125, 127
Oysters and eutrophication in estuaries
Inputs Chesapeake Bay P Sedimentation rate 1,900 0.04 cm year
0.2 cm year
5-fold inc. 3 77
Eutrophication Chesapeake Bay P Total organic carbon 1,900 0.26 mg cm
2.3 mg cm
9-fold inc. 2, 3 77
Eutrophication Chesapeake Bay P Centric/pennate diatom ratio 1,450 1:1 ratio 8:1 ratio 8-fold inc. 2, 3 77
Eutrophication Chesapeake Bay P Dinoﬂagellate cysts (Spiniferites
⬎300 50% relative abundance 80% relative abundance 1.6-fold inc. 2, 3 128
27 JULY 2001 VOL 293 SCIENCE www.sciencemag.org632
E COLOGY T HROUGH T IME
One possible explanation for Acanthaster
outbreaks is that overfishing of species that
prevy upon larval or juvenile stages of
crown-of-thorns starfish is responsible for
massive recruitment of the starfish (51). The
highly cryptic, predator-avoiding behavior of
juvenile starfish, their formidable antipreda-
tor defenses as subadults and adults, and the
reduction of some generalized predatory fish-
es on the Great Barrier Reef all point to such
a “top-down” explanation. Commercial and
recreational fishing, as well as indirect effects
of intensive trawling for prawns, are likely
explanations for decreased abundance of
predators of crown-of-thorns starfish (52).
Massive recruitment of starfish may also be
due to “bottom-up” increases in productivity
due to increased runoff of nutrients from the
land (53). In either case, the explanation is
almost certainly historical and anthropogenic,
and cannot be resolved by recent observa-
Expeditions occurred annually to northern
Australia from the Malay Archipelago
throughout the 18th and 19th centuries to
harvest an estimated 6 million sea cucumbers
each season (54 ). After European coloniza-
tion, industrial-scale fishing developed along
the Great Barrier Reef and subtropical east
Australian coast in the early to mid–19th
century (55). Whales, dugongs, turtles, pearl
oysters, and Trochus shell were each heavily
exploited only to rapidly collapse, and all
have failed to regain more than a small frac-
tion of their former abundance (55–57). Fish-
ing of pelagic and reef fishes, sharks, and
prawns has continued to the present, although
catch per unit effort has declined greatly (58).
Tropical and Subtropical Seagrass Beds
Seagrass beds cover vast areas of tropical
and subtropical bays, lagoons, and conti-
nental shelves (59). Seagrasses provide for-
age and habitat for formerly enormous
numbers of large sirenians (dugong and
manatee) and sea turtles, as well as diverse
assemblages of fishes, sharks, rays, and
invertebrates, including many commercial-
ly important species (59–62) (Fig. 1, C and
D). Like coral reefs, seagrass beds seemed
to be highly resilient to human disturbance
until recent decades when mass mortality of
seagrasses became common and wide-
spread (63–65). Examples include the die-
off of turtlegrass in Florida Bay and the
Gulf of Mexico in the 1980s (65) and the
near disappearance of subtidal seagrasses
in the offshore half of Moreton Bay near
Brisbane, Australia, over the past 20 to 30
years (63, 64 ). Proximate causes of these
losses include recent increases in sedimen-
tation, turbidity, or disease (63–65). How-
ever, extirpation of large herbivorous ver-
tebrates beginning centuries ago had al-
ready profoundly altered the ecology of
Table 1. (Continued)
observation or estimate
observation or estimate
Seagrass beds Fleets Bay, CB H Area 63 273 ha 16 ha 17-fold dec. 1, 2, 3 78
Oyster reefs Chesapeake Bay F Oyster landings 116 6.2 ⫻ 10
0.12 ⫻ 10
52-fold dec. 2 80
Oyster reefs Tangier Sound, CB F Area 122 44.6 km
0 100% loss 2 129
Anoxia Cheapeake Bay P Degree of pyritization 1,900 0.32 DOP 0.51 DOP 2-fold inc. 2, 3 77
Seagrass beds Botany Bay H Area of Posidonia beds 58 445 ha 188 ha 2.4-fold dec. 1, 2, 3 130
Eutrophication Baltic Sea P Planktic diatom relative
250 25% of total diatom
80% of total diatom
3-fold inc. 3 131
Eutrophication Baltic Sea P Total organic carbon
138 3.2 gC m
22-fold inc. 3 132
Anoxia Baltic Sea P Laminated sediments 100 5% of cores laminated 90% of cores laminated 18-fold inc. 3 85
Offshore benthic communities
Oyster reefs Foveaux Strait, NZ F Oyster landings 34 127 ⫻ 10
oysters/year* 15 ⫻ 10
8-fold dec. 2 93
Oyster reefs Foveaux Strait, NZ F Reef by-catch per station 38 1 in 4 stations had reef
1 in 7 stations had reef
2-fold dec. 2 93
Eutrophication Gulf of Mexico P Biologically bound silica 295 0.29% BSi 1.00% BSi 3.4-fold inc. 3 106
Eutrophication Gulf of Mexico P Total organic carbon
100 2.4 mg C cm
7.8 mg C cm
3.3-fold inc. 3 133
Hypoxia Gulf of Mexico P Benthic foraminifera 295 71 Ammonia–Elphidium
1.2-fold inc. 3 134
Eutrophication Adriatic Sea P Eutrophic benthic foraminifera
170 6% relative abundance 38% relative abundance 6-fold inc. 3 135
Eutrophication Adriatic Sea P Coccolithophorids 286 100 cells/g of sediment 1.6 ⫻ 10
15,700-fold inc. 3 136
*Baseline taken from peak in landings.
www.sciencemag.org SCIENCE VOL 293 27 JULY 2001 633
E COLOGY T HROUGH T IME
seagrass beds in ways that increased their
vulnerability to recent events.
Vast populations of very large green turtles
were eliminated from the Americas before the
19th century (1, 2) (Table 1). Formerly great
populations of green turtles in Moreton Bay,
Australia, also were greatly reduced by the
early 20th century (66). Moreover, there are no
estimates of abundances of turtles in Australia
at the dawn of European exploitation, so that
reported reductions must be only a small frac-
tion of the total numbers lost. All turtle species
continue to decline at unsustainable rates along
the Great Barrier Reef today (67).
Abundant green turtles closely crop
turtlegrass and greatly reduce the flux of
organic matter and nutrients to sediments
(59–62, 68). In the near absence of green
turtles today, turtlegrass beds grow longer
blades that baffle currents, shade the bot-
tom, start to decompose in situ, and provide
suitable substrate for colonization by the
slime molds that cause turtlegrass wasting
disease (65). Deposition within the beds of
vastly more plant detritus also fuels micro-
bial populations, increases the oxygen de-
mand of sediments, and promotes hypoxia
(65). Thus, all the factors that have been
linked with recent die-off of turtlegrass
beds in Florida Bay (65), except for chang-
es in temperature and salinity, can be at-
tributed to the ecological extinction of
green turtles (27).
European colonists did not exploit tropical
American manatees as systematically as they
exploited green turtles, so the data related to
fisheries are poor. We know, however, that
manatees were extensively fished by aboriginal
people and by early colonists (68). In Australia,
aboriginal people also harvested dugongs ex-
tensively long before European colonization
(3), yet the numbers reported by early colonists
were vast. Three- or four-mile-long herds com-
prising tens of thousands of large individuals
were observed in Wide Bay in about 1870 (69)
and in Moreton Bay as recently as 1893 (70).
Widespread colonial exploitation of dugongs
for their flesh and oil along the southern
Queensland coast resulted in the crash of the
dugong fishery by the beginning of the 20th
century (3) (Table 1). Ironically, scientists re-
cently reported the “discovery of a large popu-
lation” of dugongs in Moreton Bay—a mere
300 individuals (71). Further north, numbers of
dugongs in the vast southern half of the Great
Barrier Reef had dwindled to fewer than 4000
when they were first accurately counted in
1986 – 87, with a further 50 to 80% decline in
recent years (72). These increasingly fragment-
ed populations represent the last remnants of
the vast herds of the early 19th century and
The ecological implications of these re-
ductions are at least as impressive as those for
green turtles. Moderate sized herds of dug-
ongs remove up to 96% of above-ground
biomass and 71% of below-ground biomass
of seagrasses (73). Their grazing rips up large
areas of seagrass beds, providing space for
colonization by competitively inferior species
of seagrasses. Dugong grazing also produces
massive amounts of floating debris and dung
that are exported to adjacent ecosystems. The
decline in seagrasses in Moreton Bay is cer-
tainly due in large part to the dramatic decline
in water quality due to eutrophication and
runoff of sediment (63, 64). Nevertheless, as
noted for green turtles and turtlegrass in Flor-
ida Bay, the cessation of systematic plowing
of the bay floor by once abundant dugongs
must also have been a major factor.
Oysters and Eutrophication in
Temperate estuaries worldwide are undergo-
ing profound changes in oceanography and
ecology due to human exploitation and pol-
lution, rendering them the most degraded of
marine ecosystems (74–76) (Fig. 1, E and F).
The litany of changes includes increased sed-
imentation and turbidity (77); enhanced epi-
sodes of hypoxia or anoxia (74, 75, 77 ); loss
of seagrasses (78) and dominant suspension
feeders (79), with a general loss of oyster reef
habitat (80); shifts from ecosystems once
dominated by benthic primary production to
those dominated by planktonic primary pro-
duction (77); eutrophication (74–76) and en-
hanced microbial production (81); and higher
frequency and duration of nuisance algal and
toxic dinoflagellate blooms (82, 83), out-
breaks of jellyfish (79), and fish kills (83).
Most explanations for these phenomena em-
phasize “bottom-up” increases in nutrients
like nitrogen and phosphorus as causes of
phytoplankton blooms and eutrophication
(74–76), an interpretation consistent with the
role of estuaries as the focal point and sewer
for many land-based, human activities. Nev-
ertheless, long-term records demonstrate that
reduced “top-down” control resulting from
losses in benthic suspension feeders predated
The oldest and longest records come from
cores in sediments from Chesapeake Bay
(77) and Pamlico Sound (84 ) in the eastern
United States and from the Baltic Sea (85)
that extend back as far as 2500 yr B.P. (Fig.
2C and Table 1). A general sequence of
ecological change is apparent in all three
cases, but the timing of specific ecological
transitions differs among estuaries in keeping
with their unique histories of land use, ex-
ploitation, and human population growth—a
difference that rules out a simple climatic
explanation. Increased sedimentation and
burial of organic carbon began in the mid–
18th century in Chesapeake Bay, coincident
with widespread land clearance for agricul-
ture by European colonists (77). The main
ecological response was a gradual shift in the
taxa responsible for primary production that
began in the late 18th century. Seagrasses and
benthic diatoms on the bay floor declined,
while planktonic diatoms and other phyto-
plankton in the water column corresponding-
ly increased. However, anoxia and hypoxia
were not widespread until the 1930s when
phytoplankton populations and the flux of
organic matter to the bay floor increased
dramatically with concomitant loss of benthic
fauna (75, 77 ) (Fig. 2C and Table 1). Similar
changes began in the 1950s in the Baltic Sea,
with widespread expansion of the extent of
anoxic laminated sediments (74, 85), and in
the 1950s to 1970s in Pamlico Sound (84 ).
Vast oyster reefs were once prominent
structures in Chesapeake Bay (11), where
they may have filtered the equivalent of the
entire water column every 3 days (79). De-
spite intensive harvesting by aboriginal and
early colonial populations spanning several
millennia, it was not until the introduction of
mechanical harvesting with dredges in the
1870s that deep channel reefs were seriously
affected (79, 80). Oyster catch was rapidly
reduced to a few percent of peak values by
the early 20th century (79, 80) (Fig. 2C and
Table 1). Only then, after the oyster fishery
had collapsed, did hypoxia, anoxia, and other
symptoms of eutrophication begin to occur in
the 1930s (75, 77), and outbreaks of oyster
parasites became prevalent only in the
1950s (80). Thus, fishing explains the bulk
of the decline, whereas decline in water
quality and disease were secondary factors
(80). However, now that oyster reefs are
destroyed, the effects of eutrophication,
disease, hypoxia, and continued dredging
interact to prevent the recovery of oysters
and associated communities (86 ). Field ex-
periments in Pamlico Sound demonstrate
that oysters grow well, survive to maturity,
and resist oyster disease when elevated
above the zone of summer hypoxia— even
in the presence of modern levels of eu-
trophication and pollution (87).
Overfishing of oysters to the point of eco-
logical extinction is just one example in a
general pattern of removal of species capable
of top-down control of community structure
in estuaries. Dense populations of oysters and
other suspension-feeding bivalves graze
plankton so efficiently that they limit blooms
of phytoplankton and prevent symptoms of
eutrophication (88, 89), just as occurs with
grazing by zooplankton in freshwater ecosys-
tems (90). The ecological consequences of
uncounted other losses are unknown. Gray
whales (now extinct in the Atlantic), dol-
phins, manatees, river otters, sea turtles, alli-
gators, giant sturgeon, sheepshead, sharks,
and rays were all once abundant inhabitants
of Chesapeake Bay but are now virtually
27 JULY 2001 VOL 293 SCIENCE www.sciencemag.org
E COLOGY T HROUGH T IME
Offshore Benthic Communities
Continental shelves cover more of the ocean
floor than all previously discussed environ-
ments combined. Commercially important
cod, halibut, haddock, turbot, flounder,
plaice, rays, and a host of other ground fishes,
scallops, cockles, and oysters have been
fished intensively for centuries from conti-
nental shelves of Europe and North America,
and more recently throughout the world (5, 7,
10, 91). Hook-and-line fishing was replaced
by intensive use of the beam trawl during the
18th century, and industrialized fishing was
further intensified with the advent of large
steam- and diesel-powered vessels and the
otter trawl at the end of the 19th century.
Reports of severely depleted fish stocks and
shifting of fishing grounds farther and farther
from home ports into the North Sea and the
outer Grand Banks were commonplace by the
beginning of the 19th century. Scientific in-
vestigation consistently lagged behind eco-
nomic realities of depleted stocks and inexo-
rable exploitation of more-distant fishing
grounds. As late as 1883, Thomas Huxley
claimed that fish stocks were inexhaustible
(92), a view discredited by the beginning of
the 20th century (5). Today, several formerly
abundant, large fish as well as formerly dense
assemblages of suspension feeders are eco-
logically extinct over vast areas (7–10, 93).
The Primacy of Overﬁshing in Human
Disturbance to Marine Ecosystems
Overfishing of large vertebrates and shellfish
was the first major human disturbance to all
coastal ecosystems examined (Table 1). Eco-
logical changes due to overfishing are strik-
ingly similar across ecosystems despite the
obvious differences in detail (Fig. 1, A to F).
Everywhere, the magnitude of losses was
enormous in terms of biomass and abundance
of large animals that are now effectively ab-
sent from most coastal ecosystems world-
wide. These changes predated ecological in-
vestigations and cannot be understood except
by historical analysis. Their timing in the
Americas and Pacific closely tracks European
colonization and exploitation in most cases.
However, aboriginal overfishing also had ef-
fects, as exemplified by the decline of sea
otters (and possibly sea cows) in the northeast
Pacific thousands of years ago.
There are three important corollaries to
the primacy of overfishing. The first is that
pollution, eutrophication, physical destruc-
tion of habitats, outbreaks of disease, inva-
sions of introduced species, and human-
induced climate change all come much later
than overfishing in the standard sequence
of historical events (Fig. 3). The pattern
holds regardless of the initial timing of
colonial overfishing that began in the
Americas in the 16th and 17th centuries and
in Australia and New Zealand in the 19th
century. The full sequence of events is most
characteristic of temperate estuaries like
Chesapeake Bay. Not all the human distur-
bances illustrated in Fig. 3 have affected all
ecosystems yet. But wherever these events
have occurred, the standard chronological
sequence of human disturbance and modi-
fication of ecosystems is recognizable.
The second important corollary is that
overfishing may often be a necessary precon-
dition for eutrophication, outbreaks of dis-
ease, or species introductions to occur (27 ).
For example, eutrophication and hypoxia did
not occur in Chesapeake Bay until the 1930s,
nearly two centuries after clearing of land for
agriculture greatly increased runoff of sedi-
ments and nutrients into the estuary (77 ).
Suspension feeding by still enormous popu-
lations of oysters was sufficient to remove
most of the increased production of phyto-
plankton and enhanced turbidity until me-
chanical harvesting progressively decimated
oyster beds from the 1870s to the 1920s (77,
80) (Fig. 2C).
The consequences of overfishing for out-
breaks of disease in the next lower trophic level
fall into two categories. The most straightfor-
ward is that populations in the lower level
become so dense that they are much more
susceptible to disease as a result of greatly
increased rates of transmission (94). This was
presumably the case for the sea urchin Diadema
on Caribbean reefs and the seagrass Thalassia
in Florida Bay. In contrast, among oysters dis-
ease did not become important in Chesapeake
Bay until oysters had been reduced to a few
percent of their original abundance (80), a pat-
tern repeated in Pamlico Sound (86, 87) and
Foveaux Strait, New Zealand (93). Two factors
may be responsible. First, oysters may have
become less fit owing to stresses like hypoxia
or sedimentation, making them less resistant to
disease (87). Alternatively, suspension feeding
by dense populations of oysters and associated
species on oyster reefs may have indirectly
limited populations of pathogens by favoring
other plankton—an explanation that may ex-
tend to blooms of toxic plankton and most other
outbreaks of microbial populations (88).
The third important corollary is that
changes in climate are unlikely to be the
primary reason for microbial outbreaks and
disease. The rise of microbes has occurred at
different times and under different climatic
conditions in different places, as exemplified
by the time lag between events in Chesapeake
Bay and Pamlico Sound (77, 79, 80, 84 ).
Anthropogenic climate change may now be
an important confounding factor, but it was
not the original cause. Rapid expansion of
introduced species in recent decades (95)
may have a similar explanation, in addition to
increase in frequency and modes of transport.
Massive removal of suspension feeders, graz-
ers, and predators must inevitably leave ma-
rine ecosystems more vulnerable to invasion
(96, 97 ).
Synergistic Effects of Human
Ecological extinction of entire trophic levels
makes ecosystems more vulnerable to other
natural and human disturbances such as nu-
trient loading and eutrophication, hypoxia,
disease, storms, and climate change. Expan-
sion and intensification of different forms of
human disturbance and their ecological ef-
fects on coastal ecosystems have increased
and accelerated with human population
growth, unchecked exploitation of biological
resources, technological advance, and the in-
creased geographic scale of exploitation
through globalization of markets. Moreover,
the effects are synergistic, so that the whole
response is much greater than the sum of
individual disturbances (98). This is perhaps
most apparent in the rise of eutrophication,
hypoxia, and the outbreak of toxic blooms
and disease following the destruction of oys-
ter reefs by mechanical harvesting of oysters
(79, 80, 86). Other possible examples are
outbreaks of seagrass wasting disease due to
the removal of grazers of seagrasses like the
green turtle (27 ).
A striking feature of such synergistic effects
is the suddenness of the transition in abundance
Fig. 3. Historical se-
quence of human distur-
bances affecting coastal
ecosystems. Fishing (step
1) always preceded other
human disturbance in all
cases examined. This is
the basis for our hypoth-
esis of the primacy of
overﬁshing in the deterio-
ration of coastal ecosys-
tems worldwide. Subse-
quent steps 2 through 5
have not been observed in
every example and may
vary in order.
www.sciencemag.org SCIENCE VOL 293 27 JULY 2001 635
E COLOGY T HROUGH T IME
of different kinds of organisms and com-
munity composition due to threshold ef-
fects (99). Ecological diversity and redun-
dancy within trophic levels is probably the
most important reason for the delay or time
lag between the onset of fishing and the
subsequent threshold response (42, 100).
The importance of biodiversity in the form
of ecological redundancy is clearly appar-
ent for the delay in the collapse of kelp
forests in southern California compared
with Alaska after the extirpation of sea
otters. Sheephead fish, spiny lobsters, and
abalone in the more diverse Californian
kelp forests kept sea urchin populations in
check until these predators and competitors
of sea urchins had also been effectively
eliminated (8, 36). Similarly, the sea urchin
Diadema kept macroalgae in check long
after the extreme overfishing of herbivo-
rous fishes on Caribbean coral reefs (42).
A second potentially important mecha-
nism for the suddenness of ecosystem col-
lapse is the elimination of previously un-
fished refuges that were protected historically
because of distance or expense of access. For
example, reef fishes all around Jamaica in the
1960s rarely reached reproductive maturity
so that the abundant recruits of fishes on
Jamaican reefs at that time must have come
from undiscovered populations in Jamaica or
elsewhere (101). But as more and more reefs
have been overfished, the potential sources of
such recruits must have effectively disap-
peared over wider areas (102). A similar sce-
nario has been proposed for the American
lobster with regard to loss of larvae from
deep-water offshore stocks (103).
Microbialization of the Global Coastal
Most recent changes to coastal marine eco-
systems subsequent to overfishing involve
population explosions of microbes responsi-
ble for increasing eutrophication (74–76, 81),
diseases of marine species (104), toxic
blooms (82, 83), and even diseases such as
cholera that affect human health (104, 105).
Chesapeake Bay (81) and the Baltic Sea (74 )
are now bacterially dominated ecosystems
with a trophic structure totally different from
that of a century ago. Microbial domination
also has expanded to the open ocean off the
mouth of the Mississippi River (106 ) and to
the Adriatic Sea (107 ).
Nowhere is the lack of historical perspec-
tive more damaging to scientific understand-
ing than for microbial outbreaks. Plans for
remediation of eutrophication of estuaries are
still based on the belief that eutrophication is
caused only by increased nutrients without
regard to overfishing of suspension feeders.
Even more remarkable is the attribution of
the rise in marine diseases to climate change
and pollution (104 ) without regard to the
pervasive removal of higher trophic levels
and the asynchronous outbreaks of disease in
different ecosystems that belie a simple cli-
Historical Perspectives for Ecosystem
The characteristic sequence of human distur-
bance to marine ecosystems (Fig. 3) provides
a framework for remediation and restoration
that is invisible without a historical perspec-
tive. More specific paleoecological, archaeo-
logical, and historical data should be obtained
to refine the histories of specific ecosystems
and as a tool for management, but the overall
patterns are clear. The historical magnitudes
of losses of large animals and oysters were so
great as to seem unbelievable based on mod-
ern observations alone (Table 1). Even seem-
ingly gloomy estimates of the global percent-
age of fish stocks that are overfished (108)
are almost certainly far too low. The shifting
baseline syndrome is thus even more insidi-
ous and ecologically widespread than is com-
On the other hand, recognition of these
losses shows what coastal ecosystems could
be like, and the extraordinary magnitude of
economic resources that are retrievable if we
are willing to act on the basis of historical
knowledge. The central point for successful
restoration is that loss of economically im-
portant fisheries, degradation of habitat at-
tractive to landowners and tourists, and emer-
gence of noxious, toxic, and life-threatening
microbial diseases are all part of the same
standard sequence of ecosystem deterioration
that has deep historical roots (27 ). Respond-
ing only to current events on a case-by-case
basis cannot solve these problems. Instead,
they need to be addressed by a series of bold
experiments to test the success of integrated
management for multiple goals on the scale
of entire ecosystems. With few exceptions,
such as the Caribbean monk seal and Steller’s
sea cow, most species that are ecologically
extinct probably survive in sufficient num-
bers for successful restoration. This optimism
is in stark contrast with the state of many
terrestrial ecosystems where many or most
large animals are already extinct (28). More-
over, we now have the theoretical tools (109)
to roughly estimate per capita interaction
strengths of surviving individuals of now rare
animals like sea turtles, sirenians, sharks, and
large groupers. We can then use these data to
build tentative models of the consequences of
the renewed abundance of these species in
their native environments that can in turn be
used to design large-scale, adaptive experi-
ments for ecosystem restoration, exploitation,
and management (96, 108, 110).
One obviously timely and overdue exper-
iment is to attempt the amelioration of eu-
trophication, hypoxia, and toxic blooms in
Chesapeake Bay by massive restoration of oys-
ter reefs (79). Experiments in Pamlico Sound
show that this is possible (86, 87, 96 ), and
modeling of food webs suggests that even par-
tial restoration of oysters would reduce eu-
trophication substantially (110). Aquaculture of
suspension-feeding bivalves like oysters might
be promoted to reverse the effects of eutrophi-
cation and to restore water quality in degraded
estuaries. Other important examples include the
restoration of coral reefs and seagrass beds by
protection of fishes, sharks, turtles, and sire-
nians in very large reserves on the scale of all of
Florida Bay and the Florida Keys—an ap-
proach recently advocated for terrestrial ecosys-
tems (111). Once again, small-scale grazing
experiments with reef fishes (112) show that
fishes could reverse the overgrowth of corals by
macroalgae on a massive scale. The potential
for reducing diseases of corals and turtlegrass
by restoring natural levels of grazing is unprov-
en but consistent with historical evidence (27).
In summary, historical documentation of the
long-term effects of fishing provides a hereto-
fore-missing perspective for successful man-
agement and restoration of coastal marine eco-
systems. Previous attempts have failed because
they have focused only on the most recent
symptoms of the problem rather than on their
deep historical causes. Contrary to romantic
notions of the oceans as the “last frontier”
and of the supposedly superior ecological
wisdom of non-Western and precolonial
societies, our analysis demonstrates that
overfishing fundamentally altered coastal
marine ecosystems during each of the cul-
tural periods we examined. Changes in eco-
system structure and function occurred as
early as the late aboriginal and early colo-
nial stages, although these pale in compar-
ison with subsequent events. Human im-
pacts are also accelerating in their magni-
tude, rates of change, and in the diversity of
processes responsible for changes over
time. Early changes increased the sensitiv-
ity of coastal marine ecosystems to subse-
quent disturbance and thus preconditioned
the collapse we are witnessing.
References and Notes
1. J. B. C. Jackson, Coral Reefs 16, S23 (1997).
2. K. A. Bjorndal, A. B. Bolten, M. Y. Chaloupka, Ecol.
Appl. 10, 269 (2000).
3. G. C. L. Bertram, C. K. Ricardo Bertram, Biol. J. Linn.
Soc. 5, 297 (1973).
4. K. W. Kenyon, J. Mammal. 58, 97 (1977).
5. D. H. Cushing, The Provident Sea (Cambridge Univ.
Press, Cambridge, UK, 1988).
6. J. A. Estes, D. O. Duggins, G. B. Rathbun, Conserv.
Biol. 3, 252 (1989).
7. R. S. Steneck, in Proceedings of the Gulf of Maine
Ecosystem Dynamics Scientiﬁc Symposium and
Workshop (RARGOM Report 91-1, Regional Associ-
ation for Research in the Gulf of Maine, Hanover,
NH, 1997), pp. 151–165.
8. P. K. Dayton, M. J. Tegner, P. B. Edwards, K. L. Riser,
Ecol. Appl. 8, 309 (1998).
9. J. M. Casey, R. A. Myers, Science 228, 690 (1998).
10. J. A. Hutchings, Nature 406, 882 (2000).
27 JULY 2001 VOL 293 SCIENCE www.sciencemag.org636
E COLOGY T HROUGH T IME
11. E. Ingersoll, in The History and Present Condition of
the Fishery Industries, G. B. Goode, Ed. (U.S. Depart-
ment of the Interior, Tenth Census of the United
States, Washington, DC, 1881), pp. 1–252.
12. “Fishing” is the original and most general term in
the English language for hunting and gathering all
kinds of organisms in the oceans, including plants,
invertebrates, and vertebrates ranging from sharks
to true ﬁshes to whales. “Overﬁshing” is the reduc-
tion of a species well below the sustainable yield or
to such low abundance that it may not recover even
if ﬁshing ceases.
13. J. H. Brown, Macroecology (Univ. of Chicago Press,
Chicago, IL, 1995).
14. J. B. C. Jackson, BioScience 41, 475 (1991).
15. R. C. Francis, S. R. Hare, Fish. Oceanogr. 3, 279
16. A. D. MacCall, Calif. Coop. Fish. Invest. 37, 100
17. C. B. Lange, S. K. Burke, W. H. Berger Clim. Change
16, 319 (1990).
18. J. B. C. Jackson, in Biotic Interactions in Recent and
Fossil Benthic Communities, M. J. S. Tevesz, P. L.
McCall, Eds. (Plenum, New York, 1983), pp. 39 –120.
19. S. J. M. Davis, The Archaeology of Animals (Yale
Univ. Press, New Haven, CT, 1987).
20. G. Wefer, W. H. Berger, J. Bijima, G. Fischer, in Use
of Proxies in Paleoceanography: Examples from the
South Atlantic, G. Fischer, G. Wefer, Eds. (Springer-
Verlag, Berlin, 1999), pp. 1–68.
21. D. S. Jones, Paleontol. Soc. Pap. 4, 37 (1998).
22. C. W. Finkl Jr., J. Coast. Res. Spec. Iss. 17, 402
23. A. W. Crosby, Ecological Imperialism: The Biological
Expansion of Europe 900-1900 (Cambridge Univ.
Press, Cambridge, UK, 1986).
24. Biological proxies include fossils or archaeological
remains that may indicate presence, abundance,
age, growth rate, health, or body size, among other
factors, of selected taxa. Biogeochemical proxies
include concentrations of inorganic or organic con-
stituents and isotopic ratios in sediments or skele-
tons that track former abundances of unpreserved
organisms and variations in parameters such as
temperature, salinity, oxygen, nutrient chemistry,
and productivity. Physical proxies include sedimen-
tary structures and composition that reﬂect water
movements, runoff from the land, and rate of sed-
imentation. Historical proxies include demographic,
customs, and commercial records, as well as explor-
ers’ and naturalists’ descriptions of sights and
events and nautical charts marking reefs, coastal
wetlands, and other landmarks.
25. R. S. Bradley, Paleoclimatology: Reconstructing Cli-
mates of the Quaternary (Academic Press, San Di-
ego, ed. 2, 1999).
26. S. M. Kidwell, K. W. Flessa, Annu. Rev. Ecol. Syst. 26,
27. J. B. C. Jackson, Proc. Natl. Acad. Sci. U.S.A. 98,
28. D. H. Janzen, P. S. Martin, Science 215, 19 (1982).
29. P. K. Dayton, Annu. Rev. Ecol. Syst. 16, 215 (1985).
30. J. A. Estes, J. F. Palmisano, Science 185, 1058 (1974).
31. C. A. Simenstad, J. A. Estes, K. W. Kenyon, Science
200, 403 (1978).
32. E. Sala, C. F. Boudouresque, M. Harmelin-Vivien,
Oikos 82, 425 (1998).
33. J. A. Estes, M. T. Tinker, T. M. Williams, D. F. Doak,
Science 282, 473 (1998).
34. B. J. Bourque, Diversity and Complexity in Prehistoric
Maritime Societies: A Gulf of Maine Perspective (Ple-
num, New York, 1995).
35. K. Sherman, Ecol. Appl. 1, 349 (1991).
36. M. J. Tegner, P. K. Dayton, ICES (Int. Counc. Expl.
Sea) J. Mar. Sci. 57, 579 (2000).
37. J. M. Erlandson et al., Radiocarbon 38, 355 (1996).
38. A. Ogden, The California Sea Otter Trade (Univ. of
California Press, Berkeley, 1941).
39. M. J. Tegner, P. K. Dayton, Mar. Ecol. Prog. Ser. 77,
40. N. Knowlton, Proc. Natl. Acad. Sci. U. S. A. 98, 5419
41. J. Allen, C. Gosden, J. P. White, Antiquity 63, 548
42. T. P. Hughes, Science 265, 1547 (1994).
43. R. A. Kenchington, Managing Marine Environments
(Taylor and Francis, New York, 1990).
44. J. M. Pandolﬁ, J. B. C. Jackson, Ecol. Monogr. 71,49
45. R. B. Aronson, W. F. Precht, I. G. MacIntyre, Coral
Reefs 17, 223 (1998).
46. J. Geister, Proc. 3rd Int. Coral Reef Symp. 1,23
47. H. A. Lessios, Annu. Rev. Ecol. Syst. 19, 371 (1988).
48. J. E. Duerden, West Indian Bull. 1901, 121 (1901).
49. J. Sapp, What Is Natural? Coral Reef Crisis (Oxford
Univ. Press, New York, 1999).
50. R. M. Seymour, R. H. Bradbury, Mar. Ecol. Prog. Ser.
176, 1 (1999).
51. R. Ormond et al., in Acanthaster and the Coral Reef:
A Theoretical Perspective, R. H. Bradbury, Ed.
(Springer-Verlag, Berlin, 1990), pp. 189–207.
52. A. Steven, An Analysis of Fishing Activities on Possi-
ble Predators of the Crown of Thorns Starﬁsh (Acan-
thaster planci) of the Great Barrier Reef (Prepared
for the Great Barrier Reef Marine Park Authority,
Townsville, Australia, 1988).
53. C. Birkeland, Mar. Biol. 69, 175 (1982).
54. C. C. MacKnight, The Voyage to Marege: Macassan
Trepangers in Northern Australia (Melbourne Univ.
Press, Melbourne, Australia, 1976).
55. D. T. Neil, in Moreton Bay and Catchment,I.R.
Tibbetts, N. J. Hall, W. C. Dennison, Eds. (Univ. of
Queensland, Brisbane, Australia, 1998), pp. 3–54.
56. D. Jones, The Whalers of Tangalooma (Nautical As-
sociation of Australia, Melbourne, 1980).
57. R. Ganter, The Pearl-Shellers of Torres Strait: Re-
source Use, Development and Decline, 1860s–1960s
(Melbourne Univ. Press, Melbourne, Australia, 1994).
58. R. H. Quinn, Fisheries Resources of the Moreton Bay
Region (Queensland Fish Management Authority,
Brisbane, Australia, 1993).
59. J. C. Ogden, in Handbook of Seagrass Biology: An
Ecosystem Perspective, R. C. Phillips, C. P. McRoy,
Eds. (Garland STPM, New York, 1980), pp. 173–198.
60. G. W. Thayer, D. W. Engel, K. A. Bjorndal, J. Exp. Mar.
Biol. Ecol. 62, 173 (1982).
61. J. C. Ogden, L. Robinson, H. Whitlock, H. Daganhart,
R. Cebula, J. Exp. Mar. Biol. Ecol. 66, 199 (1983).
62. G. W. Thayer, K. A. Bjorndahl, J. C. Ogden, S. L.
Williams, J. C. Zieman, Estuaries 7, 351 (1984).
63. H. Kirkman, Aquat. Bot. 5, 63 (1978).
64. E. G. Abal, W. C. Dennison, Mar. Freshw. Res. 47, 763
65. J. C. Zieman, J. W. Fourqurean, T. A. Frankovich,
Estuaries 22, 460 (1999).
66. C. J. Limpus, P. J. Couper, M. A. Read, Mem. Queens-
land Mus. 35, 139 (1994).
67. C. J. Limpus, in State of the Great Barrier Reef World
Heritage Area Workshop, D. Wachenfeld, J. Oliver, K.
Davis, Eds. (Great Barrier Reef Marine Park Author-
ity, Townsville, Australia, 1995), pp. 258–265.
68. W. Dampier, A New Voyage around the World (Do-
ver, New York, 1968).
69. E. Thorne, The Queen of the Colonies (Sampson Low,
Marson, Searle, and Rivington, London, 1876).
70. G. M. Allen, Am. Comm. Int. Wildl. Protect. Spec.
Pub. 11, 1 (1942).
71. G. E. Heinsohn, R. J. Lear, M. M. Bryden, H. Marsh,
B. R. Gardner, Environ. Conserv. 5, 91 (1978).
72. H. Marsh, P. Corkeron, I. Lawler, J. Lanyon, A. Preen,
Great Barrier Reef Marine Park Authority Rep. 41,1
73. A. Preen, Mar. Ecol. Prog. Ser. 124, 201 (1995).
74. R. Elmgren, Ambio 18, 326 (1989).
75. C. B. Ofﬁcer et al., Science 223, 22 (1984).
76. S. W. Nixon, Ophelia 41, 199 (1995).
77. S. R. Cooper, G. S. Brush, Estuaries 16, 617 (1993).
78. R. J. Orth, K. A. Moore, Science 222, 51 (1983).
79. R. I. E. Newell, in Understanding the Estuary: Advanc-
es in Chesapeake Bay Research, M. P. Lynch, E. C.
Krome, Eds. (Chesapeake Bay Research Consortium,
Baltimore, MD, 1988), pp. 536–546.
80. B. J. Rothschild, J. S. Ault, P. Goulletquer, M. He´ral,
Mar. Ecol. Prog. Ser. 111, 29 (1994).
81. R. B. Jonas, Am. Zool. 37, 612 (1997).
82. T. J. Smayda, in Toxic Marine Phytoplankton,E.
Graneli et al., Eds. (Elsevier Science, New York,
1990), pp. 29 – 40.
83. J. M. Burkholder et al., Nature 358, 407 (1992).
84. S. R. Cooper, The History of Water Quality in North
Carolina Estuarine Waters as Documented in the
Stratigraphic Record (Univ. of North Carolina Water
Resources Research Institute, Raleigh, NC, 2000).
85. J. Persson, P. Jonsson, Mar. Pollut. Bull. 40, 122
86. H. S. Lenihan, C. H. Peterson, Ecol. Appl. 8, 128
87. H. S. Lenihan, F. Micheli, S. W. Shelton, C. H. Peter-
son, Limnol. Oceanogr. 44, 910 (1999).
88. C. B. Ofﬁcer, T. J. Smayda, R. Mann, Mar. Ecol. Prog.
Ser. 9, 203 (1982).
89. J. E. Cloern, Mar. Ecol. Prog. Ser. 9, 191 (1982).
90. S. R. Carpenter et al., Ecol. Monogr. 71, 163 (2001).
91. D. Pauly, V. Christensen, J. Dalsgaard, R. Froese, F.
Torres Jr., Science 279, 860 (1998).
92. T. H. Huxley, Fish. Exhib. Lit. 4, 1 (1884).
93. H. J. Cranﬁeld, K. P. Michael, I. J. Doonan, Aquatic
Conserv. Mar. Freshw. Ecosyst. 9, 461 (1999).
94. W. M. Hochachka, A. A. Dhondt, Proc. Natl. Acad.
Sci. U. S. A. 97, 5303 (2000).
95. J. T. Carlton, J. B. Geller, Science 261, 78 (1993).
96. H. S. Lenihan, Ecol. Monogr. 69, 251 (1999).
97. J. J. Stachowicz, R. B. Whitlatch, R. W. Osman,
Science 286, 1577 (1999).
98. N. Myers, Science 269, 358 (1995).
99. N. Knowlton, Am. Zool. 32, 674 (1992).
100. S. Naeem, Conserv. Biol. 12, 39 (1998).
101. J. Munro, ICLARM (Int. Cent. Living Aquat. Resour.
Manag.) Stud. Rev. 7, 1 (1983).
102. J. L. Munro, in Fish and Fisheries Series; Reef Fisher-
ies, N. V. C. Polunin, C. M. Roberts, Eds. (Chapman &
Hall, New York, 1996), pp. 1–14.
103. M. J. Fogarty, Can. Spec. Publ. Fish. Aquat. Sci. 125,
104. C. D. Harvell et al., Science 285, 1505 (1999).
105. R. R. Colwell, Science 274, 2025 (1996).
106. R. E. Turner, N. N. Rabalais, Nature 368, 619 (1994).
107. D. Justic´, in Modern and Ancient Continental Shelf
Anoxia, R. V. Tyson, T. H. Pearson, Eds. (Geological
Society, London, 1991), pp. 95–105.
108. L. W. Botsford, J. C. Castilla, C. H. Peterson, Science
277, 509 (1997).
109. R. T. Paine, Nature 355, 73 (1992).
110. R. E. Ulanowicz, J. H. Tuttle, Estuaries 15, 298
111. M. E. Soule´, J. Terborgh, Eds., Continental Conserva-
tion (Island, Washington, DC, 1999).
112. S. M. Lewis, Ecol. Monogr. 56, 183 (1986).
113. Cod body length for samples older than 50 years was
derived from cod vertebrae collected at the Turner
Farm shell midden, North Haven, Maine. Five occupa-
tions are recognized: Occupations 1 (5300 to 4900 yr
B.P.), 2 (4500 to 4200 yr B.P.), 3 (3700 to 3500 yr B.P.),
and4(⫽ Ceramic Period, 3000 to 1000 yr B.P.), and
Plow Zone (mixed from earlier occupations but mostly
postdating 1000 yr B.P.). Regr