may result from straightforward geometrical constraints. This
variation may thus be expected in more realistic, numerical simu-
lations of the geodynamo and may provide an important constraint
on those models
12–14
. A
Received 7 November 2003; accepted 2 March 2004; doi:10.1038/nature02459.
1. Merrill, R. T. & McFadden, P. L. Geomagnetic polarity transitions. Rev. Geophys. 37, 201–226
(1999).
2. Singer, B. S. & Pringle, M. S. Age and duration of the Matuyama-Brunhes geomagnetic polarity
reversal from
40
Ar/
39
Ar incremental heating analyses of lavas. Earth Planet. Sci. Lett. 139, 47–61
(1996).
3. McElhinny, M. W. & Lock, J. IAGA paleomagnetic databases with access. Surv. Geophys. 17, 575–591
(1996).
4. Channell, J. E. T. & Lehman, B. The last two geomagnetic polarity reversals recorded in high-
deposition-rate sediment drifts. Nature 389, 712–715 (1997).
5. Yamazaki, T. & Oda, H. Orbital influence on Earth’s magnetic field; 100,000-year periodicity in
inclination. Science 295, 2435–2438 (2002).
6. Oda, H., Shibuya, H. & Hsu, V. Palaeomagnetic records of the Brunhes/Matuyama polarity transition
from ODP Leg 124 (Celebes and Sulu seas). Geophys. J. Int. 142, 319–338 (2000).
7. Fisher, R. A. Dispersion on a sphere. Proc. R. Astron. Soc. A 217, 295–305 (1953).
8. Channell, J. E. T., Hodell, D. A. & Lehman, B. Relative geomagnetic paleointensity and
d
18
Oat
ODP Site 983 (Gardar Drift, North Atlantic) since 350 ka. Earth Planet. Sci. Lett. 153, 103–118
(1997).
9. Cande, S. C. & Kent, D. V. Revised calibration of the geomagnetic polarity time scale for the Late
Cretacous and Cenozoic. J. Geophys. Res. 100, 6093–6095 (1995).
10. Holt, J. W. & Kirschvink, J. L. The upper Olduvai geomagnetic field reversal from Death Valley,
California; a fold test of transitional directions. Earth Planet. Sci. Lett. 133, 475–491 (1995).
11. Clement, B. M. & Kent, D. V. Latitudinal dependency of geomagnetic polarity transition durations.
Nature 310, 488–491 (1984).
12. Dormy, E., Valet, J.-P. & Courtillot, V. Numerical models of the geodynamo and observational
constraints. Geochem. Geophys. Geosyst. 1, doi:2000GC000062 (2000).
13. McMillan, D. G., Constable, C. G., Parker, R. L. & Glatzmaier, G. A. A statistical analysis of magnetic
fields from some geodynamo simulations. Geochem. Geophys. Geosyst.[online] 28, doi:2000GC000130
(2001).
14. Coe, R. S., Hongre, L. & Glatzmaier, G. A. An examination of simulated geomagnetic reversals from a
palaeomagnetic perspective. Phil. Trans. R. Soc. 358, 1141–1170 (2000).
15. Clement, B. M. & Kent, D. V. A southern hemisphere record of the Matuyama-Brunhes polarity
reversal. Geophys. Res. Lett. 18, 81–84 (1991).
16. Theyer, F., Herrero-Bervera, E., Hsu, V. & Hammond, S. R. The zonal harmonic model of polarity
transitions; a test using successive reversals. J. Geophys. Res. B 90, 1963–1982 (1985).
17. Valet, J.-P., Tauxe, L. & Clement, B. Equatorial and mid-latitude records of the last geomagnetic
reversal from the Atlantic Ocean. Earth Planet. Sci. Lett. 94, 371–384 (1989).
18. Clement, B. M., Kent, D. V. & Opdyke, N. D. Brunhes-Matuyama polarity transition in three deep-sea
sediment cores. Phil. Trans. R. Soc. Lond. 306, 113–119 (1982).
19. Cisowski, S. M. et al. Detailed record of the Brunhes/Matuyama polarity reversal in high
sedimentation rate marine sediments from the Isu-Bonin Arc. Proc. ODP Sci. Res. 126, 341–352
(1992).
20. Zhu, R., Laj, C. & Mazuad, A. The Matuyama-Brunhes and upper Jaramillo transitions recorded in a
loess section at Weinan, north-central China. Earth Planet. Sci. Lett. 125, 143–158 (1994).
21. Okada, M. & Niitsuma, N. Detailed paleomagnetic records during the Brunhes-Matuyama
geomagnetic reversal, and a direct determination of depth lag for magnetization in marine sediments.
Phys. Earth Planet. Inter. 56, 133–150 (1989).
22. Valet, J.-P., Tauxe, L. & Clark, C. R. The Brunhes-Matuyama transition recorded from Lake Tecopa
sediments (California). Earth Planet. Sci. Lett. 87, 463–472 (1988).
23. Clement, B. M., Kent, D. V. Geomagnetic polarity transition records from five hydraulic piston core
sites in the North Atlantic. in Init. Rep. Deep Sea Drilling Project (ed. Ruddiman, W. F. et al.) 94,
831–852 (US Government Printing Office, Washington, 1987).
24. Athanossopolous, J. A Matuyama-Brunhes Polarity Reversal Record; Comparison Between Thermal and
Alternating Field Demagnetization of Ocean Sediments from the North Pacific Transect. PhD thesis,
Univ. California, Santa Barbara, 225 (1993).
25. Herrero-Bervera, E. & Theyer, F. Non-axisymmetric behaviour of Olduvai and Jaramillo polarity
transitions recorded in north-central Pacific deep-sea sediments. Nature 322, 159–162 (1986).
26. Clement, B. M. & Kent, D. V. A detailed record of the lower Jaramillo polarity transition from a
southern hemisphere, deep-sea sediment core. J. Geophys. Res. 89, 1049–1058 (1984).
27. Gee, J. S. et al. Lower Jaramillo polarity transition records from the equatorial Atlantic and Indian
oceans. Proc. ODP Sci. Res. 121, 377–391 (1991).
28. Clement, B. M. & Kent, D. V. A comparison of two sequential geomagnetic polarity transitions (upper
Olduvai and lower Jaramillo) from the Southern Hemisphere. Phys. Earth Planet. Inter. 39, 301–313
(1985).
29. Herrero-Bervera, E. & Khan, M. A. Olduvai termination; detailed palaeomagnetic analysis of a north
central Pacific core. Geophys. J. Int. 108, 535–545 (1992).
Acknowledgements R. Coe, D. V. Kent, J. Dauphin & B. Midson provided comments that
improved the manuscript. J. E. T. Channell and T. Yamazaki provided their data for this
work.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to B.M.C. (clementb@fiu.edu).
..............................................................
Effectiveness of the global
protected area network in
representing species diversity
Ana S. L. Rodrigues
1
, Sandy J. Andelman
3
, Mohamed I. Bakarr
4
,
Luigi Boitani
5
, Thomas M. Brooks
1
, Richard M. Cowling
6
,
Lincoln D. C. Fishpool
7
, Gustavo A. B. da Fonseca
1,8
, Kevin J. Gaston
9
,
Michael Hoffmann
1
, Janice S. Long
2
, Pablo A. Marquet
10
,
John D. Pilgrim
1
, Robert L. Pressey
11
, Jan Schipper
12
, Wes Sechrest
2
,
Simon N. Stuart
2
, Les G. Underhill
13
, Robert W. Waller
1
,
Matthew E. J. Watts
14
& Xie Yan
15
1
Center for Applied Biodiversity Science, and
2
IUCN-SSC/CI-CABS Biodiversity
Assessment Unit, Conservation International, Washington, DC 20036, USA
3
National Center for Ecological Analysis and Synthesis, University of California,
Santa Barbara, California 93101, USA
4
World Agroforestry Centre (ICRAF), Gigiri Nairobi, Kenya
5
Dipartimento di Biologia Animale e dell’Uomo, Universita
`
di Roma
‘La Sapienza’, 00185 Rome, Italy
6
Terrestrial Ecology Research Unit, Department of Botany, University of Port
Elizabeth, Port Elizabeth 6000, South Africa
7
BirdLife International, Cambridge CB3 0NA, UK
8
Departmento de Zoologia, Universidade Federal de Minas Gerais, Belo Horizonte
31270, Brazil
9
Biodiversity and Macroecology Group, Department of Animal and Plant Sciences,
University of Sheffield, Sheffield S10 2TN, UK
10
Center for Advanced Studies in Ecology and Biodiversity (CASEB) and
Departamento de Ecologı
´
a, Facultad de Ciencias Biolo
´
gicas, Pontificia
Universidad Cato
´
lica de Chile, Casilla 114-D, Santiago, Chile
11
New South Wales Department of Environment and Conservation, Armidale,
New South Wales 2350, Australia
12
Department of Fish and Wildlife Resources, University of Idaho, Moscow,
Idaho 83844, USA
13
Avian Demography Unit, Department of Statistical Sciences, University of Cape
Town, Rondebosch 7701, South Africa
14
194W Hill Street, Walcha, New South Wales 2354, Australia
15
Institute of Zoology, Chinese Academy of Sciences, Beijing 100080, China
.............................................................................................................................................................................
The Fifth World Parks Congress in Durban, South Africa,
announced in September 2003 that the global network of pro-
tected areas now covers 11.5% of the planet’s land surface
1
. This
surpasses the 10% target proposed a decade earlier, at the Caracas
Congress
2
, for 9 out of 14 major terrestrial biomes
1
.Such
uniform targets based on percentage of area have become deeply
embedded into national and international conservation plan-
ning
3
. Although politically expedient, the scientific basis and
conservation value of these targets have been questioned
4,5
.In
practice, however, little is known of how to set appropriate
targets, or of the extent to which the current global protected
area network fulfils its goal of protecting biodiversity. Here, we
combine five global data sets on the distribution of species and
protected areas to provide the first global gap analysis assessing
the effectiveness of protected areas in representing species
diversity. We show that the global network is far from complete,
and demonstrate the inadequacy of uniform
—
that is, ‘one size fits
all’
—
conservation targets.
Systematic approaches to conservation planning have been devel-
oped over the last two decades to guide the efficient allocation of the
scarce resources available for protecting biodiversity
6
. Gap analysis
is a planning approach based on assessment of the comprehensive-
ness of existing protected area networks and identification of gaps in
coverage
7,8
. It has also been developed into a formal method
now applied by the US Geological Survey National Gap Analysis
Program
9
and others. Numerous gap analyses at regional scales
reveal that coverage of biodiversity by existing networks of pro-
tected areas is inadequate
10,11
. Furthermore, many such networks are
letters to nature
NATURE | VOL 428 |8 APRIL 2004 | www.nature.com/nature640
© 2004
Nature
Publishing
Group
skewed towards particular ecosystems, often those that are less
economically valuable, leaving others inadequately protected
12
.At
the global scale, however, the degree to which biodiversity is
represented within the existing network of protected areas is
unknown.
In this analysis, we considered a species to be a ‘covered species’
if any protected area overlapped any extent of its mapped distri-
bution, and otherwise to be a ‘gap species’. Overall, 1,424 gap species
(12% of all species analysed) were identified (Table 1). Protected
areas may not retain all of their species if they are too small to
maintain viable populations
13
or if they are used extractively
14
.Of
the covered species, 1,423 were not represented in any protected
area larger than 1,000 ha and in stricter conservation classifications
(The World Conservation Union (IUCN) categories I–IV
15
).
Threatened and restricted-range species are those of most conserva-
tion concern
16–18
. Sets of species with smaller median range sizes
tend to have a higher proportion of gap species (Table 1). Hence,
amphibians are the least represented taxon and, within any given
taxon, threatened species (which tend to have smaller ranges) have
proportionally higher numbers of gap species than do all species
considered together. Overall, 20% of all threatened species analysed
were identified as gap species.
The number of covered species is an overestimate, mainly because
of two unrealistic assumptions. First, all protected areas are con-
sidered to be adequate for protecting every species, whereas in
reality even those classified in IUCN categories I–IV vary substan-
tially in the degree of effectiveness and enforcement
19
. Second, it
assumes that species can be protected equally effectively in any part
of their range, regardless of habitat suitability, and by the protection
of any fraction of that range, regardless of viability constraints. In
practice, simple presence within a protected area is insufficient to
ensure the long-term persistence of many species, particularly those
with demanding habitat or area requirements
13
, and does not
consider threats such as global climate change
20
.
As species are only considered to be gap species if they are not
touched by any protected area, concentrations of gap species in a
Table 1 Numbers of gap species in the current protected area network and in randomly selected networks
Taxon Median range size
(km
2
)
Numbers of gap species
Current network
(all PAs)
Current network
(PAs .1,000 ha and IUCN I–IV)
Model I
(equal area sites)
Model II
(variable area sites)
Model III
(tropical bias)
...................................................................................................................................................................................................................................................................................................................................................................
All species
Mammals (n ¼ 4,735) 247,341 258 (5.5%) 644 (13.5%) 297.7 (6.3%) 342.3 (7.2%) 226.6 (4.8%)
Turtles (n ¼ 273) 309,172 21 (7.7%) 48 (17.6%) 24.6 (9.0%) 26.5 (9.7%) 23.8 (8.7%)
Amphibians (n ¼ 5,454) 7,944 913 (16.7%) 1,718 (31.5%) 1,230.2 (22.6%) 1,507.8 (27.7%) 804.2 (14.7%)
Threatened species
Mammals (n ¼ 1,063) 22,902 149 (14.0%) 314 (29.6%) 191.8 (18.0%) 218.2 (20.5%) 151.6 (14.3%)
Birds (n ¼ 1,171) 4,015 232 (19.8%) 437 (37.3%) 349.7 (29.9%) 409.6 (35.0%) 275.4 (23.5%)
Turtles (n ¼ 119) 167,611 12 (10.1%) 32 (26.9%) 15.9 (13.3%) 17.3 (14.6%) 15.5 (13.1%)
Amphibians (n ¼ 1,543) 896 411 (26.6%) 767 (49.7%) 604.5 (39.2%) 740.0 (48.0%) 423.3 (27.4%)
All species analysed
(n ¼ 11,633) 38,229 1,424 (12.2%) 2,847 (24.5%) 1,902.3 (16.4%) 2,286.2 (19.7%) 1,330.3 (11.4%)
...................................................................................................................................................................................................................................................................................................................................................................
The total numbers of species and their respective median range size are given for comparative purposes. Values in parentheses are the percentage of all species/threatened species analysed within a given
taxon. PAs, protected areas.
Figure 1 Density map of gap species per half-degree cell, created by overlaying the ranges of all species not covered by any protected area.
letters to nature
NATURE | VOL 428 |8 APRIL 2004 | www.nature.com/nature 641
© 2004
Nature
Publishing
Group
given region may be explained by sparse protected area coverage
and/or by a concentration of narrowly distributed species. The
global distribution of gap species (Fig. 1) is influenced more
strongly by the latter. Indeed, within a given biome
21
, the percentage
of species that are gaps is highly significantly correlated with the
level of endemism, independent of the percentage of area protected
(Fig. 2a, b). Across countries, the percentage of gap species decreases
with percentage of area protected, but is more strongly correlated
with levels of national endemism (Fig. 2c, d). Consequently,
although in some regions the absence of protected areas allows for
relatively widespread gap species (notably in Somalia), the map of
gap species mainly reflects the presence of narrowly distributed
species (Fig. 1). The regions highlighted include many widely
recognized centres of endemism
16,22
, such as Yunnan province and
the mountains surrounding the Sichuan basin in southern China,
the Western Ghats of India, Sri Lanka, the islands of Southeast Asia
and Melanesia, the Pacific islands, Madagascar, the Cameroon
highlands, Mesoamerica, the tropical Andes, the Caribbean, and
the Atlantic Forest of South America. Most of these are montane or
insular regions in the tropics.
These results have implications for global conservation planning
strategies, as they clearly demonstrate that the percentage of area
already protected in a given country or biome is a very poor
indicator of additional conservation needs. Contrary to frequent
recommendations
1,23
, current protection levels should not be used
as a significant criterion to guide priorities for allocation of future
conservation investments. Indeed, the regions with greatest need for
expansion of the global protected area network are not necessarily
those with a lower percentage of their area protected; rather, they
typically are those with higher levels of endemism
24
. Conversely,
uniform targets based on percentage of area protected (except for
100%) cannot be used as a ceiling to distinguish between regions
sufficiently protected and those that need additional protection
4–5
.
Global conservation strategies based on the recommendation
that 10% (or other similar targets) of each country or biome be
protected will not be effective because they are blind to the fact that
biodiversity is not evenly distributed across the planet
25
; by the same
token, neither should protected areas be. Indeed, a network with the
same total area as the existing one but evenly distributed across the
world would perform less adequately than the current network in
representing species of mammals, amphibians, turtles and threat-
ened birds (Table 1). The better performance of the current network
indicates uneven distribution of protected areas relative to bio-
diversity pattern. Indeed, the current network is significantly (albeit
not overwhelmingly) biased towards sites with higher richness of all
species, restricted-range species and threatened species. This may be
the legacy of decisions to locate some protected areas in better
sites, and/or be symptomatic of higher levels of biodiversity loss
outside protected areas
19
. Nonetheless, the current global network
could still perform better in terms of species coverage. For example,
a network biased towards the tropics (to match their higher level
of endemism) would have fewer gap species than the current
network, and far fewer gap species than a random unbiased network
(Table 1).
Our results demonstrate that if the conservation goal is species
representation, then the expansion of the global network of pro-
tected areas must account for biodiversity patterns, rather than rely
on general percentage-based targets that are formed largely by
political and feasibility considerations
4–5
. Given the increasing
threats to biodiversity, such expansion should be made strategically
by focusing on those regions that would contribute most to the
global system and prioritizing, within those, the regions where
the urgency for conservation action is greatest
22
. Conservation
strategies must also address the complexity of natural ecosystems,
including genetic and phylogenetic diversity, and ecological and
evolutionary processes
26
.
The existing protected area network provides an invaluable
service in shielding habitat from destructive use and hence in
reducing biodiversity loss
19
. However, our global gap analysis clearly
demonstrates that the global protected area network is still far from
complete, even for terrestrial vertebrates, the best known and most
popular of all species groups
27
. Of the species considered, at least
12% are not represented in any protected area, despite the extremely
strict assumptions applied for identifying gap species. It is likely that
other taxa with high levels of endemism, such as plants and insects,
are even less well represented, given the tendency for sets of species
with smaller range sizes to have higher proportions of gap species.
Protected areas are not the only tactic available to conservation
planners, but they are highly cost effective in protecting biodiver-
sity
28
. Advances in data availability and in the science of conserva-
tion planning enable us to act strategically in the face of increasing
human pressure. Clearly, the task ahead is as urgent as it is
challenging, as much biodiversity remains to be protected. A
Methods
Data
Data on the global distribution of protected areas were obtained from the 2003 World
Database on Protected Areas
29
. Distribution maps were obtained for 11,633 species of
terrestrial vertebrates: 4,735 terrestrial mammals, compiled by the IUCN Global Mammal
Assessment; 1,171 globally threatened birds
17
; 273 freshwater turtles and tortoises
30
;and
5,454 amphibians, compiled by the IUCN Global Amphibian Assessment. These species
data also include assessments of conservation status, with 1,063 mammals, 1,171 birds,
119 turtles and 1,543 amphibians having been listed as globally threatened by the IUCN
Red List
17–18
. See Supplementary Information for more details.
Randomly distributed networks
Two null models were created to simulate a network of protected areas with similar
characteristics to the existing one, but evenly spread around the world: Model I (equal area
sites), 69,794 circles, of the same size as the mean area of a protected site, and 11,119 points
were randomly spread around the world’s land surface (excluding Antarctica); Model II
(variable area sites), 69,794 circles, with the same distribution of sizes as the current
protected area network, and 11,119 points were randomly spread around the world’s land
surface (excluding Antarctica).
Of all species that are restricted to either the tropical or the non-tropical regions (that
is, excluding species that span both), 75.8% are found in the tropics; however, only 45.8%
of the global protected area network is in the tropics. Therefore, we considered a third
model in which the percentage of the global protected area in the tropics was increased to
match its level of endemism: Model III (tropical bias), 69,794 circles, of the same size as the
mean area of a protected site, and 11,119 points were distributed such that 75.8% of each
occurred in the topics, having random distributions within tropical and non-tropical
areas.
Figure 2 Percentage of gap species in relation to endemism levels and percentage of area
protected across biomes and countries. a–d, Relationships between: percentage of
species in each biome that are endemic and percentage that are gap species (a)(n ¼ 16,
r ¼ 0.72, P , 0.005); percentage of each biome’s protected area and percentage of
gap species (b)(P . 0.5); percentage of species in each country that are endemic and
percentage that are gap species (c)(n ¼ 247, r ¼ 0.69, P , 0.001); percentage of
each country’s protected area and percentage of gap species (d)(n ¼ 247, r ¼ 0.15,
P , 0.05).
letters to nature
NATURE | VOL 428 |8 APRIL 2004 | www.nature.com/nature642
© 2004
Nature
Publishing
Group
Sixty replicates were obtained for each of these randomly distributed networks. These
were then overlaid with species distributional data to analyse the number of gap species in
each case. See Supplementary Information for the confidence intervals for each of the
models.
Richness of protected and unprotected cells
The richness of each quarter-degree cell touching land (outside Antarctica) was calculated
for all species, restricted-range species
16
(occupying #50,000 km
2
) and threatened species.
Cells touching protected areas were considered ‘protected’. Protected cells are significantly
(P , 0.001) biased towards higher richness of all, restricted-range and threatened species.
See Supplementary Information for a comparison of frequency distributions.
Received 21 December 2003; accepted 11 February 2004; doi:10.1038/nature02422.
1. Chape, S., Fish, L., Fox, P. & Spalding, M. United Nations List of Protected Areas (IUCN/UNEP, Gland,
Switzerland/Cambridge, UK, 2003).
2. The World Conservation Union. Parks For Life: Report of the IVth World Congress on National Parks
and Protected Areas (IUCN, Gland, Switzerland, 1993).
3. Kamden-Toham, A. et al. Forest conservation in the Congo Basin. Science 299, 346 (2003).
4. Soule
´
, M. E. & Sanjayan, M. A. Conservation targets: do they help? Science 279, 2060–2061 (1998).
5. Pressey, R. L., Cowling, R. M. & Rouget, M. Formulating conservation targets for biodiversity pattern
and process in the Cape Floristic Region, South Africa. Biol. Conserv. 112, 99–127 (2003).
6. Margules, C. R. & Pressey, R. L. Systematic conservation planning. Nature 405, 243–253 (2000).
7. Scott, J. M. et al. Gap analysis
—
a geographic approach to protection of biological diversity. Wildl.
Monogr. 123, 1–41 (1993).
8. Lacher, T. E. Jr in GIS Methodologies for Developing Conservation Strategies (eds Savitsky, B. G. &
Lacher, T. E. Jr) 199–209 (Columbia Univ. Press, New York, 1998).
9. Jennings, M. D. Gap analysis: Concepts, methods, and recent results. Landscape Ecol. 15, 5–20
(2000).
10. Scott, J. M. et al. Nature reserves: Do they capture the full range of America’s biological diversity? Ecol.
Appl. 11, 999–1007 (2001).
11. Andelman, S. J. & Willig, M. R. Present patterns and future prospects for biodiversity in the Western
Hemisphere. Ecol. Lett. 6, 818–824 (2003).
12. Pressey, R. L. Ad hoc reservations
—
Forward or backward steps in developing representative reserve
systems? Conserv. Biol. 8, 662–668 (1994).
13. Newmark, W. D. Insularization of Tanzanian parks and the local extinction of large mammals.
Conserv. Biol. 10, 1549–1556 (1996).
14. Peres, C. A. & Lake, I. R. Extent of nontimber resource extraction in tropical forests: accessibility to
game vertebrates by hunters in the Amazon basin. Conserv. Biol. 17, 521–535 (2003).
15. The World Conservation Union, Guidelines for Protected Area Management Categories (IUCN
CNPPA/WCMC, Gland, Switzerland/Cambridge, UK, 1994).
16. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. Endemic Bird Areas of the World
—
Priorities
for Biodiversity Conservation (BirdLife International, Cambridge, UK, 1998).
17. BirdLife International. Threatened Birds of the World (Lynx Edicions/BirdLife International,
Barcelona/Cambridge, UK, 2000).
18. The World Conservation Union. IUCN Red List of Threatened Species [online] khttp://
www.redlist.orgl (2003).
19. Bruner, A. G., Gullison, R. E., Rice, R. E. & Fonseca, G. A. B. Effectiveness of parks in protecting
tropical biodiversity. Science 291, 125–128 (2001).
20. Thomas, C. D. et al. Extinction risk from climate change. Nature 427, 145–148 (2004).
21. Olson, D. M. et al. Terrestrial ecoregions of the world: A new map of life on earth. Bioscience 51,
933–938 (2001).
22. Myers, N., Mittermeier, R. A., Mittermeier, C. G., Fonseca, G. A. B. & Kent, J. Biodiversity hotspots for
conservation priorities. Nature 403, 853–858 (2000).
23. Green, M. J. B. & Paine, J. State of the World’s Protected Areas at the End of the Twentieth Century
(WCMC, Cambridge, UK, 1997).
24. Rodrigues, A. S. L. & Gaston, K. J. How large do reserve networks need to be? Ecol. Lett. 4, 602–609
(2001).
25. Gaston, K. J. Global patterns in biodiversity. Nature 405, 220–227 (2000).
26. Cowling, R. M. & Pressey, R. L. Rapid plant diversification: planning for an evolutionary future. Proc.
Natl Acad. Sci. USA 98, 5452–5457 (2001).
27. Gaston, K. J. & May, R. M. Taxonomy of taxonomists. Nature 356, 281–282 (1992).
28. Balmford, A. et al. Economic reasons for conserving wild nature. Science 297, 950–953 (2002).
29. World Database on Protected Areas. World Database on Protected Areas (IUCN-WCPA/UNEP-
WCMC, Washington DC, 2003).
30. Iverson, J. B., Kiester, A. R., Hughes, L. E. & Kimerling, A. J. The EMYSystem World Turtle Database
2003 [online] khttp://emys.geo.orst.edul (2003).
Supplementary Information accompanies the paper on www.nature.com/nature.
Acknowledgements We thank the Moore Family Foundation, the Howard Gilman Foundation
and the National Center for Ecological Analysis and Synthesis of the University of California Santa
Barbara for support. The analysis was possible thanks to the combined effort of the thousands of
individuals and hundreds of institutions who collected and compiled the data, or provided
financial support for such efforts. We are grateful to the numerous individuals who contributed to
this analysis, especially to K. Buhlmann, S. Butchart, N. Cox, P. P. van Dijk, J. Iverson, R. Kiester,
T. Lacher and B. Young. H. Possingham made valuable comments on the manuscript. Figure 1 was
generated by J. Seeber.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to A.S.L.R.
(a.rodrigues@conservation.org).
..............................................................
Spatial structure often inhibits
the evolution of cooperation
in the snowdrift game
Christoph Hauert & Michael Doebeli
Departments of Zoology and Mathematics, University of British Columbia,
6270 University Boulevard, Vancouver, British Columbia V6T 1Z4, Canada
.............................................................................................................................................................................
Understanding the emergence of cooperation is a fundamental
problem in evolutionary biology
1
. Evolutionary game theory
2,3
has become a powerful framework with which to investigate this
problem. Two simple games have attracted most attention in
theoretical and experimental studies: the Prisoner’s Dilemma
4
and the snowdrift game (also known as the hawk–dove or chicken
game)
5
. In the Prisoner’s Dilemma, the non-cooperative state is
evolutionarily stable, which has inspired numerous investi-
gations of suitable extensions that enable cooperative behaviour
to persist. In particular, on the basis of spatial extensions of the
Prisoner’s Dilemma, it is widely accepted that spatial structure
promotes the evolution of cooperation
6–8
. Here we show that no
such general predictions can be made for the effects of spatial
structure in the snowdrift game. In unstructured snowdrift
games, intermediate levels of cooperation persist. Unexpectedly,
spatial structure reduces the proportion of cooperators for a wide
range of parameters. In particular, spatial structure eliminates
cooperation if the cost-to-benefit ratio of cooperation is high.
Our results caution against the common belief that spatial
structure is necessarily beneficial for cooperative behaviour.
The Prisoner’s Dilemma illustrates that cooperating individuals
are prone to exploitation, and that natural selection should favour
cheaters. In this game, two players simultaneously decide whether to
cooperate or defect. Cooperation results in a benefit b to the
recipient but incurs a cost c to the donor (b . c . 0). Mutual
cooperation thus pays a net benefit of R ¼ b 2 c, whereas mutual
defection results in payoff P ¼ 0 for both players. With unilateral
cooperation, defection yields the highest payoff, T ¼ b,atthe
expense of the cooperator bearing the cost S ¼ 2c. It follows that
it is best to defect regardless of the co-player’s decision. Thus,
defection is the evolutionarily stable strategy, even though all
individuals would be better off if they all cooperated. This outcome
is a simple consequence of the ranking of the four payoff values:
T . R . P . S. Despite this seemingly convincing argument,
many natural species show altruism, with individuals bearing
costs to the benefit of others: vampire bats share blood
9
, alarm
calls warn from predators
10
, monkeys groom each other
11
, and fish
inspect predators preferably in pairs
12
.
In field and experimental studies it is often difficult to assess the
fitness payoffs for different behavioural patterns, and even the
proper ranking of the payoffs is challenging
13,14
. This has led to a
considerable gap between theory and experimental evidence, and to
an increasing discomfort with the Prisoner’s Dilemma as the only
model to discuss cooperative behaviour
15,16
. The snowdrift game is a
viable and biologically interesting alternative. It differs from the
Prisoner’s Dilemma in that the payoffs P and S have a reverse order:
T . R . S . P. This changes the situation fundamentally and
leads to persistence of cooperation.
To illustrate the snowdrift game, imagine two drivers that are
caught in a blizzard and trapped on either side of a snowdrift. They
can either get out and start shovelling (cooperate) or remain in the
car (defect). If both cooperate, they have the benefit b of getting
home while sharing the labour c. Thus, R ¼ b 2 c/2. If both defect,
they do not get anywhere and P ¼ 0. If only one shovels, however,
they both get home but the defector avoids the labour cost and gets
letters to nature
NATURE | VOL 428 |8 APRIL 2004 | www.nature.com/nature 643
© 2004
Nature
Publishing
Group
