Download species–area relationship

FURTHER READING
Coleman, P. J. . Stratigraphical and structural notes on the British
Solomon Islands with reference to the first geological map. British Solomon Islands Geological Record (1959–1962) : –.
Hughes, G. W., P. M. Craig, and R.A. Dennis. . Geology of the outer
Eastern Islands. Geological Survey Division Solomon Islands Bulletin .
Kroenke, L. W. . Solomon Islands: San Cristobal to Bougainville
and Buka, in Cenozoic tectonic development of the southwest Pacific. L.
W. Kroenke, ed. Technical Bulletin . Suva, Fiji: Committee for Coordination of Joint Prospecting for Mineral Resources in South Pacific
Offshore Areas, ch .
Mahoney, J., G. Fitton, P. Wallace, and the Leg  Scientific Party. .
ODP Leg : basement drilling on the Ontong Java Plateau. JOIDES
Journal .: – and covers.
Reagan, A. J., and H. M. Griffin. . Bougainville before the conflict.
Canberra: Pandanus Books/ANU.
Vedder, J. G., and T. S. Bruns. . The geology and offshore resources of Pacific
island arcs: Solomon Islands and Bougainville. Earth Science Series . Houston: Circum-Pacific Council for Energy and Mineral Resources.
Vedder, J. G., K. S. Pound, and S. Q. Boundy, eds. . The geology
and offshore resources of Pacific island arcs: central and western Solomon
Islands. Earth Science Series . Houston: Circum-Pacific Council for
Energy and Mineral Resources.
REFERENCES
Davies, H. L., et al. . Geology of Oceania, in Encyclopedia of Geology, vol. 4. R. C. Selley, L. R. M. Cocks, and I. R. Plimer, eds. Oxford:
Elsevier, –.
Smith, W. H. F., and D. T. Sandwell. . Global seafloor topography from
satellite altimetry and ship depth soundings. Science : –.
SOUTH GEORGIA
SEE ATLANTIC REGION
SOUTH SANDWICH ISLANDS
SEE ATLANTIC REGION
FIGURE 1 A good example of the species–area relationship is the num-
ber of land plant species on the Galápagos Islands. Top: Map of the
Galápagos Islands showing the number of land plant species. Note that
the largest island contains the most species and that numbers tend to
be higher on large islands than on small islands. However, several devi-
SPECIES–AREA
RELATIONSHIP
ations from this pattern exist, such as the “unexpected” high number
of species (319) on the medium-sized island at the southern end of the
archipelago, suggesting that factors other than area may sometimes
be important in determining the number of species on islands. Bottom:
Relationship between the number of species and island area. Regres-
DAVID A. SPILLER AND THOMAS W. SCHOENER
University of California, Davis
For over a century, ecologists have been captivated by the
tendency for number of species within a taxonomic group
to increase with island area. This “species–area relationship” has been found for a broad range of organisms in
numerous archipelagoes around the world. A partial list of
studies demonstrating the relationship includes land plants
on the Galápagos (Fig. ) and Aleutian Islands; insects on
sion line is the least-squares estimate: log(number of species) = 1.46 +
0.33 log(area). Data taken from Hamilton et al. (1964).
the Tuscan Islands and on subantarctic islands; reptiles on
islands in the Gulf of California and in the West Indies;
birds on the Canary, Solomon, and Aegean Islands; and
mammals on islands in the Philippines and North American Great Lakes. Although most studies are of higher
organisms, even protozoans and diatoms have shown the
relationship. To explain the occurrence of the species–area
relationship, ecologists have proposed several hypotheses
S P E C I E S – A R E A R E L AT I O N S H I P
Gillespie08_S.indd 857
857
4/23/09 12:58:39 PM
on causal mechanisms, which shed light on the processes
that structure biological communities. Conservation biologists have applied the relationship to the design of nature
reserves.
STATISTICAL MODELS
The species-area relationship can be estimated statistically
using the power equation
S = k Az
()
where S = number of species, A = area of island, and k
and z are fitted parameters. Logarithmic transformation of both sides of the power equation yields the linear
equation
log S = log k + z log A
()
in which z is the slope of the estimated species–area relationship (i.e., the rate of increase of log S with log A).
Because the scales are logarithmic, the z-value is independent of scale, allowing comparisons between different studies even when the units for area are different. A
second equation, with S instead of log S in Equation , is
often used instead.
Log-transformed data on number of land plant species
and area for the Galápagos Islands (Fig. , bottom) show
a positive linear relationship with a slope (z-value) of ..
Similarly, many other studies have found that the data fit
the log-log model with z-values usually ranging from .
to .. Preston developed a mathematical explanation
for the prevalence of the linear log-log relationship, with
specific assumptions about the distribution of species
abundances and other biological processes, in which the
z-value is expected to be approximately .. However,
a detailed analysis of  studies by Connor and McCoy
showed that the data often fit other statistical models just
as well or better, suggesting that biological interpretation of parameters in statistical models should be made
cautiously.
viduals, then the number of species should increase with
island area. An analogy would be a bowl of colored marbles
with ten marbles of each of ten different colors. The larger
the handful of marbles you take, the more different colors
you will get on average. This simple explanation may serve
as a null model to compare with the other five hypotheses,
all of which incorporate biological processes.
Larger Populations and Less Extinction on Larger
Islands
The hypothesis that populations are larger on larger
islands, implying lower extinction rates, is an integral part
of the equilibrium theory of island biogeography developed by MacArthur and Wilson in . According to
this theory, the number of species present on an island
is determined by the dynamic equilibrium between the
rate of immigration of species not already on the island
from a source pool and the rate of extinction of species
already on the island (Fig. A). Because larger islands
WHY DOES THE NUMBER OF SPECIES
INCREASE WITH ISLAND AREA?
Although the species–area relationship is one of the surest
generalizations in ecology, there has been much debate
on the causal factors and processes. There are six major
hypotheses, as follows.
FIGURE 2 Graphical representations of the MacArthur–Wilson equi-
librium model of island biogeography. (A) The effect of island area;
Random Sampling
Assuming that the individuals on a given island are a
random sample from a nearby mainland or other source
containing all species, and larger islands contain more indi-
858
Gillespie08_S.indd 858
S = equilibrium number of species on a small island, L = equilibrium
number of species on a large island, P = number of species in the pool
on the mainland (source). (B) The effect of distance of the island from
the source of immigration; N = equilibrium number of species on a near
island, F = equilibrium number of species on a far island.
S P E C I E S – A R E A R E L AT I O N S H I P
4/23/09 12:58:42 PM
contain more individuals (larger population size) in each
species, and extinction rate is inversely proportional to
population size, extinction rate of species on small islands
is higher than on large islands, so the equilibrium number
of species is positively related to island area. In  Simberloff provided experimental evidence for this hypothesis
by showing that the number of species on mangrove islets
decreased when he reduced the area of islets. In addition
to the area effect, the equilibrium theory contains a distance effect; immigration rates are higher for islands near
the source pool than for those farther away (Fig. B).
Larger Interception Area and More Immigration
on Larger Islands
In addition to the lower extinction rate in the equilibrium model, a larger island may also be a bigger “target”
for colonists; this causes the immigration curve of a large
island to be higher than that of a small island, thereby
increasing the equilibrium for the large island.
Higher Habitat Diversity on Larger Islands
Because the number of different types of habitats may
increase with island area and different species often occur
in different habitats, more species occur on larger islands.
Hence, the species-area relationship may be indirect via the
positive effect of area on habitat diversity. This hypothesis
is most popular among ecologists familiar with the natural histories of the species in their studies. Early on, Watson found that habitat diversity was a better predictor of
Aegean bird species number than by area. Very recently,
Morrison found that the number of ant species on islets in
the Bahamas was predicted by number of land plants species better than was area; in this case, different plant species
may serve as different types of habitats or provide different
types of food resources for ants. Returning to Galápagos
land plants, Hamilton and collaborators showed in 
that elevation (a proxy for habitat diversity) was a better
predictor of number of species than was area.
rium as in the MacArthur–Wilson model. In addition to
the effect of episodic major disturbances, chronic noncatastrophic disturbances, such as wind and salt spray,
may affect small islands more than large ones because of
the greater perimeter-area ratio on small islands. Only
species that can tolerate these harsh conditions could persist on small islands.
Greater Speciation on Larger Islands
The larger the island, the more likely that geographic barriers exist that cause isolation between viable subpopulations of a species; this can lead to allopatric speciation.
The effect may be geometric, as the opportunity for speciation is itself proportional to species number, causing an
especially high species–area slope.
Lomolino’s Combined Model
Lomolino proposed in  a general model for the species–
area relationship that integrates disturbance and equilibrium
dynamics along with speciation on islands. In this model the
species–area relationship has three regions (Fig. ): () On
small islands, the stochastic effects of disturbances are predominant, making the relationship between number of species and island area variable and unpredictable (an example
is given in the later discussion of spiders in the Bahamas). ()
On medium to large islands, the deterministic effects of area
on extinction/immigration and habitat diversity predominate, and number of species increases at a decreasing rate as
the number of species in the source pool is approached. ()
The largest islands contain barriers isolating species populations, leading to allopatric speciation. Losos and Schluter
assessed the speciation component for Anolis lizards of the
West Indies. For these relatively poor dispersers, the sharp
increase in the species–area slope due to speciation obliterates the flattening portion (region ) of Lomolino’s model.
Lower Abiotic Disturbance on Larger Islands
The impact of abiotic disturbances, such as hurricanes
and tsunamis, may be more devastating on small islands
than on large islands, exterminating species mostly on the
former and thereby causing the positive species–area relationship. Such an effect could go well beyond the effect
of population size on extinction, as equivalently sized
populations could be more vulnerable on smaller islands.
Whittaker pointed out that such disturbances can affect
island communities for decades or even centuries, during
which time species would not be at a dynamic equilib-
FIGURE 3 Lomolino’s general species–area model containing three
different regions (see text for explanation). Modified from Lomolino
and Weisen (2001).
S P E C I E S – A R E A R E L AT I O N S H I P
Gillespie08_S.indd 859
859
4/23/09 12:58:43 PM
ECOLOGICAL CORRELATES OF VARIATION IN
THE SPECIES–AREA SLOPE
Comparisons among different kinds of organisms have
revealed substantial variation in the slope of the species–area relationship. Wright showed that on islands
in the West Indies, the slope was greater for non-flying
mammals than for bats or birds, suggesting that dispersal
ability influences the relationship. A plausible explanation is that islands are more isolated for species with low
dispersal ability, so most of them do not occur on small
islands, whereas species with high dispersal ability are frequently found on small islands because their immigration
rates are high even though they may not persist for long.
Schoener suggested that z-values should be greater for species occurring at low density (e.g., territorial carnivores)
than for those occurring at high density because the lowdensity species can only persist on larger islands. Holt and
collaborators developed a mathematical model predicting
that the z-value for species at the top of the food chain is
higher than for species in lower trophic levels.
SPIDERS ON BAHAMIAN ISLANDS
Studies of web spiders occurring on islands in the Bahamas by Schoener and Spiller illustrate the species–area
relationship and address several of the issues discussed
above (Fig. ). Complete censuses of all web spiders
occurring over the entire areas of  islands were conducted annually over a ten-year period. Number of
species (mean over time) was positively correlated with
island area for four reasons. First, larger islands tended to
have larger populations, which in turn had lower extinction rates than did small populations, increasing their
number of species. Second, larger islands have a higher
immigration rate: the “target effect.” Third, some species
are habitat generalists (e.g., Argiope argentata), occurring
in areas with high or low vegetation, whereas others are
more specialized (e.g., Gasteracantha cancriformis), living
in only high vegetation (Fig. ). Small islands tend to
have only low vegetation, whereas large islands contain
areas with both low and high vegetation. Therefore, only
FIGURE 5 (A) Argiope argentata. (B) Gasteracantha cancriformis.
Photographs by D. Spiller.
generalists occur on small islands, whereas both generalists and specialists occur on large islands with more types
of habitats. Fourth, tropical storms can affect smaller
islands more drastically: During several recent hurricanes, smaller Bahamian islands, which are lower, were
completely inundated by high water, apparently killing
all spiders. Note that variation in the numbers of spider
species on small islands is relatively high as in Lomolino’s
model. Another factor that can affect number of species
is the presence of lizards, which are major predators of
spiders and which occurred on about half of the study
islands: Islands with lizards tended to have fewer spider
species (Fig. ). This “lizard effect” is more apparent for
larger islands than for smaller islands, and the slope of
the species-area relationship is greater for islands without
lizards (Fig. ). Hence, the z-value is greater for islands
without lizards, on which spiders occupy a higher level in
the food web, as in Holt’s model.
FIGURE 4 Aerial photograph showing a portion of the Bahamian spi-
der study area.
860
Gillespie08_S.indd 860
FIGURE 6 Illustration of a lizard eating a spider by G. Dan.
S P E C I E S – A R E A R E L AT I O N S H I P
4/23/09 12:58:44 PM
hood, may be increased with a certain amount of reservearea fragmentation. In conclusion, several small preserves
distributed over a large or varied region may sometimes
be the better conservation strategy.
SEE ALSO THE FOLLOWING ARTICLES
Extinction / Fragmentation / Galápagos Islands, Biology / Island
Biogeography, Theory of / Spiders
FURTHER READING
FIGURE 7 Species–area relationship for Bahamian orb spiders on 37
islands without lizards and 27 islands with lizards. Regression lines
for islands without and with lizards are respectively: log(number of
species) = -0.85 + 0.41log(area), log(number of species) = -0.28 +
0.16log(area). Unpublished data (T. Schoener and D. Spiller).
Connor, E. F., and E. D. McCoy. . The statistics and biology of the
species-area relationship. American Naturalist : –.
Hamilton, T. H., R. H. Barth, and I. Rubinoff. . The environmental control of insular variation in bird species abundance. Proceedings
of the National Academy of Sciences of the United States of America :
–.
Lomolino, M. V., and M. D. Weisen. . Toward a more general species-area relationship. Journal of Biogeography : –.
MacArthur, R. H., and E. O. Wilson. . The theory of island biogeography. Princeton, NJ: Princeton University Press.
Quinn, J. P., and S. P. Harrison. . Effects of habitat fragmentation and
isolation on species richness: evidence from biogeographic patterns.
Oecologia : –.
Whittaker, R. J. . Disturbed island ecology. Trends in Ecology and
Evolution : –.
THE DESIGN OF NATURE RESERVES
In addition to “true islands” (bodies of land surrounded
by water), the species–area relationship has been well documented for many types of “island analogues” (patches of
suitable habitat for species isolated by unsuitable habitat)
on mainlands. Knowledge of the factors that shape the
species–area relationship can be used to evaluate different strategies for designing nature reserves that maintain
the highest number of species. Of course, the ideal strategy would be to have many huge preserves, but this may
not be feasible. Hence, the problem that conservation
biologists have pondered is whether one large reserve or
several small reserves with the same total area is better.
The answer to this question, sometimes referred to as the
SLOSS (single large or several small) debate, is not obvious: A single large reserve will have a lower per-species
extinction rate than will any smaller reserve, but the more
reserves there are, the less chance there will be for a species to disappear simultaneously from all of them. In a
review of land plants, insects, and vertebrates on islands,
Quinn and Harrison showed that total numbers of species on groups of several small islands were higher than
on a comparable area consisting of only one or a few large
islands. Among other explanations, they suggest that the
several small islands were distributed over a larger region
and contained more types of habitats, enabling more species to exist on the entire set of islands. Similarly, genetic
diversity within a species, which reduces extinction likeli-
SPIDERS
MIQUEL A. ARNEDO
University of Barcelona, Spain
The ability to produce silk is a distinctive feature of spiders. Silk-mediated airborne dispersal has allowed spiders
to colonize even the most remote archipelagoes. Spiders
on islands have served as models for the study of the evolutionary and ecological underpinnings of biodiversity.
Because of their generalist predatory habits, introduced
spiders may pose a serious threat to islands’ native fauna.
DEFINING A SPIDER
Spiders (order Araneae) comprise a megadiverse group of
arthropods that includes close to , species distributed
in  families. The origin of spiders can be traced back to
the Devonian, about  million years ago, representing
some of the earliest evidence of terrestrial life on Earth. As
the dominant non-vertebrate predators in most terrestrial
ecosystems, spiders have enormous ecological importance.
ACCESSING ISLANDS
Dispersal capabilities and generalist predatory habits
make spiders formidable pioneers. Spiders have a unique
SPIDERS
Gillespie08_S.indd 861
861
4/23/09 12:58:48 PM