Whittaker 2005

Diversity and Distributions, (Diversity Distrib.
) (2005) 11, 3–23
Conservation Biogeography: assessment

Blackwell Publishing, Ltd.
SPECIAL
PAPER
and prospect
Robert J. Whittaker*, Miguel B. Araújo, Paul Jepson, Richard J. Ladle,
James E. M. Watson and Katherine J. Willis
Biodiversity Research Group, School of ABSTRACT

Geography and the Environment, University of
There is general agreement among scientists that biodiversity is under assault on a
Oxford, Mansfield Road, Oxford, OX1 3TB, UK
global basis and that species are being lost at a greatly enhanced rate. This article
examines the role played by biogeographical science in the emergence of conservation
guidance and makes the case for the recognition of Conservation Biogeography as a
key subfield of conservation biology delimited as: the application of biogeographical
principles, theories, and analyses, being those concerned with the distributional dynamics
of taxa individually and collectively, to problems concerning the conservation of bio-
diversity. Conservation biogeography thus encompasses both a substantial body of
theory and analysis, and some of the most prominent planning frameworks used in
conservation. Considerable advances in conservation guidelines have been made over
the last few decades by applying biogeographical methods and principles. Herein we
provide a critical review focussed on the sensitivity to assumptions inherent in the
applications we examine. In particular, we focus on four inter-related factors: (i) scale
dependency (both spatial and temporal); (ii) inadequacies in taxonomic and distri-
butional data (the so-called Linnean and Wallacean shortfalls); (iii) effects of model
structure and parameterisation; and (iv) inadequacies of theory. These generic problems
are illustrated by reference to studies ranging from the application of historical bio-
geography, through island biogeography, and complementarity analyses to bioclimatic
envelope modelling. There is a great deal of uncertainty inherent in predictive analyses in
conservation biogeography and this area in particular presents considerable challenges.
Protected area planning frameworks and their resulting map outputs are amongst
the most powerful and influential applications within conservation biogeography,
and at the global scale are characterised by the production, by a small number of
prominent NGOs, of bespoke schemes, which serve both to mobilise funds and
channel efforts in a highly targeted fashion. We provide a simple typology of pro-
tected area planning frameworks, with particular reference to the global scale, and
provide a brief critique of some of their strengths and weaknesses. Finally, we discuss
the importance, especially at regional scales, of developing more responsive analyses
and models that integrate pattern (the compositionalist approach) and processes
(the functionalist approach) such as range collapse and climate change, again noting
the sensitivity of outcomes to starting assumptions. We make the case for the greater
engagement of the biogeographical community in a programme of evaluation and
refinement of all such schemes to test their robustness and their sensitivity to alter-
*Correspondence: Robert J. Whittaker,
Biodiversity Research Group, School of
native conservation priorities and goals.
Geography and the Environment, University of
Keywords
Oxford, Mansfield Road, Oxford, OX1 3TB, UK.
E-mail: [email protected] Conservation biogeography, models, protected area frameworks, scale, sensitivity
Order of co-authors arranged alphabetically. analysis, uncertainty.
guidance for biodiversity conservation (e.g. Dasmann, 1972,

INTRODUCTION
1973; Diamond, 1975), in our view it has done so as something
Although biogeographical science has played its part alongside of a poor relation — a Cinderella within Conservation Biology. In
other subfields of biology in the emergence of current scientific the present article, we argue that biogeography can now cast
© 2005 Blackwell Publishing Ltd www.blackwellpublishing.com/ddi 3

R. J. Whittaker et al.
aside its metaphorical rags as it emerges as an area of central the conservation message to policy makers (compare: Gillson &
importance to conservation planning. In part this has been Willis, 2004; Hambler, 2004; Willis et al., 2004a,b). As applies in
driven by the advent of extensive computerised biogeographical climate change research, we must also have, as a discipline, the
data bases and powerful computer-based analytical tools, which confidence to pursue critical analyses in order to improve the
have enabled rapid progress in many areas of the field, both biogeographical advice being offered. A key objective of our paper
pure and applied; and in part, it reflects theoretical and concep- is therefore to assess the sensitivity to assumptions inherent in
tual advances (e.g. Williams et al., 2000a; Lomolino & Heaney, the applications of Conservation Biogeography we review. In
2004). particular, we focus on (i) scale dependency; (ii) the Linnean and
Yet, we must also recognise that the underlying species distri- Wallacean shortfalls; (iii) model structure and parameterisation;
butional and other data often remain highly problematic, protocols and (iv) adequacy of theory. This is followed by a brief assessment
for analysis are still in development, and we have only recently of key research issues in the use of biogeography in developing
begun the task of systematically analysing the sensitivity of our and refining protected area planning frameworks.
analyses to the starting assumptions. Accordingly, we argue that
there is a need for more biogeographers to get engaged with the
WHAT IS CONSERVATION BIOGEOGRAPHY?
problems of conservation science, and for the injection of more
biogeography into training for conservation scientists (cf. Lourie Herein we define Conservation Biogeography as ‘the application
& Vincent, 2004). As a step towards these goals, we first offer a of biogeographical principles, theories, and analyses, being those
definition of the field of Conservation Biogeography, placing it in concerned with the distributional dynamics of taxa individually
the context of the closest surrounding disciplinary areas, and sec- and collectively, to problems concerning the conservation of biodi-
ond, highlight what we believe to be key generic research needs versity’. It is not a new area of scientific enquiry, but the conjunc-
within the field. tion of the words conservation and biogeography appears to have
We are within a crucial phase in the development of conserva- arisen very recently, in a meeting leading up to the first confer-
tion theory and strategy. There is general agreement that biodi- ence of the International Biogeography Society (held in 2003)
versity is under assault on a global basis and that species are being and the nearest to a book of the title that we are aware of is Applied
lost at a greatly enhanced level (Lawton & May, 1995; Royal Soci- Biogeography (Spellerberg & Sawyer, 1999). As an applied and
ety, 2003). In response, over the last decade, several prominent interdisciplinary science concerned with the conservation of
international non-governmental organisations (NGOs) have nature, Conservation Biogeography can be seen as a sub-discipline
engaged in developing regional, continental, and particularly of both Biogeography and Conservation Biology. We briefly con-
global schemes (e.g. Dinerstein et al., 1995; Long et al., 1996; sider the origins of these related endeavours in order of their
Olson & Dinerstein, 1998; Myers et al., 2000) to capture and emergence.
prioritize substantial new flows of conservation investment Biogeography is the study at all possible scales of analysis of the
(Dalton, 2000; Myers & Mittermeier, 2003). These schemes are distribution of life across space, and how, through time, it has
outputs from analyses that are essentially biogeographical in changed, within which a major concern has always been the dis-
nature, and they are having an immense influence in the organ- tribution and dynamics of diversity, frequently codified in terms
isation and prioritisation of conservation efforts (Jepson & simply of numbers of species, or proportions of endemic spe-
Whittaker, 2002a, Myers & Mittermeier, 2003). As such they are cies. The discipline has deep scientific roots, with some of the
perhaps the most prominent of recent conservation biogeographical major themes already established as areas of enquiry by the early
developments: but many other approaches also warrant attention. 1800s (Brown & Lomolino, 1998; Lomolino et al., 2004).
Thus, in the following, we illustrate our case by brief reference By contrast, nature conservation can be thought of as a social
to a wide range of biogeographical applications rather than movement working to develop or reassert certain values in society
providing an in-depth review of each. Even so, there are areas of concerning the human–nature relationship ( Jepson & Canney,
conservation biogeography, which, for reasons of space, get little 2001; Jepson & Whittaker, 2002b). The modern conservation
or no mention. movement emerged in the late 19th century in response to fun-
In climate change research, a mature debate has developed damental changes in world views concerning the human–nature
based on assessing how sensitive future climate scenarios are to relationship, emanating largely from elite society of the American
starting assumptions and model parameters (e.g. Gerber et al., East Coast and Western Europe (Jepson & Whittaker, 2002b).
2004). In this, conservation biogeography lags some way behind. The movement was motivated both by a desire to preserve sites
Indeed there appears to be a certain reluctance to entertain criti- with special meaning for intellectual and aesthetic contempla-
cal debate on the merits of particular schemes lest it detract tion of nature and by acceptance that the human conquest of
policy makers from their implementation or lead to loss of public nature carries with it a moral responsibility to ensure the survival
support. In illustration, critics of Conservation International’s of threatened life forms. The movement gained new momentum
hotspots scheme (see, e.g. Humphries, 2001; Brummitt & Lughadha, in the second half of the 20th century when science further
2003; Ovadia, 2003) have recently been labelled ‘dissidents’ by the expanded understandings of the society–nature relationship
scheme’s originator (Myers & Mittermeier, 2003). Similarly, (Frank et al., 1999; Adams, 2004).
acknowledgement of the non-pristine nature of many rain forest Motivated by, but distinct from the nature conservation move-
areas stirs controversy in part because of the dilution in clarity of ment, conservation biology is the name given for applied research
4 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Conservation Biogeography
designed to inform management decisions concerning the con- From this brief outline we highlight three points. First, the
servation of biodiversity. As such, its roots lie largely within the foundation of biogeography as a discipline substantially predates
mid-20th century: it gained huge momentum during the 1970s the emergence of conservation biology. Second, conservation
and early 1980s when it was formally identified as a sub- biogeography forms an important, distinctive (but not entirely
discipline, with dedicated journals and textbooks (e.g. Soulé, 1986; distinct) subset of conservation biology. Third, the motivating
Primack, 2002). It is a large field, but if subdivided by scale of force for these scientific endeavours is a diverse social movement,
application, we might recognise the following subdivisions of representing varied values.
relevant theory. First, the development and evaluation of biolog-
ical theory spanning population biological and genetic process,
SOCIAL VALUES AND CONSERVATION
and concerned with minimum viable populations, genetic ero-
BIOGEOGRAPHY
sion from small populations, competitive influence of invasive
species, behavioural ecology, etc., i.e. concerned with processes We are concerned in this article with the scientific underpinnings
in which biogeography is generally not prominent (e.g. see of conservation decision making. Yet it is important to recognise
Primack, 2002). Second, theory concerning processes at the that such scientific guidance, and the language in which it is
local-landscape scale, including the foundational influence of couched (Trudgill, 2001), is not value-free and that there is a
MacArthur & Wilson’s dynamic (equilibrium) theory of island debate to be had concerning what properties of nature we wish as
biogeography, the derivative Single Large or Several Small reser- a society to foster. Much of the scientific guidance and of current
ves (SLOSS) debate, habitat corridors, metapopulation theory, conservation practice assumes that this debate has a particular
and nestedness (reviewed by Whittaker, 1998): i.e. bridging outcome, without paying much attention to the possible validity
ecology and biogeography. Third, applications on a yet coarser of alternative value systems (but see: Redford et al., 2003). So, for
scale in part are concerned with mapping and modelling biogeo- example, Conservation International’s hotspots (Myers et al.,
graphic patterns and in part invoke historical biogeographic 2000) delimit as priority regions areas that possess > 1.5% of
theory concerned with the distribution and explanation of geo- global plant diversity uniquely within their bounds, and which
graphical patterns in diversity, which can be referenced back to have lost > 70% of their original habitat. In doing so, CI are nec-
the founding fathers of the discipline of biogeography (Alfred essarily placing higher value on those species that on a regional scale
Russel Wallace, Philip Sclater, etc.). We see such coarser scale work co-occur with many other range-restricted plant species, than
on the geography of nature as being unambiguously within they do on species occurring in other systems. Other value systems
the heartland of biogeography (cf. Lomolino et al., 2004) and exist, such as those emphasising the importance of intact
despite its undoubted importance within conservation science, megafaunal assemblages, aesthetic and cultural significance of land-
we argue that it is here that the ‘Cinderella’ tag applies to conser- scapes, ecosystem health or biotic integrity (cf. Callicott et al., 1999;
vation biogeography, and here where there is greatest need for Redford et al., 2003; Adams, 2004), and these value systems motivate
critical attention to our science, and for greater interaction conservation action in many nations, especially at local scales but
between those involved in theory and application (cf. Lourie & also globally (Jepson & Canney, 2001, 2003; Trudgill, 2001).
Vincent, 2004). The decision to adopt a particular set of values does not lie
Conservation biogeography, the application of biogeography within the bounds of science, and although conservation scien-
in conservation, is thus separable from the application of ecology tists are well placed to contribute to debate, there is an important
and macroecology most clearly at coarser scales. Whilst the distinction between the processes leading to the adoption of a set
use of zoogeographic regions, areas of endemism, geographic of values, and the process of deriving the scientific guidelines
patterns in species richness, or phylogeographic structure to implement these values. So, whilst there may be pragmatic
for conservation prioritisation purposes are readily identifiable reasons to opt for a ‘silver-bullet’ approach (e.g. Myers et al., 2000),
as conservation biogeography, applications at increasingly conservation biogeographers should be in the business of pro-
fine spatial scales, for example focused on habitat corridors viding alternative scenarios addressing differing end goals (cf.
or metapopulation dynamics can be seen as simultaneously Williams et al., 2000a; Dimitrakopoulos et al., 2004). Thus,
drawing from traditions in both ecology and biogeography. alongside the programme of exploring sensitivity to the more
Similarly, macroecological analyses [referring to the analysis prosaic assumptions concerning data quality, modelling effects,
of the emergent statistical properties of ecological datasets etc., set out below, should be added sensitivity of outcomes to
(Brown, 1995)] may also be based on both ‘ecological’ traits (e.g. different societal objectives.
growth rates, propagule size, breeding system, body size) and
‘biogeographical’ traits (e.g. range size, region of origin). In illus-
KEY ASSUMPTIONS AND SENSITIVITIES
tration, efforts to develop explanatory and predictive models of
invasiveness of non-native species have been made that mine
(i) The Linnean and Wallacean shortfalls
both sets of traits, frequently finding a biogeographical signal in
the resulting models (Dehnen-Schmutz, 2004; PyÍek et al., 2004), Our knowledge of the overall biodiversity of the planet remains
indicative that such analyses draw from both ecological and bio- woefully inadequate, with estimates of global biodiversity rang-
geographical traditions within conservation science to varying ing over one or possibly two orders of magnitude (Groombridge,
degrees. 1992; Brown & Lomolino, 1998; Groombridge & Jenkins, 2000).
Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 5

Much of the diversity we do know about has yet to be formally areas) for the Brazilian Amazon. The validity of these patterns is
described and catalogued. In general, these problems — the so-called critical to debates on the Quaternary history of the Amazon (e.g.
Linnean shortfall (Brown & Lomolino 1998) — appear to be of Bush, 1994; Colinvaux et al., 2001). Nelson et al. (1990) quanti-
increasing severity as the organisms decrease in size and com- fied collecting intensity within the herbarium collection at
plexity, e.g. from vertebrate groups, to invertebrates, down to Manaus by recording the number of specimens of the virtually
nematodes (e.g. see: Medellín & Soberón, 1999; Ødegaard et al., ubiquitous tree genus Inga per 1° grid cell. They found that
2000; Lambshead & Boucher, 2003). In so far as we do know our vouchers of Inga varied from 0 to > 320 per cell, and that only
species, we also have, for many taxa, inadequate knowledge of one of the seven previously proposed Brazilian CORE areas was
their global, regional, and even local distributions, a problem not a focus of collecting activity: in short, the biogeographic pat-
that Lomolino (2004) has labelled the Wallacean shortfall. Many tern is confounded by collecting intensity. Despite recent interest
areas of the world remain seriously under-collected for most in artefactual effects in diversity gradient analyses (e.g. Diniz-
taxa, with the result that even for higher plants, reliable, system- Filho et al., 2003; Colwell et al., 2004), formal analyses of possible
atic species range maps — the necessary basis for robust analyses artefacts in collecting effort and data quality underpinning
of diversity patterns — are available only for a fraction of the coarse-scale strategic conservation schemes appear to be rare
earth’s surface. Moreover, much of our distributional data and (but see, e.g. Prinzing et al., 2003). It would seem worthwhile to
many of our compilations pertain to political geographical units explore the sensitivity of such data and schemes to the problems
(states) lacking biological meaning, which have gone their own inherent in the Wallacean shortfall.
ways historically in gathering botanical and zoological data, and
which vary hugely in area: factors liable to produce serious
(ii) Scale dependency
artefacts when data are combined for analysis.
The following examples illustrate some of these problems,
Spatial scale: extent and focus
beginning with an esoteric choice: marine nematodes. Nematodes
have been identified as a possibly hyper-diverse group, especially There are numerous metrics of diversity, and they have been
benthic nematodes in the deep sea (i.e. estimates of > 1 m species applied on widely varying scales, often inconsistently (Whittaker
have been made). Yet, to date, very few marine nematodes have been et al., 2001). A crucial distinction is that between (a) the geo-
described to species level, and these high values are derived using graphical extent of a study system, being the space over which
extrapolation techniques from local sampling. A recently col- observations are made, e.g. a hillside, a state, a continent; and (b)
lected transect spanning 3000 km of the abyssal plain of the north- the grain (focus) of the data, being the contiguous area over which
central Pacific ocean indicated a regional richness of only a few a single observation is made, or at which data are aggregated for
hundred species (Lambshead & Boucher, 2003). Intriguingly, analysis, e.g. a light trap, 1 ha plot, or latitude–longitude grid cell.
fully 71% of the species in the two northern stations of the Often the term scale, or spatial scale, is used to refer to both of
transect were also collected from the (more species rich) south- these quite distinct properties of study systems, but they are not
ern stations. Not only is there a pronounced Linnean shortfall in equivalent or substitutable (Palmer & White, 1994; Nekola &
marine nematodes, but the relationships between alpha, beta and White, 1999; Whittaker et al., 2001; Whittaker & Heegaard,
gamma richness patterns now appear to be radically different 2003). This is shown by analyses using nested sets of grid cells
between the shallow and deep seas, i.e. there is an equally large that reveal different patterns of diversity by varying grain and
Wallacean shortfall, and initial assumptions now seen unfounded. holding extent constant (Stoms, 1994; Rahbek & Graves, 2000,
On the basis of these and other recent data, Lambshead & 2001; Koleff & Gaston, 2002). Similarly, when scale (i.e. grain) is
Boucher (2003) suggest that a radical reassessment downwards held constant, but geographic extent is varied, again different
of the global diversity of marine nematodes is necessary. patterns can result (e.g. Hawkins et al., 2003). The failure to
While the social and economic value placed on marine nema- control for scale (e.g. by holding area constant in analysis) and
todes and their conservation may currently be limited, the same the confounding of grain and extent within the diversity litera-
cannot be said of wild plants. The Atlas Florae Europaeae (AFE) ture has undoubtedly hindered progress in the development of
is a project, launched in 1965 as a collaborative effort of theory. These scale issues have a clear relevance to the develop-
European botanists, with the aim of mapping and providing taxo- ment of strategic conservation frameworks, in so far as they
nomic notes for the entire European flora. Some forty years on, make use of, or are dependent upon, mapping diversity patterns
twelve volumes have been published. Yet, as of 2003, the AFE was (Williams et al., 2000a; Redford et al., 2003). This is shown by
able to claim only that the maps cover families that include more Lennon et al.’s (2001) analyses of British birds, in which spatial
than 20% (our emphasis) of the vascular plants of the European patterns of richness using a 10-km grain grid system were found
flora. Moreover, electronic versions of early maps have only to be statistically unrelated to those using a 90-km grain system:
recently become available. We are still a long way off being able to with coarsening scale the locations of the most species-rich
take full advantage of the collections residing in herbaria and quadrats moved north. This indicates that reserve selection
museums, notwithstanding massive improvements in computer criteria based on species distribution and diversity data may be
power, etc. (Wheeler et al., 2004). sensitive to the scale of analysis employed. Relatively little work
Nelson et al. (1990) provide a tropical botanical example in has been done on this problem, although it may have a strong
their critique of ‘centres of richness and endemism’ (CORE influence (Stoms, 1994; Araújo, 2004).

population size is good news. Of course, this may be a misleading

Range restricted species — and thresholds of rarity
example, and it must be balanced against the knowledge that, for
Rarity, as Darwin commented, is often a precursor to extinction. example, about half of Hawaii’s indigenous bird species have
Rarity can mean that a species occurs at low density, and it can gone extinct since human colonization, and that population data
mean that a species occupies a small geographic range. Given indicate many island species to be seriously threatened (Whit-
these two forms of rarity (and there are others: Gaston, 1994) a taker, 1998).
species can be: (i) abundant and widespread; (ii) abundant but The threshold of 50,000 km2 has been widely adopted in con-
localised; (iii) low density but widespread; or (iv) low density and servation prioritisation analyses (e.g. Long et al., 1996; Rodrigues
localised. It has long been known that a high proportion of et al., 2004), and for particular taxa and contexts is a reasonable
species have relatively small geographic ranges, and few are approximation (e.g. for freshwater fish species in North America,
widespread. Brown’s (1995) graphical macroecological analyses of Rosenfield, 2002). In the IUCN red list criteria, attribution can
range size and density indicate that within taxonomically or eco- default to range size if other data are lacking, in which case, the
logically similar species there tends to be a positive correlation thresholds used are 100 km2 (extent of occurrence) for critically
between these two properties. This means that species occupying endangered, 5000 km2 for endangered, and 20,000 km2 for
large ranges tend to be more abundant throughout those ranges vulnerable (version 2.3; www.redlist.org/info/categories_
than range-restricted species, suggesting the corners of the plot criteria1994.html). This three-point scale recognises that on
represented by conditions (ii) and (iii) to be very unusual. When average, threat increases incrementally as range size reduces. For
examining whole faunas, such as the Breeding Bird Survey data those schemes employing a single range size, whether 50,000 km2
for North American land birds, a lot more scatter is evident than or another, it would appear valuable to undertake analyses for
within similar groups of species, but still the condition of different taxa and biogeographical contexts of the sensitivity of
‘abundant and localised’ (i.e. ii) is extremely rare (Brown, 1995, prioritisation analyses to the adoption of that threshold (Williams
Fig. 6.2; and see Gaston, 1994). It would therefore seem that con- et al., 2000a). Moreover, to refine the meaning of ‘range-restricted’
servation biogeographers are right to focus attention on range- for conservation prioritisation, we need more empirical data to
restricted species as being a priori of conservation concern translate range size distributions into population estimates (c.f.
(Rosenfield, 2002). www.birdlife.org /datazone) across a large sample of species for
Yet, what constitutes a small range, and how might the range different contexts, e.g. mainland versus island systems.
size/abundance pattern vary in relation to biogeographic context
(e.g. continental versus island archipelago contexts) or for taxa
Temporal scale
of differing body sizes? In the North American breeding bird
survey data set analysed by Brown (1995), range size varies from In conservation biogeography we are — or should be — concerned
c. 10,000 km2 to over 10 million km2 and fewer than 10 species with both temporal and spatial scale dependency (Gillson & Willis,
have ranges < 100,000 km2. By way of contrast, the land area of 2004). By analogy to the terms used in considering spatial scale,
the Canary Islands is only approximately 7500 km2 and that of the the study ‘grain’ in temporal analyses would refer to the interval
Hawaiian archipelago is 16,640 km2. Assuming that the typical between data points, and the ‘extent’ to the temporal depth of the
pattern of many species with small ranges exists within these analysis. Although fine resolution data often reveal oscillations
essentially self-contained biogeographic areas, then many native hidden in coarse resolution assessments, typically, it is the sensi-
species have tiny ranges, and even the most abundant species on tivity of forward projections to the length of the data series used
these islands will have smaller ranges than the ‘localised, low that is of greatest concern in the present context. Conservation-
density’ rarities of the North American data set. In conservation, a ists aim to ensure the survival of species in the long term. Hence
common threshold for defining range-restricted (sometimes it is important to understand the temporal dynamics of bio-
‘local endemic’) species is that the range size is < 50,000 km2 geographical processes and whether short-term trends can be
(from Terborgh & Winter, 1983). On this basis, the entire endemic projected into the future (Gillson & Willis, 2004). Yet, the rapidity of
biota of island archipelagos such as Hawaii, Galapagos and the contemporary changes such as habitat fragmentation is such that
Canaries would be classed as of conservation concern. This attempts typically have to be made to project far into the future
would be unduly pessimistic. A simple illustration is provided by on the basis either of snapshot data or of very short temporal
the case of the Tanimbar corella (Cacatua goffini) — a parrot series. There are many attendant difficulties in doing so, as for
species endemic to the Tanimbar islands (Indonesia), a group of instance, the ecosystem dynamics of forest fragments may respond
66 islands of a total land area of about 5400 km2. It was initially to forcing via subtle alterations to the population dynamics of
listed as ‘threatened’ in 1989 on the basis of small global range the dominant organisms (e.g. forest trees) on decadal or multi-
and concern that it was being traded at a possibly unsustainable ple-decadal timescales (Bush & Whittaker, 1991; Whittaker,
rate, but field survey on the largest island, Yamdena (3250 km2) 1998; Willis et al., 2004a).
subsequently produced a population estimate of 231,500 (± 33,000) A classic illustration both of genuine reasons for concern and
for that island alone, suggesting that the initial categorisation had of the uncertainty involved in projections is provided by the
been unwarranted (Jepson et al., 2001). That an island species tropical moist forests of the Atlantic seaboard of Brazil, which
with a global range one-tenth the size of the Terborgh & Winter have been reduced over the past few centuries to only an estimated
(1983) threshold for range-restricted species can have a healthy 12% of their former cover. Simple species–area calculations, based

on island biogeography theory, predict about 50% extinctions. key to understanding such range shifts but is a difficult process to
To date, none have been documented with certainty, although model. Recent work by Kullman (e.g. 1998) indicates that
species-level assessments show many species to be IUCN listed as traditional pollen analytical techniques underestimate the ability
vulnerable, endangered or critically endangered based on range, of some temperate species to respond rapidly to climate warming
or population estimates (see www.birdlife.org/datazone). Brooks events. Such insights, and others derived from phylogeography,
and Balmford (1996) compare losses of birds projected using the demonstrating genetic legacies from the restriction and sub-
species–area approach with those listed as ‘threatened’ and find sequent emergence of populations from refugial areas, can con-
congruence, i.e. they argue that the predictions are basically cor- tribute considerably to an improved theoretical understanding of
rect, but that there is a substantial lag between the habitat loss/ responses to climate forcing, and to the parameterisation of models
fragmentation process and global extinction of the species. It forecasting future changes (Fig. 1; Willis & Whittaker, 2000;
should of course be recognised that species–area analyses based Bush, 2002; Schmitt & Krauss, 2004).
on island theory provides only a coarse, stochastic simulation of
impacts of fragmentation, and the actual drivers of extinction
(iii) Effects of model structure and parameterisation
may be structured, and in some instances avoidable by appropri-
ate management. Conservation action in the Atlantic forests has In this section, we focus on the sensitivity of conservation-
likely already been and will continue to be crucial to staving off the biogeographical analyses to how we construct and parameterise
predictions of the species–area models (Birdlife EBA factsheet our models. The taxonomic resolution at which we set priorities
75, www.birdlife.org/datazone/ visited July 2004). This also is one example. Endemism is usually considered in schemes such
means that in many cases it is impossible to distinguish between as Endemic Bird Areas (EBAs) (Long et al., 1996) at a species level,
inadequacies of theory (reviewed in Whittaker, 1998) and suc- but could be analysed at family, genus, species or subspecies levels,
cessful implementation of conservation strategies. each implying a diminishing degree of taxonomic distinctiveness
Considering Quaternary-scale studies, it is difficult to obtain or evolutionary distance. All of the regional and/or global prior-
high precision data on past ecosystem dynamics and how they itisation schemes (as Table 1) that we are aware of operate at the
relate to environmental driving forces, particularly at a mecha- species level. How might they change if conducted at the family
nistic level. Techniques such as pollen analysis offer the best level, or the subspecies level? Modelling of how spatial patterns of
resolution, and have greatly improved our understanding of plant richness varies at species, genus and family levels suggests
responses to climatic forcing. In essence, systems respond highly that whilst species and genus level patterns are closely related,
individualistically, such that past community signatures exist for shifting to the family level may produce significant differences
which there are no modern-day analogues (e.g. Bush, 1994, 2002). (e.g. O’Brien et al., 1998, 2000; Dimitrakopoulos et al., 2004).
Yet, palaeoecologists still debate the respective roles of changing An illustration of the effects of adopting different modelling
moisture regimes, temperature regimes and CO2 concentrations approaches is provided by da Silva et al.’s (2004) analyses of
in driving the changes (Bennett & Willis, 2000; Colinvaux et al., endemism in passerine birds of the Atlantic forests of eastern
2001). There is also a considerable degree of uncertainty concern- South America: a well-recognised biogeographical unit. The steps
ing the migrational lags evident in range adjustments to climate in their analysis were as follows: (a) define operational geo-
forcing. Long distance dispersal at low density is likely to be the graphical units (OGU); (b) construct a data matrix; (c) perform
Figure 1 Some of the steps, choices and

assumptions involved in modelling species
losses resulting from future climate change
using the bioclimatic envelope modelling
approach (e.g. Pearson & Dawson, 2003;
Thomas et al., 2004, Thuiller et al., 2004). Not
all studies involve all elements, e.g. land-use
data, or dispersal models, but these
components are important for increased
realism (Pearson & Dawson, 2004). Original.

Table 1 A typology of protected area planning frameworks of global remit. Biogeographical representation approaches set out to identify
representative examples of ecosystem types identified using established biogeographical or ecological frameworks (e.g. Dasmann, 1972, 1973;
Udvardy, 1975; Dinerstein et al., 1995). In contrast to this largely zonal approach, hotspot approaches are azonal, and target places rich in species
and under threat (e.g. Long et al., 1996; Myers et al., 2000). The Global 200 Ecoregions project combines elements of both hotspot and
representation approaches (Olson & Dinerstein, 1998). The approach designated ‘important areas’ uses criteria such as numbers of wintering
wildfowl congregating in the site, or importance of site as a stage in a migration route, to identify important areas for prioritisation (e.g. http://
www.birdlife.net/action/science/sites/ visited September 2004). Source: P. Jepson & R.J. Whittaker (unpublished)
Global protected area planning frameworks
Representation Hotspots Important Areas
Basic idea An example of each Maximize number of species Select a suite of sites that together
‘saved’ given available resources protect key attributes of concern
Questions arising What are the units of nature? Where are places rich in What are the key attributes & how can
What do we add next? species and under threat? they be assessed?
Parameters Vegetation formations Species richness Threatened species

Faunal regions Species endemism Endemic species
Ecoregions Rate of habitat loss Species assemblages
% original habitat Congregating species
Methods % dissimilarity Species counts Application of varied

Controlling factors Threat /diversity criteria
indexes
Schemes Biogeographic Biodiversity Hotspots Important Bird Areas

provinces Endemic Bird Areas
WWF Ecoregions Key Biodiversity Areas
Global 200 Ecoregions
a parsimony analysis of the data matrix to produce an area clado- biogeographers to adopt flexible and inclusive approaches to
gram; (d) delimit the OGU or groups of OGUs defined by at least their work. Whilst the goal for an international NGO may appro-
two endemic species; and (e) map the species endemic to each priately be the production of a single map for conservation
OGU or groups of OGUs to delineate the boundaries of each area prioritisation (Redford et al., 2003), the broader conservation
of endemism. OGUs can be outputs from previous schemes, e.g. biogeography community should be concerned with providing
ecoregions, biomes, provinces, or as in this case, grid cells (24 sets of alternative maps, based on varying parameters and choices.
cells spanning 1°). Their analysis identified four areas of ende- A key distinction within spatial diversity analyses is that between
mism (Pernambuco, central Bahia, coastal Bahia and, Serra do richness and endemism, i.e. between the number of species per
Mar) that exhibit some congruence with those identified for unit area and the number of local /regional endemic species per
woody plants, bamboos, and swallowtail butterflies, suggesting unit area. Areas rich in (restricted-range) endemics are often also
that they may have some generality. However, as they also point species rich, but there is high variability at regional scales (Williams
out, they contrast with previous analyses of vertebrate data, in et al., 2000a), and patterns in richness and endemism are not
identifying varying numbers of areas of endemism, with incom- necessarily positively related (see maps in Huntley, 1996). Whittaker
plete overlap with their own. Whilst the differences arise in part et al. (2001, 2003) argue that explaining patterns of endemicity
from improvements in data, they in large part reflect model requires theories focussing on evolution and historical contingen-
effects. In particular, their use of Parsimony Analysis of Ende- cies, whilst species richness patterns can often be related to con-
mism is controversial in that the methods are subject to heated temporary ecological processes and controls. This decoupling is of
dispute amongst historical biogeographers (e.g. see Brooks & van more than merely theoretical interest, as prioritisation schemes based
Veller, 2003; Ebach et al., 2003). The degree of difference in the on maximising richness may produce different outcomes from ones
outputs between da Silva et al.’s analyses and Stattersfield et al.’s based on maximising numbers of endemics (Williams et al., 2000a).
revised (1998, and see Long et al., 1996) scheme is not huge but This is illustrated by Dimitrakopoulos et al. (2004) who, in
it is noteworthy in that the latter’s Endemic Bird Areas approach essence, provide a sensitivity analysis for prioritising plant diver-
has already provided a key strategic planning tool for the deploy- sity in Crete at species, genus and family levels using a grid of
ment regionally and globally of the conservation advocacy efforts cells each of 8.25 km2 (Fig. 2). They used hotspot analysis and
of Birdlife International. There would again seem to be scope complementarity algorithms to define priority areas for conser-
for further analyses of the sensitivity of this form of ‘hotspot’ vation to compare with the recently designated Natura 2000
mapping outcome to model assumptions, and this may require Special Areas of Conservation (SACs) in Europe. The Natura

Figure 2 Alternative maps of potential

conservation prioritisation of plants in Crete,
based on 162 grid cells of 8.25 km2. Locations
of: (a) species richness hotspots; (b) species
threatspots and endemic hotspots; (c) grid cells
chosen to represent all genera at least once; and
(d) grid cells chosen to represent all families at
least once. The asterisks indicate the Natura
2000 Special Areas of Conservation. (From
Dimitrakopoulos et al., 2004, Figs 1 and 2). See
text for further details.
2000 network was based on an area-selection strategy mostly spatial overlap between rarity hotspots, endemic hotspots and
concerned with representativeness, quality of habitat, size, density threatspots is only about 33%, whilst that between the comple-
of listed species, and the degree of their isolation. In their com- mentary sets calculated for all species and the various categories
parative analysis, Dimitrakopoulus et al. scored each cell for: of hotspot is less than 9%. In addition, the congruence with the
species richness, a vulnerability index (based on IUCN listings), Natura 2000 sites is also low — these sites cover a substantial area,
a taxonomic distinctiveness index (reflecting proportionality of located in 32 grid cells, but the overlap between them and the
species-genus-family membership), a rarity index (reflecting species various hotspot categories is at best only 14%. This analysis indi-
of restricted range within Crete), and endemicity (i.e. island cates that Natura 2000 sites in Crete do not provide a particularly
endemics). When these differing indices are used to identify good representation of the plant diversity of the island. The
hotspot areas (i.e. the top 5% of cells) and complementary sets analysis also shows that model effects, viz: (a) hotspot versus
(i.e. the minimum number of cells containing all target taxa), complementarity algorithms; (b) taxonomic resolution; and (c)
they produce quite varied outcomes (Fig. 2). For instance, the priority given to richness, taxonomic distinctiveness, and rarity,

Figure 3 Sensitivity to thresholds in

modelling representation in Protected Areas.
Coloured cells show species richness
(maximum values in red) of plant species not
represented in reserves using 10 × 10 km grid
cells (for mainland Portugal). A cell is
switched on as reserved dependent on how
much of its area is covered by a reserve, using
thresholds varying from 1% up to 99%. Grey
cells are those that have at least some reserved
area within them (figure modified from
Araújo, 2004, Fig. 2; reproduced with
permission from Elsevier).
provide for a highly diverse range of conservation solutions. Such questions concerning the paradigmatic structures within which
model effects are likely to be typical of reserve selection analyses conservation biogeography has so far been constructed. For
(e.g. Fig. 3). A further problem that has been identified in analy- instance, in respect of spatial scale, it has long been appreciated
ses of this sort is that there is often a multiplicity of different but that species richness of otherwise equivalent sample areas
equally efficient complementarity solutions, meaning that increases as larger areas are sampled. Thus comparative analyses
comparisons such as that carried out for Crete should be based on of richness — whether exploring causation in species diversity or
multiple runs to ensure that the differences resulting are robust (e.g. assessing conservation importance — which fail to hold area
Gaston et al., 2001; Hopkinson et al., 2001; Araújo et al., 2005). effectively constant, are fatally compromised (Whittaker et al.,
There is now a substantial body of work showing variability of 2001; Brummitt & Lughadha, 2003; Ovadia, 2003). One reason
area selection outcomes as a function of starting assumptions they continue to feature in the literature is that insufficient efforts
and goals (e.g. Pressey & Nicholls, 1989; Williams et al., 1996, have been made to develop diversity theory structured with
2000a, 2000b; Rodrigues et al., 1999; Andelman & Fagan, 2000; respect to scale (Whittaker et al., 2001, 2003; Willis & Whittaker,
Dimitrakopoulos et al., 2004). However, these studies typically 2002; Rosenzweig, 2003; Whittaker, 2004).
concern simple problems at a regional scale, ignoring issues of In respect of temporal perspectives, despite a number of calls
range dynamism, and being restricted to single taxa. Whilst some for change, a paradigmatic structure persists within conservation
studies have improved on these static approaches (e.g. Araújo & that the environment has two ‘states’: pre- and post-anthropogenic
Williams, 2000; Araújo et al., 2002a,b; Cabeza, 2003; Cabeza et al., impact. Prior to anthropogenic impact the ecosystems is viewed
2004), if we begin to consider more sophisticated problems, as an undisturbed natural system in equilibrium with the envi-
such as the conservation implications of climate change, with all ronment: it is ‘pristine’ and change is slow. In contrast, the post-
the associated uncertainties, then the various model effects mul- anthropogenic period is seen as a time of rapid change, where a
tiply (Fig. 1; Araújo et al., 2004a; Meir et al., 2004; Williams et al., threshold is exceeded and the environment irrevocably altered
2005). The use of so-called bioclimate envelope models to capture for the worse. Yet there are now numerous studies to indicate that
the relationship between species distributions and climate space and this structure is far too simplistic. For instance, some of the sys-
thereby to model future responses to climate change has advanced tems associated with the labels ‘pristine’ and ‘fragile’ turn out to
considerably in the last few years. Yet, the sensitivity to key assump- have recovered from past intensive human use (Willis et al.,
tions is such that the resulting bioclimate envelope modelling 2004a,b). The use of such value-laden terms is part and parcel of
projections are so uncertain as to effectively compromise their this two-state worldview (cf. Stott, 1998; Trudgill, 2001), and it is
practical value (compare Buckley & Roughgarden, 2004; Thomas a paradigm we surely need to break away from in recognition of
et al., 2004; Thuiller et al., 2004) notwithstanding the wider the capacity of biogeographical phenomena to be variously
impact they may have (Ladle et al., 2004; Hannah & Phillips, 2004). equilibrial and non-equilibrial dependent on the properties under
These models should be seen as providing merely interesting scrutiny, the geographical context, and the spatial and temporal
‘what if ’ scenarios, until such time as we find ways of assessing scales of the assessment (as Whittaker, 2004).
model variability and reliability quantitatively (Hannah et al., 2002; One prominent contribution of biogeography to conservation
Pearson & Dawson, 2003, 2004). biology has been the use of island theory, in which the dynamic,
equilibrium theory of MacArthur &Wilson (1967) has been
seminal. We take this body of theory in illustration of the impor-
(iv) Adequacy of theory
tance of improved theoretical foundations. In the early 1970s,
In addition to practical issues concerning e.g. finding the most Diamond (1975) and others set out a theoretical framework for
efficient tree building or complementarity algorithms, it is also fragmentation research derived from MacArthur–Wilson theory,
important to consider the adequacy of biogeographic theory. which has spawned a host of studies of species area relationships
Whilst this theme is only partially separable from the foregoing (SPARs) for terrestrial ‘habitat islands’. The premise is that
sections, it is important to focus explicitly on some of the deeper habitat fragmentation will produce predictable patterns of species

loss as previously equilibrial patches become ‘supersaturated’ profound importance for understanding the eventual impact of
and thereafter ‘relax’, i.e. shed species. Diamond drew on sup- fragmentation on species losses, just as was suggested by MacArthur
porting quantitative evidence for relaxation coming largely from & Wilson (1967). In other words, as systems move from the sample–
‘land-bridge islands’, the assumption being that prior to sea level area to archipelagic curve they should shed species, and should
rise at the start of the Holocene, these islands would have been do so at a predictable rate.
connected to the mainland, and would have shared close to the Unfortunately, island theory does not tell us how species
full quota of species now present on the mainland. The discrepancy should be distributed across a system of fragments, and this is
between this assumed starting value and that derived from crucial to the issue of whether it is better to advocate one very
species–area plots for ‘control’ areas is taken to represent the large reserve, or several smaller ones of the same total area. To
number of extinctions, with the speed of equilibration being seen some extent, Diamond himself acknowledged this when he wrote
as an inverse function of area. The possibilities that the observed ‘separate reserves in an inhomogeneous region may each favour the
discrepancies reflect the influence of variables other than area survival of a different group of species; and … even in a homogeneous
and isolation, or that biotas may not initially be in the assumed region, separate reserves may save more species of a set of vicariant
equilibrium richness state prior to fragmentation (Boecklen & similar species, one of which would ultimately exclude the others
Gotelli, 1984; Boecklen, 1986) were recognized but downplayed. from a single reserve’ (Diamond, 1975). If we take two extremes,
Diamond made a number of fairly bold statements about our systems could be either perfectly nested, or perfectly non-nested.
ability to predict not only total numbers of losses (the oft-cited If nested, it means that each smaller set of species is a perfect sub-
50% species remaining in 10% of the area figure) but also rates of set of the next larger: in which case the total number of species
losses. Then, as now, such predictions appear to be derived largely held by the system of fragments is the number held in the largest
from inferences based on studies of actual islands, and of the (richest) patch. If the fragment system is perfectly non-nested,
form of SPARs (e.g. Rosenzweig, 1995, 2003) rather than on then the number held in the system as a whole is the sum of the
studies of recorded losses from fragmented habitats within large number held in each patch, as each contains unique species (Fig. 4).
landmasses (but see Stouffer & Bierregaard, 1996). Empirical work to date indicates that most systems tend signifi-
It is disappointing that we still know so little about the power cantly towards being nested, but are not perfectly nested (Whit-
and timescale of ‘species relaxation’. There have paradoxically taker, 1998). But, how sensitive are these observations, of
been many studies of fragmentation, but remarkably few appear SPAR z-values and of compositional structure, to the scale of our
to have had ‘before’ and ‘after’ data or have undertaken repeat system, i.e. to the grain and extent of the data? As fragments are
surveys of the same fragments over a lengthy period (but see
Honnay et al., 1999; Gonzalez, 2000; Bierregaard et al., 2001). As
a result, inferences have to be made based on the use of the
species–area relationship, typically assuming: a fixed z value (slope)
of 0.25 (e.g. Brooks et al., 2002), that the system was essentially
equilibrial prior to fragmentation, and that the properties of the
matrix habitats can be ignored. Each of these assumptions is
problematic (Whittaker, 1998; Watson et al., 2005).
An important contribution to our understanding of SPARs
is provided by Rosenzweig (1995, 2003), who demonstrated the
importance of separating out different categories of curve (and
see Scheiner, 2003, 2004; Gray et al., 2004a,b). First, we may re-
cognize and discount species-accumulation or collectors curves
(Gray et al., 2004b), which describe the increase in total number
of species with increasing collection effort within an area. Rosen-
zweig recognizes three scales of ‘true’ species–area curves (whereby
number of species for each area is plotted against area), which he
terms the sample–area, archipelagic and interprovincial. Respec- Figure 4 A variant on the theme of the design principles for nature
tively, these are (i) sample–area: constructed from considering reserves, originally derived from consideration of island theory by
different sized (non-nested) samples from within non-isolated Diamond (1975) and others. Here we posit a choice between a single
habitat within the same region; (ii) archipelagic: from islands of large reserve and a network of smaller reserves of approximately the
an archipelago; and (iii) interprovincial: constructed from all same total area. In the perfectly nested system, the largest reserve
contains all the species in the reserve network. In the perfectly non-
species found in separate biological provinces, i.e. more or less
nested system, the set of smaller reserves contains more species than
self-contained regions whose species are predominantly native
the largest reserve and no species in common with it. Empirical data
in origin. Rosenzweig argues that the z-values (slopes of the tend to support nestedness as a general property of small island/
log-richness–log-area relationship) vary systematically between habitat island data sets sampled from within a landscape or small
these three forms, (i) sample–area, typically between 0.1 and 0.2; region. The degree of nestedness must diminish if islands/habitat
(ii) archipelagic, typically between 0.25 and 0.55; (iii) interpro- islands from across larger areas, or distinctive biogeographical areas
vincial, 0.6 –0.9. If these ranges are indeed typical, then they have are considered.

considered across a wider area, encompassing more diverse

microenvironments and more biogeographic history, so we cap-
ture more beta diversity, and nestedness must diminish in a hab-
itat island data set. Thus, the answer to the SLOSS debate can be
expected to change from an emphasis on large reserves to an
emphasis on many, comparatively smaller reserves, as the grain
and extent increase above the local-landscape scale.
Moreover, it has recently been argued that the approach to
model fitting for SPARs should be based on an updated consider-
ation of ecology rather than statistics, i.e. that we should examine
the fit of theoretical (mechanistic) models against empirical data sets
rather than setting out to find the best straight-line fit through Figure 5 Insular distribution of a focal species (dark symbols
the data (Tjørve, 2003). Lomolino (2000) has also questioned the indicate presence, open circles indicate absence), highlighting the
possible effect of matrix properties on the presence of a species. The
approach of using data transformation in search of the best linear
grey remnants and dashed line indicate that a species would inhabit
fit. He argues that there may be good ecological reasons to posit
these remnants when in a landscape with matrix composition ‘A’ but
more complex scale-dependent relationships, and that untrans-
would not in matrix composition ‘B’. From Watson (2004), after
formed SPARs should exhibit a sigmoidal form, with a phase Lomolino (2000).
across low values of area where species numbers scarcely increase,
followed by a rapid increase with area and a subsequent flattening
as the number of species approaches the richness of the species pool. Perhaps, the whole paradigm of viewing habitat islands and
The existence of a ‘small island’ effect appears well established reserves as islands has become part of the problem? Applied bio-
(Lomolino 2000, 2002; Lomolino & Weiser, 2001), and thus must geographers have been so focussed on the fragments that they
have some influence on z-values in standard linear SPAR regres- have paid insufficient attention to studying and developing
sions, but Williamson et al. (2001, 2002) contend that there is no theory encompassing the matrix. In his original 1975 paper
evidence of an upper asymptote. It is noteworthy that the form Diamond noted other ecological features that influence the like-
and biological meaning of SPARs remains the subject of so much lihood of persistence in reserves of different sizes, and captured some
debate and confusion. One important reason is that empirical of this signal by the use of ‘incidence functions’ — a static analysis of
data sets each have their own unique combination of variation in the distribution of a species across a system of isolates. Recent
contributory environmental variables (e.g. area, elevation, climate, work by Watson (2004; Watson et al., 2005) has demonstrated
habitat type, isolation, disturbance histories, etc.). Each of which that at least in some ecological contexts, matrix effects can be
has relevance, but not necessary significantly so over all scales pronounced. Analyses of incidence of 27 woodland bird species
within the empirical data gathered. Perhaps unsurprisingly, few in three landscapes from the Canberra area of the Australian
SPAR analyses have attempted to tackle such threshold responses Capital Territory, in relation both to area and isolation show a
in respect of the complex array of potentially important, remarkably varied response not just between species, but
interacting variables. It has thus proven difficult to identify between populations of the same species in different landscapes;
general patterns, and to develop general models through this where the principal apparent difference is not in properties of the
approach. woodland habitat islands, but in the characteristics of the matrix
Yet, numerous habitat island data sets have been collected, and habitats (defined as urban, peri-urban, and rural) that contain
so what is needed is not necessarily new field efforts, but more the habitat islands (Figs 5 & 6). Although the premises from
concerted efforts in analysis and synthesis, to tease out the scale which we arrive at these conclusions may differ, we are in agree-
sensitivity of habitat island data sets, particularly, the form of ment with Rosenzweig (2003) and others (e.g. Renjifo, 2001;
their SPARs, and their compositional structure, and how these Lindenmayer & Franklin, 2002; Wethered & Lawes, 2002) in
responses vary between different higher taxa (e.g. butterflies, all arguing that greater attention needs to be given to developing
invertebrates, higher plants, birds). Some theoreticians, notably guidelines for so-called ‘extensive’ conservation efforts encom-
Rosenzweig (2003), see dire implications in the continued process passing processes across whole landscapes — i.e., conservation
of ecosystem fragmentation and area reduction as systems move within the matrix in which the reserves are embedded.
from samples to archipelagic to interprovincial SPAR forms. Yet,
it is also important to recognise that whilst on a global scale bio-
PROTECTED AREA PLANNING FRAMEWORKS
diversity is suffering attrition (i.e. species are going extinct at well
— MAPPING PRESENT AND FUTURE
above background rates), at local and regional scales, patterns of
BIOGEOGRAPHIES
change are quite variable. The introduction of non-native species
and alterations of habitat typically result in the loss of particular
A typology of frameworks with particular reference
sets of species at the local scale, yet at the regional scale — and
to the global scale
sometimes locally also — the net effect may commonly be increases
in species richness (McKinney, 2002; Sax et al., 2002; Sax & It is important to recognise that in practice, the designation of
Gaines, 2003). How can such observations be reconciled? protected areas has been and will continue to be influenced by

Figure 6 Incidence curves for six area-

sensitive woodland species in different
landscape matrices in the ACT region of
southeastern Australia. The three woodland
landscapes surveyed were: agricultural (n = 55
fragments), modified (peri-urban) (n = 42)
and urban (n = 30). Incidence functions were
developed using the results of a logistic
regression based on the presence/absence of
each species in woodland remnants in each
landscape. From Watson (2004).
diverse and complex environmental, socio-economic and politi- (Demeritt, 2001), and the choice of criteria and variables used to
cal factors (e.g. Sutherland, 1998; Jepson & Whittaker, 2002b). create the maps will be strongly influenced by what the creators
But in the following account, in the interests of simplification, we of the map value (Williams et al., 2000a).
focus exclusively on the scientific principles underpinning the In the 1960s there was widespread support within the conser-
designation of protected areas. vation community for the goal of establishing a worldwide
The concept that the world is a patchwork of regions differing network of natural reserves encompassing representative areas of
in their biological makeup has profoundly influenced how the world’s ecosystems. Under the aegis of the IUCN, Dasmann
humanity perceives and interacts with the natural world. The (1972, 1973) and Udvardy (1975) put this ‘biogeographic repre-
concept of biological regions emerged as a result of European sentation principle’ into practice, by extending and combining
expansion in the tropics from the 16th century onwards and the earlier maps of faunal regions (Wallace) and vegetation zones
desire to understand the diversity, richness and patterns of nature (Clements) to create a nested hierarchy of biological regions (Table 1,
for religious, intellectual and economic purposes (Lomolino et al., Jepson & Whittaker, 2002a). Their work is foundational within
2004). The realisation that nature differed in different parts of conservation biogeography and provided the framework for
the world, and the subsequent spatial representations of these a massive and rapid expansion of protected areas globally.
differences, played an important role in the planning and execu- The Dasmann–Udvardy framework subdivides the globe into
tion of the European colonial endeavour, the creation of national faunal regions (biotic realms) within which a biome classifica-
identities, and the development of international tourism among tion system is applied, with biotic provinces delineated by subdi-
other things. Maps play an important role in the communication viding a physiognomically defined climax vegetation type on the
of biogeographic theory and are powerful tools within conserva- basis of a distinctive fauna. Areas with less than 65% of their spe-
tion biogeography (cf. Dalton, 2000; Myers & Mittermeier, cies in common are delimited as separate faunal provinces, and
2003). But, maps are only a symbolic representation of nature this essentially zonal scheme is afforced by the recognition of

azonal features such as high mountains and mountainous islands. the scheme must be transparent and explicit. Second, the data on
As originally developed, the approach was confined to bird and which the frameworks are built must be available and of a quality
mammal data and to a coarse scale of application, but it was sub- sufficient to the task, involving both biogeographical data and
sequently refined for use at a finer scale, by using the same algo- environmental data. Third, schemes once developed should be
rithm for smaller geographical units (Mackinnon & Wind, 1981; subject to further, independent tests of their efficiency and
Mackinnon & Mackinnon, 1986a,b; Jepson & Whittaker, 2002a). robustness (as e.g. Stoms, 1994; Williams et al., 1999). In prac-
This scheme provides a transparent methodology with a clearly tice, the broad framework for biogeographic representation is
defined purpose, and in principle the analysis can be repeated to generally agreed at the global scale (although, see Cox, 2001) and
assess the implications of, for instance, using updated distributional it is arguably at the meso-scale that improvements in data quality
data, a different system of biogeographic regions (cf. Cox, 2001), might produce significant alterations in the maps and adjust-
or threshold similarity value. ments in conservation efforts.
Since the development of the biogeographic regions approach, An alternative approach to reliance on biogeographic data
the tools for biogeographic research have undergone a radical sensu stricto was formalised by Faith & Walker (1996a,b), who
transformation, as computerised databases and advanced spatial based their approach on the correspondence between increasing
analytical approaches have been developed. Whilst the Dasmann– numbers of (and/or increasing difference between) environmen-
Udvardy approach remains an important foundational scheme tal domains and increased biodiversity. They noted that sampling
at a global scale (Table 1), it has in practical terms been super- environmental pattern, or compositional variation within one
seded at this scale by other global schemes developed and pro- indicator group, would predict compositional variation within
moted by major conservation NGOs, whilst within regions a far other groups if they spanned the entire range of habitats or envi-
greater array of approaches have been developed both by NGOs ronments available. Hence, sampling environmental pattern (or
and other actors from the scientific community (e.g. Redford assemblage pattern) itself would be an alternative to selecting
et al., 2003). In Table 1 we identify the WWF Ecoregions scheme areas using species distribution data directly. Whilst this approach
as a prominent global scheme based on the principle of biogeo- has some merits, its effectiveness in achieving pre-determined
graphic representation (e.g. Dinerstein et al., 1995). It incorpo- goals has yet to be consistently demonstrated (compare Faith and
rates data on biogeography, habitat type and elevation to identify Walker, 1996a,b; Ferrier & Watson, 1997; Araújo et al., 2001,
biogeographic units at a finer scale than the original Dasmann– 2004b). For example, Araújo et al. (2001) showed that European
Udvardy framework. The approach aims to meet the goals of plants exhibited consistent non-random positive patterns of re-
representation and also to draw up natural units within which presentation with conservation areas selected to maximise pattern
ecological flows and linking processes are maintained. In this variation among environmental domains, but that terrestrial
way, the approach combines both compositionalist and func- vertebrates (especially reptiles and amphibians) were consist-
tionalist perspectives (as Callicott et al., 1999; Williams & Araújo, ently under-represented with this approach. They argued that
2000, 2002). The WWF Ecoregions approach, specifically as the degree of success of environmental-diversity strategies would
applied within Indonesia, has been criticised by Jepson & Whit- be contingent on the contribution of historical biogeography
taker (2002a) for a variety of reasons, including data deficiencies shaping current distributions. According to this view, taxa with
and a lack of transparency as to the criteria involved in fine-scale poor dispersal abilities (e.g. reptiles and amphibians) would be
designations. No schemes are entirely free of such problems, and in clear disadvantage compared with those of good dispersal
Wikramanayake et al. (2002) provide a defence of the approach abilities (e.g. plants), which would more easily achieve conditions
in an equally forthright reply. The scheme’s proponents are of of quasi-equilibrium with current environmental conditions,
course driven by the heavy responsibility of getting on with the and hence would be better predicted by environmental-based
job of putting theory into practice, and effecting conservation surrogates. Although evidence of the ability of environmental-
actions on the ground. But, it would be naïve to think of any such and assemblage-based models to represent biodiversity at a rate
map as being the final say in land-use planning and designation, higher than expected by chance is currently lacking (Araújo et al.,
as such battles have to be constantly re-fought and occasionally 2003), the approach is still strongly advocated by its proponents
land abandonment provides new opportunities for ecological (e.g. Faith, 2003; Faith et al., 2004). This is one example where
restoration and/or for inclusion of new areas into protected area conservation values, epistemology, theories of equilibrium in
networks (cf. Meir et al., 2004). Magnusson (2004) argues that as biogeography and numerical ecological techniques are inter-
ecoregion approaches are being heavily promoted, more rigor- mingled in a complex but possibly important debate within
ous tests of their delineation are urgently needed. He was writing Conservation Biogeography.
about an attempt to test the validity of Bailey’s foundational Hotspot approaches at the global scale contrast with biogeo-
ecoregions scheme, but the same case can be made for the WWF graphical representation approaches by focusing on the richness
scheme. Magnusson suggests a simple approach involving look- and endemism of areas combined with a measure of the threat to
ing for natural breaks in species distributions across ecoregion biodiversity (Table 1). There are a number of prominent global
boundaries, and a more sophisticated approach to the problem is schemes, including: the IUCN–WWF Centres of Plant Diver-
illustrated by Williams (1996) and Williams et al. (1999). sity, Birdlife International’s Endemic Bird Areas (e.g. Long et al.,
To improve further on the alternative representation frame- 1996; Stattersfield et al., 1998), and Conservation International’s
works available requires several components. First, the criteria of (CI) hotspots (Myers et al., 2000). Applications of the hotspots

approach are based, partly through necessity, on a limited array ter serve these goals (cf. Mace et al., 2000; Williams et al., 2000a;
of taxa. For instance, the 25 CI hotspots are delimited solely on Rodrigues et al., 2004). Moreover, as such large sums and efforts
the basis of plant endemism and habitat conversion statistics. are involved, it is surely crucial that we provide the checks and
Myers et al. (2000) claim that the hotspots hold 44% of the balances to ensure that the resources are deployed effectively, and
world’s plant species and 35% of the vertebrates in 12% of the one part of this is about a continuing process of reviewing the
land area of the earth. These taxa, however, represent a relatively scientific basis (Meir et al., 2004).
small proportion of the species on earth, and it remains to be To complete our typology of global protected area approaches
established what proportion of, for example, insects, are found (Table 1), we have designated a third type of approach — ‘impor-
within these areas. The areas delimited are also far too coarse, of tant areas’ — as being distinct from representation and hotspot
themselves, to guide issues such as reserve placements. Rather they approaches. As exemplified by Birdlife International’s Important
provide a global ‘cookie cutter’ of large areas deemed most deserv- Bird Areas (IBAs) scheme, important areas are key sites for con-
ing of conservation funding and attention. The approach has servation, small enough to be conserved in their entirety and
been remarkably successful in terms of fund raising, and if often already part of a protected-area network. The stated criteria
successful in implementation could make a significant contribu- for selection of IBA sites are that they do one or more of the fol-
tion to reducing global biodiversity loss (Myers & Mittermeier, lowing things: (i) they hold significant numbers of one or more
2003). globally threatened species; (ii) they are one of a set of sites that
The CI hotspots approach can be criticised on several grounds, together hold a suite of restricted-range species or biome-
including: (i) that it serves only a limited set of values (Jepson & restricted species; and (iii) they have exceptionally large numbers
Canney, 2001); (ii) that it sends out a powerful if entirely unin- of migratory or congregatory species (http://www.birdlife.net /
tentional signal that biodiverse areas excluded from the list of action /science /sites/ visited September 2004). The scheme pro-
hotspots do not matter in conservation terms (Bates & Demos, vides a fine-scale network of sites, below the level of resolution
2001); (iii) that it is a very coarse-resolution and simplistic ana- of Birdlife’s Endemic Bird Areas approach: for instance there
lysis (Mace et al., 2000); and (iv) that there seems to be no under- are some 4000 IBA sites in Europe alone, in contrast to 218 EBAs
lying base map, and certainly nothing approaching an equal-area globally.
grid cell system is involved, so that the basis for delimiting the In moving from the global to the regional scale of application,
boundaries of the areas selected appears essentially arbitrary. it is clear that there are numerous approaches to mapping and
Moreover, in examining the applicability of the approach to the prioritising putative or existing protected areas (e.g. Pressey &
marine realm, Hughes et al. (2002) find that first, richness peaks Nicholls, 1989; Margules & Pressey, 2000; Williams et al., 2000b;
do not coincide with centres of high endemicity in Indo-Pacific Cabeza & Moilanen, 2001; Williams & Araújo, 2002; Dimitrako-
corals and reef fishes, and second, that fish and coral endemicity poulos et al., 2004). Typically these approaches take grid-cell
patterns are at variance with one another. They therefore argue based species range data and use reiterative computer algorithms
against marine prioritisation by ‘hotspots’, instead calling for a to select complementary sets of cells that achieve a predeter-
focus on preserving connectivity and genetic diversity of widely mined goal. These approaches have the advantage of making
dispersed species combined with intensive protection of quite goals, values and priorities explicit before area priorities are
localised areas of high endemicity. defined. They are quantitative, repeatable and ensure that effi-
Criticisms have also been levelled at the way in which CI’s 25 cient and effective solutions are obtained for a set of pre-selected
selected terrestrial hotspots have been ranked internally (Brummitt goals and data (Williams et al., 2000a). Unfortunately, they are
& Lughadha, 2003; Ovadia, 2003), with Brummitt & Lughadha also very data-hungry and highly sensitive to data quality
correctly arguing that the use of unscaled species–area ratios to (Flather et al., 1997; Freitag et al., 1998; Araújo, 2004; Araújo
select the ‘hottest hotspots’ is inappropriate. Myers & Mitter- et al., 2004b). Hence, quantitative ‘gap analyses’ have until very
meier (2003) sidestep the scientific criticism in their response to recently been restricted to problems of regional extent. The
Brummitt & Lughadha (2003), pointing to the urgency of action application of this approach at a global scale by Rodrigues et al.
for conservation, the $750 m raised in support funds for CI’s (2004) using data for several groups of terrestrial vertebrates
hotspots scheme, and expressing frustration at the slowness of thus represents an important development, supplementing the
the scientific community to engage with their approach. The ten- approaches identified in Table 1.
sion between the priorities of major NGOs and academic debate
is understandable. The point of our listing the above criticisms of
INTEGRATING PATTERN AND PROCESS
the WWF Ecoregions and CI hotspots programmes is not to
undermine them as a basis for taking conservation action now, In an earlier section of this paper we discussed how varied the
but to encourage an ongoing effort from conservation biogeo- protected area system outcomes can be from taking different
graphers to update, revise and test protected area planning frame- approaches to selecting areas, dependent for instance on the taxo-
works at global and also regional scales. We see this as valuable nomic level, and whether the emphasis is on richness, endemism,
because first, the broader goals of nature conservation are such or perceived threats. This type of approach can be characterised
that we need alternative sets of maps, capturing and representing as compositionalist, as such analyses provide an essentially static
different value sets, and second, because as our data, theory, and assessment of varying aspects of the present composition of the
models improve, we can refine our planning frameworks to bet- biota and seek to preserve the option value that comes from variety

(Williams & Araújo, 2000). We can contrast these approaches insights into the dynamics of range change should be fed into
with more ‘functionalist’ approaches (Williams & Araújo, 2000, conservation planning (Channell & Lomolino, 2000; Araújo
2002), which focus on the potential viability of protected area et al., 2002b; Ceballos & Ehrlich, 2002; Rodríguez, 2002), but it is
networks projected into the future. Recently, attempts have been clearly an area worthy of more research efforts by conservation
made within these grid-cell based ‘complementarity’ analyses to biogeographers. To come full circle with an earlier theme, it is
combine compositionalist and functionalist criteria based on well established in the biogeographical and palaeoecological
assessments of the likelihood of persistence of populations in literature that across much of the globe ecological systems have
occupied grid cells into the future (e.g. Cowling et al., 1999; been continually resorted and reconstituted in response to pro-
Araújo et al., 2002a; Williams & Araújo, 2002; Rouget et al., nounced climatic change over the Quaternary period. Species
2003). In particular, accelerated climate change creates an urgent respond individualistically to these changes, and so species range
need to model species distributions and biogeographic patterns locations, shapes and sizes have fluctuated through time. A com-
under multiple future scenarios (above). bination of documentary and palaeoecological evidence provides
Another interesting theme combining process with pattern the potential for improved insights into how species ranges change
emerges from recent work on the geography of range collapse as in size and configuration in response to both anthropogenic and
species head towards extinction. On the grounds that the centres natural forcing over a range of time frames.
of the range might be anticipated to provide optimal conditions
for a species and the margins suboptimal habitat, it was thought
CONCLUSIONS
likely that species would persist longest near their range centres
(e.g. Brown, 1984; Curnutt et al., 1996). This idea has led to the Biogeographical traditions of thinking and analysis are impor-
development of quantitative reserve-selection methods favour- tant in many areas of conservation science, from coarse to fine
ing conservation of species within the core of their current distri- scales of resolution of pattern, but are most evident and most sig-
butions (e.g. Araújo & Williams, 2000). Empirical evidence in nificant working from the landscape, up to regional and thence
support of this approach was offered in the form of observations to global scales of analysis. In the present paper, we have defined
that extinction probabilities of British-bird species in a particular conservation biogeography in broad terms, but have necessarily
20-years’ period were more likely in the edges of most species illustrated the paper with only a limited set of topics, paying
ranges than in the core (Araújo et al., 2002a). This approach to particular attention to strategic conservation planning. We think
reserve selection corrects the tendency of complementarity-based the latter focus is justified because of the contribution of bio-
approaches to select ‘ecotonal’ areas in the intersection of many geographical analyses in this important arena. We have also
species ranges (Gaston et al., 2001; Araújo & Williams, 2001). stressed that in this, as in other areas of conservation science, the
Paradoxically, some authors argue that reserves should indeed be approach taken in the analysis and interpretation necessarily
located in areas of transition (Smith et al., 2001; Spector, 2002) involves the adoption of particular systems of social values, and
because this is where species’ adaptation to future environmental that it is important to recognise this fact and to strive to provide
changes is more likely. This discussion flags some of the most dif- guidance serving alternate valuations of nature and natural
ficult debates in conservation biogeography: although it is true resources.
that species in transition zones are more likely to be subject to By selecting four principal themes by which to organise issues
selective pressures that may favour genotypes with higher capa- concerning the sensitivity of conservation biogeography to
city to adapt in suboptimal conditions (e.g. produced by adverse assumptions we hope to have highlighted important areas for
climatic change conditions), it is also true that core populations future work, but have doubtless thereby neglected others. For
are generally better adapted to cope with current environmental instance, we have not explicitly discussed the many technical
conditions. Ideally both core and peripheral populations should issues concerned with the proper analysis of spatially distributed
be conserved, but conservation resources are scarce and this data, and especially the concerns over spatial autocorrelation,
degree of representation may not be achievable for more than a and other artefacts such as the so-called geometric constraints
handful of species (Araújo, 2002). problem (see, e.g. Diniz-Filho et al., 2003; Hawkins & Porter,
To complicate the issue further, Channell & Lomolino (2000) 2003; Colwell et al., 2004; Dark, 2004).
have undertaken analyses of 309 declining species, using historical To conclude this assessment, we may ask ‘Where do we stand
data, and have found that the predominant pattern is consistent in conservation biogeography?’ At one extreme we have authors
with the ‘contagion hypothesis’ of range collapse, i.e. the extinc- who articulate the ‘crisis’ nature of the biodiversity conservation
tion process first impacts a peripheral population, but then moves problem. They argue that we have all the data we need in order to
to drive more central populations to extinction, so that the last make the key strategic decisions on mapping the biosphere and
populations to persist are to be found (perhaps in isolates) prioritising areas for protection at global and regional scales
towards a point on the historic range margin. Their analyses are alike, we have to take decisions now to save as much as we can,
based on recorded patterns of range change across continents and we have long established principles to guide reserve network
and islands over long periods (centuries) and it remains to be design within regions, in terms of reserve sizes, connectivity,
shown if similar patterns and processes operate at fine scales of corridors, etc. Unsurprisingly, NGO-based proponents of global
space and time (e.g. Rodríguez, 2002) or under conditions of protected area planning frameworks take this sort of position,
particular directional shifts in climate. We are unsure how these and emphasise the importance of developing ‘… a “grand synthesis”

of global priority-setting approaches so that we can speak with Andelman, S.J. & Fagan, W.F. (2000) Umbrellas and flagships:
one voice to decision makers and donors about the fate of the efficient conservation surrogates or expensive mistakes?
biota on our planet’ (Wikramanayake et al., 2002, 242). Proceedings of the National Academy of Sciences, USA, 97,
At the other end of this debate we may imagine the sceptic who 5954–5959.
contends that we do not yet know to within one, possibly two Araújo, M.B. (2002) Biodiversity hotspots and zones of ecologi-
orders of magnitude what the present global species tally is, we cal transition. Conservation Biology, 16, 1662–1663.
lack systematic species range data of global extent (even for the Araújo, M.B. (2004) Matching species with reserves — uncertain-
best known taxa), that patterns of endemism are (in places) ties from using data at different resolutions. Biological
strongly correlated with patterns of collecting intensity, that the- Conservation, 118, 533–538.
oretical guidelines for reserve systems are based on equivocal Araújo, M.B., Cabeza, M., Thuiller, W., Hannah, L. & Williams, P.H.
empirical data, and that there is huge uncertainty attached to our (2004a) Would climate change drive species out of reserves?
assumptions and models for protected area planning, and that in An assessment of existing reserve selection methods. Global
any event, no single conservation map can serve the diverse goals Change Biology, 10, 1618–1626.
society may require. Moreover, such a sceptic may observe that Araújo, M.B., Humphries, C.J., Densham, P.J., Lampinen, R.,
international conservation NGOs have their own agendas, and Hagemeijer, W.J.M., Mitchell-Jones, A.J. & Gasc, J.P. (2001)
need prioritisation schemes to ‘sell’ in the market place of biodi- Would environmental diversity be a good surrogate for species
versity fund raising. These organisations have simultaneously diversity? Ecography, 24, 103–110.
engaged in developing the science, and advocating the outcomes, Araújo, M.B., Densham, P.J. & Humphries, C.J. (2003) Predicting
and have vested interests in defending the schemes they have species diversity with ED: the quest for evidence. Ecography,
developed and branded. Whilst we have highlighted many uncer- 26, 380–383.
tainties and critical questions within the present article, we do Araújo, M.B., Densham, P.J. & Williams, P.H. (2004b) Represent-
not identify ourselves with such an extreme sceptical position: ing species in reserves from patterns of assemblage diversity.
specifically we do not question the need for conservation action, Journal of Biogeography, 31, 1037–1050.
we applaud the efforts of those in the NGO community who are Araújo, M.B., Thuiller, W., Williams, P.H. & Reginster, I. (2005)
so determined to do something about the problem, and we also Downscaling European species atlas distributions to a finer
welcome the emphasis on local and national leadership inherent, resolution: implications for conservation planning. Global
for example, in the recently floated multi-NGO Key Biodiversity Ecology and Biogeography, in press. doi: 10.1111/j.1466-
Areas initiative. 822X.2004.00128.x.
We believe that the best defence against such sceptical views Araújo, M.B. & Williams, P. (2000) Selecting areas for species
gaining hold (and they can readily be found in the wider dis- persistence using occurrence data. Biological Conservation, 96,
course of the global internet) is for a broad section of the biogeo- 331–345.
graphy community to engage in independent development and Araújo, M.B. & Williams, P.H. (2001) The bias of complementa-
evaluation of the scientific guidelines for conservation, spanning rity hotspots toward marginal populations. Conservation
all aspects of conservation biogeography, and especially focussed Biology, 15, 1710–1720.
on assessing the sensitivity of conservation guidelines to the Araújo, M.B., Williams, P.H. & Fuller, R.J. (2002a) Dynamics of
assumptions inherent in our theories, models and analyses. We extinction and the selection of nature reserves. Proceedings of
hope that this paper will make a small contribution towards the Royal Society of London, Series B, 269, 1971–1980.
developing this research programme. Such a programme offers Araújo, M.B., Williams, P.H. & Turner, A. (2002b) A sequential
enticing intellectual challenges and is of the greatest societal rele- approach to minimise threats within selected conservation
vance. In short, we see conservation biogeography as having a central areas. Biodiversity and Conservation, 11, 1011–1024.
role to play in the conservation endeavour in the coming decades. Bates, J.M. & Demos, T.C. (2001) Do we need to devalue Amazo-
nia and other large tropical forests? Diversity and Distributions,
7, 249–255.
ACKNOWLEDGEMENTS
Bennett, K.D. & Willis, K.J. (2000) Effect of global atmospheric
We thank for discussion and /or comment on the manuscript: carbon dioxide on glacial-interglacial vegetation change.
D. Brockington, J.M. Fernández-Palacios, S. Ferrier, N. O’Dea, Global Ecology and Biogeography, 9, 355–361.
R.G. Pearson, J.P. Rodríguez, P.H. Williams and two anonymous Bierregaard, R.O., Gascon, C., Lovejoy, T.E. & Mesquita, R.
reviewers. (2001) Lessons from Amazonia: the ecology and conservation of
a fragmented forest. Yale University Press, New Haven.
Boecklen, W.J. (1986) Effects of habitat heterogeneity on the
REFERENCES
species–area relationships of forest birds. Journal of Biogeo-
Adams, W. (2004) Against extinction: the past and future of con- graphy, 13, 59 –68.
servation. Earthscan, London. Boecklen, W.J. & Gotelli, N.J. (1984) Island biogeographic
AFE. (2003) Atlas florae Europaeae project, http:// theory and conservation practice: species–area or specious-area
www.fmnh.helsinki.fi/map/afe/E_afe.htm [accessed 22nd Sept relationships? Biological Conservation, 29, 63 –80.
2003]. Brooks, D.R. & van Veller, M.C.P. (2003) Critique of parsimony

analysis of endemicity as a method of historical biogeography. ments for a sustainable system of conservation areas in the
Journal of Biogeography, 30, 819 – 825. species-rich Mediterranean-climate desert of southern Africa.
Brooks, T. & Balmford, A. (1996) Atlantic forest extinctions. Diversity and Distributions, 5, 51 –71.
Nature, 380, 115. Cox, C.B. (2001) The biogeographic regions reconsidered.
Brooks, T.M., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, Journal of Biogeography, 28, 511–523.
G.A.B., Rylands, A.B., Konstant, W.R., Flick, P., Pilgrim, J., Curnutt, J.L., Pimm, S.L. & Maurer, B.A. (1996) Population
Oldfield, S., Magin, G. & Hilton-Taylor, C. (2002) Habitat loss variability of sparrows in space and time. Oikos, 76, 131–
and extinction in the hotspots of biodiversity. Conservation 144.
Biology, 16, 909 –923. Dalton, R. (2000) Biodiversity cash aimed at hotspots. Nature,
Brown, J.H. (1984) On the relationship between abundance and 406, 818.
distribution of species. The American Naturalist, 124, 255–279. Dark, S.J. (2004) The biogeography of invasive alien plants in
Brown, J.H. (1995) Macroecology. The University of Chicago California: an application of GIS and spatial regression analy-
Press, Chicago. sis. Diversity and Distributions, 10, 1– 9.
Brown, J.H. & Lomolino, M.V. (1998) Biogeography, 2nd edn. Dasmann, R.F. (1972) Towards a system for classifying natural
Sinauer Press, Sunderland, Massachusetts. regions of the world and their representation by national parks
Brummitt, N. & Lughadha, E.N. (2003) Biodiversity: where’s hot and reserves. Biological Conservation, 4, 247–255.
and where’s not. Conservation Biology, 17, 1442 – 1448. Dasmann, R.F. (1973) A system for defining and classifying natural
Buckley, L.B. & Roughgarden, J. (2004) Effects of changes in regions for the purposes of conservation. World Conservation
climate and land use. Nature, (01 July 2004); doi:10.1038/ Union, Morges, Switzerland.
nature02717. Dehnen-Schmutz, K. (2004) Alien species reflecting history:
Bush, M.B. (1994) Amazonian speciation — a necessarily complex medieval castles in Germany. Diversity and Distributions, 10,
model. Journal of Biogeography, 21, 5 – 17. 147–151.
Bush, M.B. (2002) Distributional change and conservation on Diamond, J.M. (1975) The Island Dilemma: lessons of modern
the Andean flank: a palaeoecological perspective. Global biogeographic studies for the design of nature reserves. Biolog-
Ecology and Biogeography, 11, 463 – 473. ical Conservation, 7, 129–146.
Bush, M.B. & Whittaker, R.J. (1991) Krakatau: Colonization Demeritt, D. (2001) Scientific forest conservation and the statis-
patterns and hierarchies. Journal of Biogeography, 18, 341– tical picturing of Nature’s limits in the progressive-era United
356. States. Environment and Planning D: Society and Space, 19,
Cabeza, M. (2003) Habitat loss and connectivity of reserve 431–459.
networks in probability approaches to reserve design. Ecology Dinerstein, E., Olson, D.M., Graham, D.J., Webster, A.L.,
Letters, 6, 665 –672. Pimm, S.A., Bookbinder, M.A. & Ledec, G. (1995) A conserva-
Cabeza, M. & Moilanen, A. (2001) Design of reserve networks tion assessment of the terrestrial Ecoregions of Latin America and
and the persistence of biodiversity. Trends in Ecology and the Caribbean. The World Bank, Washington, D.C.
Evolution, 16, 242 – 247. Diniz-Filho, J.A.F., Bini, L.M. & Hawkins, B.A. (2003) Spatial
Cabeza, M., Araújo, M.B., Wilson, R.J., Thomas, C.D., autocorrelation and red herrings in geographical ecology.
Cowley, M.J.R. & Moilanen, A. (2004) Combining probabil- Global Ecology and Biogeography, 12, 53 –64.
ities of occurrence with spatial reserve design. Journal of Dimitrakopoulos, P.G., Memtsas, D. & Troumbis, A.Y. (2004)
Applied Ecology, 41, 252 – 262. Questioning the effectiveness of the Natura 2000 Special Areas
Callicott, J.B., Crowder, L.B. & Mumford, K. (1999) Current of Conservation strategy: the case of Crete. Global Ecology and
normative concepts in conservation. Conservation Biology, Biogeography, 13, 199–207.
13, 22 – 35. Ebach, M.C., Humphries, C.J. & Williams, D.M. (2003) Phylo-
Ceballos, G. & Ehrlich, P.R. (2002) Mammal population losses genetic biogeography deconstructed. Journal of Biogeography,
and the extinction crisis. Science, 296, 904 – 907. 30, 1285–1296.
Channell, R. & Lomolino, M.V. (2000) Trajectories to extinction: Faith, D.P. (2003) Environmental diversity (ED) as a surrogate
spatial dynamics of the contraction of geographical ranges. information for species-level biodiversity. Ecography, 26, 374–
Journal of Biogeography, 27, 169 – 179. 379.
Colinvaux, P.A., Irion, G., Rasanen, M.E., Bush, M.B. & de Faith, D.P., Ferrier, S. & Walker, P.A. (2004) The ED strategy:
Mello, J.A.S.N. (2001) A paradigm to be discarded: geological how species-level surrogates indicate general biodiversity pat-
and paleoecological data falsify the Haffer & Prance refuge terns through an ‘environmental diversity’ perspective. Journal
hypothesis of Amazonian speciation. Amazoniana, 16, 609– of Biogeography, 31, 1207–1217.
646. Faith, D.P. & Walker, P.A. (1996a) How do indicator groups pro-
Colwell, R.K., Rahbek, C. & Gotelli, N.J. (2004) The mid-domain vide information about the relative biodiversity of different
effect and species richness patterns: what have we learned so sets of areas?: on hotspots, complementarity and pattern-based
far? American Naturalist, 163, E1 – E23. approaches. Biodiversity Letters, 3, 18 –25.
Cowling, R.M., Pressey, R.L., Lombard, A.T., Desmet, P.G. & Faith, D.P. & Walker, P.A. (1996b) Environmental diversity: on
Ellis, A.G. (1999) From representation to persistence: require- the best-possible use of surrogate data for assessing the relative

biodiversity of sets of areas. Biodiversity and Conservation, 5, Hawkins, B.A., Porter, E.E. & Diniz-Filho, J.A.F. (2003) Produc-
399 –415. tivity and history as predictors of the latitudinal diversity gradient
Ferrier, S. & Watson, G. (1997) An evaluation of the effectiveness of terrestrial birds. Ecology, 84, 1608–1623.
of environmental surrogates and modelling techniques in predict- Honnay, O., Hermy, M. & Coppin, P. (1999) Nested plant com-
ing the distribution of biological diversity. Report produced for munities in deciduous forest fragments: species relaxation or
the Department of Environment, Sports and Territories, nested subsets? Oikos, 84, 119–129.
Australia. Hopkinson, P., Travis, J.M.J., Evans, J., Gregory, R.D., Telfer, M.G.
Flather, C.H., Wilson, K.R., Dean, D.J. & McComb, W.C. (1997) & Williams, P.H. (2001) Flexibility and the use of indicator
Identifying gaps in conservation networks: of indicators and taxa in the selection of sites for nature reserves. Biodiversity
uncertainty in geographic-based analysis. Ecological Applica- and Conservation, 10, 271–285.
tions, 7, 531–542. Hughes, T.P., Bellwood, D.R. & Connolly, S.R. (2002) Biodiver-
Frank, D.J., Hironaka, A., Meyer, J.W., Schoffer, E. & Tuma, N.B. sity hotspots, centres of endemicity, and the conservation of
(1999) The rationalization and organisation of Nature in coral reefs. Ecology Letters, 5, 775–784.
World culture. Constructing World culture: international Humphries, C.J. (2001) Hotspots: going off the boil? Diversity
non-government organisations since 1875 (ed. by J. Boli and and Distributions, 7, 104–105.
G.M. Thomas), pp. 81 – 96. Stanford University Press, Stan- Huntley, B.J. (1996) Biodiversity conservation in the New South
ford, California. Africa. Biodiversity, science and development: towards a new
Freitag, S., Nicholls, A.O. & van Jaarseveld, A.S. (1998) Dealing partnership (ed. by F. di Castri and T. Younès), pp. 282– 303.
with established reserve networks and incomplete distribution CABI, Wallingford, UK.
data sets in conservation planning. South African Journal of Jepson, P., Brickle, N. & Chayadin, Y. (2001) The conservation
Science, 94, 79 – 86. status of Tanimbar corella and blue-streaked lory on the Tanim-
Gaston, K.J. (1994) Rarity. Chapman & Hall, London. bar Islands, Indonesia: results of a rapid contextual survey.
Gaston, K.J., Rodrigues, A.S., van Rensburg, B.J., Koleff, P. & Oryx, 35, 224–235.
Chown, S.L. (2001) Complementary representation and zones Jepson, P. & Canney, S. (2001) Biodiversity hotspots: hot for
of ecological transition. Ecology Letters, 4, 4 – 9. what? Global Ecology and Biogeography, 10, 225–227.
Gerber, S., Joos, F. & Prenctice, I.C. (2004) Sensitivity of a Jepson, P. & Canney, S. (2003) Values-led conservation. Global
dynamic global vegetation model to climate and atmospheric Ecology and Biogeography, 12, 271–274.
CO2. Global Change Biology, 10, 1223 – 1239. Jepson, P. & Whittaker, R.J. (2002a) Ecoregions in context: a cri-
Gillson, L. & Willis, K.J. (2004) ‘As Earth’s testimonies tell’: tique with special reference to Indonesia. Conservation Biology,
wilderness conservation in a changing world. Ecology Letters, 16, 42 –57.
7, 990 –998. Jepson, P. & Whittaker, R.J. (2002b) Histories of Protected Areas:
Gonzalez, A. (2000) Community relaxation in fragmented internationalism of conservationist values and their adoption
landscapes: the relation between species richness, area and age. in the Netherlands Indies (Indonesia). Environment and
Ecology Letters, 3, 441 – 448. History, 8, 129–172.
Gray, J.S., Ugland, K.I. & Lambshead, J. (2004a) Species accumu- Koleff, P. & Gaston, K.J. (2002) The relationships between local
lation and species area curves — a comment on Scheiner (2003). and regional species richness and spatial turnover. Global Ecology
Global Ecology and Biogeography, 13, 473 – 476. and Biogeography, 11, 363–375.
Gray, J.S., Ugland, K.I. & Lambshead, J. (2004b) On species- Kullman, L. (1998) Palaeoecological, biogeographical and palae-
accumulation and species-area curves. Global Ecology and oclimatological implications of early Holocene immigration of
Biogeography, 13, 567 – 568. Larix sibirica Ledeb into the Scandes Mountains, Sweden.
Groombridge, B. (ed.) (1992) Global biodiversity: status of the Global Ecology and Biogeography Letters, 7, 181–188.
Earth’s living resources. (A report compiled by the World Con- Ladle, R., Jepson, P., Araújo, M.B. & Whittaker, R.J. (2004)
servation Monitoring Centre). Chapman and Hall, London. Dangers of crying wolf over risk of extinctions. Nature,
Groombridge, B. & Jenkins, M.D. (2000) Global biodiversity: 482, 799.
Earth’s living resources in the 21st Century. World Conservation Lambshead, P.J.D. & Boucher, G. (2003) Marine nematode deep-
Press, Cambridge. sea biodiversity — hyperdiverse or hype? Journal of Biogeography,
Hambler, C. (2004) Virgin rainforests and conservation. Science, 30, 475–485.
305, 943. Lawton, J.H. & May, R.M. (eds) (1995) Extinction rates. OUP, Oxford.
Hannah, L., Midgley, G.F. & Millar, D. (2002) Climate change- Lennon, J.J., Koleff, P., Greenwood, J.J.D. & Gaston, K.J. (2001)
integrated conservation strategies. Global Ecology and Biogeog- The geographical structure of British bird distributions: diver-
raphy, 11, 485 –495. sity, spatial turnover and scale. Journal of Animal Ecology, 70,
Hannah, L. & Phillips, B. (2004) Extinction-risk coverage is 966–979.
worth inaccuracies. Nature, 430, 141. Lindenmayer, D.B. & Franklin, J.F. (2002) Conserving forest
Hawkins, B.A. & Porter, E.E. (2003) Does herbivory depend biodiversity: a comprehensive multiscale approach. Island Press,
on plant diversity? The case of California butterflies. The Washington.
American Naturalist, 161, 40 – 49. Lomolino, M.V. (2000) Ecology’s most general, yet protean

pattern: the species–area relationship. Journal of Biogeography, Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca,
29, 555 –557. G.A.B. & Kent, J. (2000) Biodiversity hotspots for conservation
Lomolino, M.V. (2002) ‘… there are areas too small, and areas priorities. Nature, 403, 853–859.
too large, to show clear diversity patterns …’ R.H. MacArthur Nekola, J.C. & White, P.S. (1999) The distance decay of similarity in
(1972: 191). Journal of Biogeography, 29, 555 –557. biogeography and ecology. Journal of Biogeography, 26, 867– 878.
Lomolino, M.V. (2004) Conservation biogeography. Frontiers of Nelson, B.W., Ferreira, C.A.C., de Silva, M.F. & Kawasaki, M.L.
Biogeography: new directions in the geography of nature (ed. (1990) Endemism centres, refugia and botanical collection
by M.V. Lomolino and L.R. Heaney), pp. 293–296. Sinauer density in Brazilian Amazonia. Nature, 345, 714–716.
Associates, Sunderland, Massachusetts. O’Brien, E.M., Field, R. & Whittaker, R.J. (2000) Climatic gra-
Lomolino, M.V. & Heaney, L.R. (eds) (2004) Frontiers of dients in woody plant (tree and shrub) diversity: water-energy
Biogeography: new directions in the geography of nature. Sinauer dynamics, residual variation, and topography. Oikos, 89, 588–
Associates, Sunderland, Massachusetts. 600.
Lomolino, M.V., Sax, D. & Brown, J.H. (eds) (2004) Foundations O’Brien, E.M., Whittaker, R.J. & Field, R. (1998) Climate and
of Biogeography. Chicago University Press, Chicaco. woody plant diversity in southern Africa: relationships at
Lomolino, M.V. & Weiser, M.D. (2001) Towards a more general species, genus and family levels. Ecography, 21, 495–509.
species–area relationship: diversity on all islands, great and Ødegaard, F. Diserud, O.H., Engen, S. & Aagaard, K. (2000) The
small. Journal of Biogeography, 28, 431– 445. magnitude of local host specificity for phytophagous insects
Long, A.J., Crosby, M.J., Stattersfield, A.J. & Wege, D.C. (1996) and its implications for estimates of global species richness.
Towards a global map of biodiversity: patterns in the distribu- Conservation Biology, 14, 1182–1186.
tion of restricted-range birds. Global Ecology and Biogeography Olson, D.M. & Dinerstein, E. (1998) The Global 200: a represen-
Letters, 5, 281–304. tation approach to conserving the Earth’s most biologically
Lourie, S.A. & Vincent, A.C.J. (2004) Using biogeography to help valuable ecoregions. Conservation Biology, 12, 502–515.
set priorities in marine conservation. Conservation Biology, 18, Ovadia, O. (2003) Ranking hotspots of varying sizes: a lesson
1004 –1020. from the nonlinearity of the species–area relationship. Conser-
MacArthur, R.H. & Wilson, E.O. (1967) The theory of island vation Biology, 17, 1440–1441.
biogeography. Princeton University Press, Princeton. Palmer, M.W. & White, P.S. (1994) Scale dependence and the
Mace, G.M., Balmford, A., Boitani, L., Cowlishaw, G., Dobson, A.P., species–area relationship. American Naturalist, 144, 717–
Faith, D.P., Gaston, K.J., Humphries, C.J., Vane-Wright, R.I., 740.
Williams, P.H., Lawton, J.H., Margules, C.R., May, R.M., Pearson, R.G. & Dawson, T.P. (2003) Predicting the impacts of
Nicholls, A.O., Possingham, H.P., Rahbek, C. & van Jaarseveld, climate change on the distribution of species: are bioclimate
A.S. (2000) It’s time to work together and stop duplicating envelope models useful? Global Ecology and Biogeography, 12,
conservation efforts. Nature, 405, 393. 361–371.
MacKinnon, J. & MacKinnon, K. (1986a) Review of the protected Pearson, R.G. & Dawson, T.P. (2004) Bioclimate envelope models:
area systems of the Afrotropical realm. IUCN and UNDP, Gland, what they detect and what they hide — response to Hampe
Switzerland. (2004). Global Ecology and Biogeography, 13, 471–473.
MacKinnon, J. & MacKinnon, K. (1986b) Review of the protected Pressey, R.L. & Nicholls, A.O. (1989) Efficiency in conservation
areas system in the Indo-Malayan Realm. IUCN, CNPPA and evaluation: scoring vs. iterative procedures. Biological Conser-
UNEP, Cambridge & Gland. vation, 50, 199–218.
MacKinnon, J. & Wind, J. (1981) Birds of Indonesia. Special Primack, R.B. (2002) Essentials of conservation biology, 3rd edn,
report F0/INS/78/061, available from the Food and Agriculture Sinauer Press, Sunderland, Massachusetts.
Organization of the United Nations, Rome. Prinzing, A., Klotz, S., Stadler, J. & Brandl, R. (2003) Woody
Magnusson, W.E. (2004) Ecoregion as a Pragmatic Tool. Conser- plants in Kenya: expanding the higher-taxon approach. Biological
vation Biology, 18, 4. Conservation, 203, 307–314.
Margules, C.R. & Pressey, R.L. (2000) Systematic conservation PyÍek, P., Richardson, D.M. & Williamson, M. (2004) Predicting
planning. Nature, 405, 243 –253. and explaining plant invasion through analysis of source area
McKinney, M.L. (2002) Do human activities raise species rich- floras: some critical considerations. Diversity and Distribu-
ness? Contrasting patterns in United States plants and fishes. tions, 10, 179–187.
Global Ecology and Biogeography, 11, 343 –348. Rahbek, C. & Graves, G.R. (2000) Detection of macro-ecological
Medellín, R.A. & Soberón, J. (1999) Predictions of mammal patterns in South American hummingbirds is affected by
diversity on four land masses. Conservation Biology, 13, 143–149. spatial scale. Proceedings of the Royal Society of London, Series
Meir, E., Andelman, S. & Possingham, H.P. (2004) Does conser- B, 267, 2259–2265.
vation planning matter in a dynamic and uncertain world? Rahbek, C. & Graves, G.R. (2001) Multiscale assessment of
Ecology Letters, 7, 615 – 622. patterns of avian species richness. Proceedings of the National
Myers, N. & Mittermeier, R.A. (2003) Impact and acceptance of Academy of Sciences, USA, 98, 4534–4539.
the hotspots strategy: response to Ovadia and to Brummitt Redford, K.H., Coppolillo, P., Sanderson, E.W., da Fonseca, G.A.B.,
and Lughadha. Conservation Biology, 17, 1449 –1450. Dinerstein, E., Groves, C., Mace, G., Maginnis, S., Mittermeier, R.A.,

Noss, R., Olson, D., Robinson, J.G., Vedder, A. & Wright, M. Smith, T.B., Kark, S., Schneider, C.J., Wayne, R.K. & Moritz, C.
(2003) Mapping the conservation landscape. Conservation (2001) Biodiversity hotspots and beyond: the need for preserv-
Biology, 17, 116 –131. ing environmental transitions. Trends in Ecology and Evolution,
Renjifo, L.M. (2001) Effect of natural and anthropogenic land- 16, 431.
scape matrices on the abundance of subAndean bird species. Soulé, M.E. (ed.) (1986) Conservation biology: the science of scar-
Ecological Applications, 11, 14 –31. city and diversity. Sinauer Press, Sunderland, Massachusetts.
Rodrigues, A.S.L., Andelman, S.J., Bakarr, M.I., Boltani, L., Spector, S. (2002) Biogeographic crossroads as priority areas for
Brooks, T.M., Cowling, R.M., Fishpool, L.D.C., da Fonseca, biodiversity conservation. Conservation Biology, 16, 1480–
G.A.B., Gaston, K.J., Hoffmann, M., Long, J.S., Marquet, P.A., 1487.
Pilgrim, J.D., Pressey, R.L., Schipper, J., Sechrest, W., Stuart, S.N., Spellerberg, I.F. & Sawyer, J.W.D. (1999) An introduction to
Underhill, L.G., Waller, R.W., Watts, M.E.J. & Yan, X. (2004) applied biogeography. Cambridge University Press, Cambridge.
Effectiveness of the global protected area network in represent- Stattersfield, A.J., Crosby, M.J., Long, A.J. & Wege, D.C. (1998)
ing species diversity. Nature, 428, 640 – 643. Endemic bird areas of the World: priorities for conservation.
Rodrigues, A.S.L., Tratt, R., Wheeler, B.D. & Gaston, K.J. (1999) Birdlife International, Cambridge.
The performance of existing networks of conservation areas in Stoms, D.M. (1994) Scale dependence of species richness maps.
representing biodiversity. Proceedings of the Royal Society of Professional Geographer, 46, 346–358.
London, Series B, 266, 1453 –1460. Stouffer, P.C. & Bierregaard, R.O., Jr. (1996) Effects of forest
Rodríguez, J.P. (2002) Range contraction in declining North fragmentation on understorey hummingbirds in Amazonian
American bird populations. Ecological Applications, 12, 238– Brazil. Conservation Biology, 9, 1085–94.
248. Stott, P. (1998) Biogeography and ecology in crisis: the urgent
Rosenfield, J.A. (2002) Pattern and process in the geographical need for a new metalanguage. Journal of Biogeography, 25, 1–2.
range of freshwater fishes. Global Ecology and Biogeography, 11, Sutherland, W.J. (ed.) (1998) Conservation science and action.
323 –332. Blackwell Science Ltd, Oxford.
Rosenzweig, M.L. (1995) Species diversity in space and time. Terborgh, J. & Winter, B. (1983) A method for siting parks and
Cambridge University Press, Cambridge. reserves with special reference to Colombia and Ecuador.
Rosenzweig, M.L. (2003) Reconciliation ecology and the future Biological Conservation, 27, 45–58.
of species diversity. Oryx, 37, 194 –205. Thomas, C.D., Cameron, A., Green, R.A. et al. (2004) Extinction
Rouget, M., Cowling, R.M., Pressey, R.L. & Richardson, D.M. risk from climate change. Nature, 427, 145–148.
(2003) Identifying spatial components of ecological and evolu- Thuiller, W., Araújo, M.B., Pearson, R.G., Whittaker, R.J.,
tionary processes for regional conservation planning in the Brotons, L. & Lavorel, S. (2004) Uncertainty in predictions of
Cape Floristic Region, South Africa. Diversity and Distribu- extinction risk. Nature (01 July 2004); doi:10.1038/nature02716.
tions, 9, 191–210. Tjørve, E. (2003) Shapes and functions of species–area curves: a
Royal Society (2003) Measuring Biodiversity for Conserva- review of possible models. Journal of Biogeography, 30, 827–835.
tion. Policy document 11/03. http://www.royalsoc.ac.uk/ Trudgill, S. (2001) Psychobiogeography: meanings of nature and
document.asp?id=1474 motivations for a democratized conservation ethic. Journal of
Sax, D.F. & Gaines, S.D. (2003) Species diversity: from global Biogeography, 28, 677–698.
decreases to local increases. Trends in Ecology and Evolution, Udvardy, M.D.F. (1975) A classification of the biogeographical
18, 561–566. provinces of the World. IUCN, Morges, Switzerland.
Sax, D.F., Gaines, S.D. & Brown, J.H. (2002) Species invasions Watson, J.E.M. (2004) Bird responses to habitat fragmentation:
exceed extinctions on islands worldwide: a comparative illustrations from Madagascan and Australian case studies.
study of plants and birds. American Naturalist, 160, 766– Unpublished D.Phil. thesis, University of Oxford.
783. Watson, J.E.M., Whittaker, R.J. & Freudenberger, D. (2005) Bird
Scheiner, S.M. (2003) Six types of species–area curves. Global community responses to fragmentation: how consistent are
Ecology and Biogeography, 12, 441– 447. they across landscapes? Journal of Biogeography, 32, in press.
Scheiner, S.M. (2004) A mélange of curves — Further dialogue Wethered, R. & Lawes, M.J. (2003) Matrix effects on bird assem-
about species–area relationships. Global Ecology and Biogeo- blages in fragmented Afromontane forests in South Africa.
graphy, 13, 479 –484. Biological Conservation, 114, 327–340.
Schmitt, T. & Krauss, J. (2004) Reconstruction of the coloniza- Wheeler, Q.D., Raven, P.H. & Wilson, E.O. (2004) Taxonomy:
tion route from glacial refugium to the northern distribution impediment or expedient? Science, 303, 285.
range of the European butterfly Polyommatus coridon Whittaker, R.J. (1998) Island biogeography: ecology, evolution, and
(Lepidoptera: Lycaenidae). Diversity and Distributions, 10, conservation. Oxford University Press, Oxford.
271–274. Whittaker, R.J. (2004) Dynamic hypotheses of richness on
da Silva, J.M.C., Cardoso de Sousa, M. & Castelleti, C.H.M. islands and continents. Frontiers of Biogeography: new
(2004) Areas of endemism for passerine birds in the Atlantic directions in the geography of nature (ed. by M.V. Lomolino
forest, South America. Global Ecology and Biogeography, 13, and L.R. Heaney), pp. 211–231. Sinauer Press, Sunderland,
85 –92. Massachusetts, USA.

Whittaker, R.J. & Heegaard, E. (2003) What is the observed rela- ness hotspots, rarity hotspots and complementary areas for
tionship between species richness and productivity? Comment. conserving diversity using British birds. Conservation Biology,
Ecology, 84, 3384 –3390. 10, 15–74.
Whittaker, R.J., Willis, K.J. & Field, R. (2001) Scale and species Williams, P.H., Hannah, L., Andelman, S., Midgley, G.F.,
richness: towards a general, hierarchical theory of species Araújo, M.B., Hughes, G., Manne, L.L., Martinez-Meyer, E. &
diversity. Journal of Biogeography, 28, 453 – 470. Pearson, R.G. (2005) Planning for climate change: identifying
Whittaker, R.J., Willis, K.J. & Field, R. (2003) Climatic-energetic minimum-dispersal corridors for the Cape Proteaceae.
explanations of diversity: a macroscopic perspective. Macro- Conservation Biology, in press.
ecology: concepts and consequences, British Ecological Society Williams, P.H., Humphries, C., Araújo, M.B., Lampinen, R.,
Symposia Series (ed. by T.M. Blackburn and K.J. Gaston), Hagemeijer, W., Gasc, J.-P. & Mitchell-Jones, T. (2000a) Ende-
pp. 107–129. Blackwell Publishing, Oxford. mism and important areas for representing European bio-
Wikramanayake, E., Dinerstein, E., Loucks, C., Olson, D., diversity: a preliminary exploration of atlas data for plants
Morrison, J., Lamoreux, J., McKnight, M. & Hedao, P. (2002) and terrestrial vertebrates. Belgian Journal of Entomology, 2,
Ecoregions in ascendence: reply to Jepson and Whittaker. 221–46.
Conservation Biology, 16, 238 –243. Williams, P.H., de Klerk, H.M. & Crowe, T.H. (1999) Interpret-
Williams, P.H. (1996) Mapping variations in the strength and ing biogeographical boundaries among Afrotropical birds:
breadth of biogeographic transition zones using species spatial patterns in richness gradients and species replacement.
turnover. Proceedings of the Royal Society of London, Series B, Journal of Biogeography, 26, 459–474.
263, 579 – 588. Williamson, M., Gaston, K.J. & Lonsdale, W.M. (2001) The
Williams, P.H. & Araújo, M.B. (2000) Integrating species and species–area relationship does not have an asymptote! Journal
ecosystem monitoring for identifying conservation priorities. of Biogeography, 28, 827–830.
European Conservation, 4, 17–18. Williamson, M., Gaston, K.J. & Lonsdale, W.M. (2002) An
Williams, P.H. & Araújo, M.B. (2002) Apples, oranges, and asymptote is an asymptote and not found in species–area
probabilities: Integrating multiple factors into biodiversity relationships. Journal of Biogeography, 29, 1713.
conservation with consistency. Environmental Modeling and Willis, K.J., Gillson, L. & Brncic, T.M. (2004a) How virgin is
Assessment, 7, 139 –151. virgin rainforest? Science, 304, 402–403.
Williams, P.H., Burgess, N. & Rahbek, C. (2000b) Flagship spe- Willis, K.J., Gillson, L. & Brncic, T.M. (2004b) Virgin rainforests
cies, ecological complementarity, and conserving the diversity and conservation: Reply. Science, 305, 944.
of mammals and birds in sub-Saharan Africa. Animal Conser- Willis, K.J. & Whittaker, R.J. (2000) The refugial debate. Science
vation, 3, 249 –260. 287, 1406–1407.
Williams, P.H., Gibbons, D.W., Margules, C.R., Rebelo, A.G., Willis, K.J. & Whittaker, R.J. (2002) Species diversity — scale
Humphries, C.J. & Pressey, R.L. (1996) A comparison of rich- matters. Science, 295, 1245–1248.

Whittaker 2005

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Whittaker 2005

Uploaded by

Copyright:

Available Formats

Diversity and Distributions, (Diversity Distrib.

) (2005) 11, 3–23

Conservation Biogeography: assessment

Biodiversity Research Group, School of ABSTRACT

guidance for biodiversity conservation (e.g. Dasmann, 1972,

© 2005 Blackwell Publishing Ltd www.blackwellpublishing.com/ddi 3

4 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 5

6 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

population size is good news. Of course, this may be a misleading

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 7

Figure 1 Some of the steps, choices and

8 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Global protected area planning frameworks

Representation Hotspots Important Areas

Parameters Vegetation formations Species richness Threatened species

Methods % dissimilarity Species counts Application of varied

Schemes Biogeographic Biodiversity Hotspots Important Bird Areas

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 9

Figure 2 Alternative maps of potential

10 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Figure 3 Sensitivity to thresholds in

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 11

12 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

considered across a wider area, encompassing more diverse

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 13

Figure 6 Incidence curves for six area-

14 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 15

16 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 17

18 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 19

20 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 21

22 Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd

Diversity and Distributions, 11, 3–23, © 2005 Blackwell Publishing Ltd 23

You might also like