April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
321
Ecological Applications, 12(2), 2002, pp. 321–334
q 2002 by the Ecological Society of America
A CONCEPTUAL GENEALOGY OF FRAGMENTATION RESEARCH:
FROM ISLAND BIOGEOGRAPHY TO LANDSCAPE ECOLOGY
YRJÖ HAILA1
Department of Regional Studies and Environmental Policy, University of Tampere, P.O. Box 607, 33101 Tampere, Finland
Abstract. The concept of habitat fragmentation has become an important theme in
conservation research, and it is often used as if fragmentation were a unitary phenomenon.
However, the concept is ambiguous, and empirical studies demonstrate a wide variety of
direct and indirect effects, sometimes with mutually opposing implications. The effects of
fragmentation vary across organisms, habitat types, and geographic regions. Such a contrast
between a schematic concept and multifaceted empirical reality is counterproductive. I
analyzed the stabilization of the schematic view of fragmentation by the early 1980s, using
a genealogical narrative as a methodological approach. The main assumptions behind the
schematic view were: (1) fragments are comparable to oceanic islands; (2) habitats surrounding fragments are hostile to a majority of the organisms; and (3) natural pre-fragmentation conditions were uniform. The stabilization loop of this view was supported by
the reduction of empirical research to species–area curve fitting, which always produced
expected results. I present a model of the dynamics of fragmentation research that shows
the schematic, island-biogeographic view as an ‘‘intellectual attractor.’’ Since the 1980s,
the theoretical presuppositions of the schematic view have been challenged, and empirical
research has become multifaceted. Fragments of a particular habitat type are viewed as
elements in a heterogeneous landscape rather than ‘‘islands’’ surrounded by a hostile ‘‘sea.’’
However, the island metaphor is still used in conservation contexts in the shape of species–
area curves. It is backed by a presupposition that human-influenced environments are
essentially different from so-called ‘‘natural’’ environments, but this is unfounded. My
suggestion is that our perspective should be broadened still further so that habitat fragmentation is viewed as a particular form of human-induced environmental degradation; I
discuss both theoretical and practical implications of this suggestion.
Key words: dynamics of research; ecological theory; ecology and environmentalism; environmental degradation; genealogy; habitat degradation; habitat fragmentation; island biogeography;
landscape ecology.
Instead of an essentially continuous forest cover,
with infrequent meadow-like openings along watercourses or small grasslands where fires had been
unusually severe, the landscape now presents the
aspect of a savana, with isolated trees, small clumps
or clusters of trees, or small groves scattered in a
matrix of artificial grassland of grains and pasture
grasses, unstable and frequently devoid of plant cover as a result of regular plowing.
—Curtis (1956)
INTRODUCTION
‘‘Habitat fragmentation’’ has become a major theme
that conservation biologists consider when characterizing human-induced ecological deterioration of the environment, as is shown by any essay collection or textbook on conservation biology, from Soulé (1986) onManuscript received 31 May 2000; revised 6 February 2001;
accepted 12 February 2001; final version received 27 February
2001. For reprints of this Invited Feature, see footnote 1, p. 319.
1 E-mail: yrjo.haila@uta.fi
wards, or declaration on the research needs in ecology
defined by environmental issues (Lubchenko et al.
1991). This observation, however, presents us with a
dilemma: the notion of ‘‘habitat fragmentation’’ is conceptually ambiguous, for three main reasons (Haila
1999a). (1) Habitat fragmentation always consists of
both reduction in the area of the original habitat and
change in the spatial configuration of what remains,
but these give rise to different ecological processes and
should, for the sake of clarity, be held separate (Haila
and Hanski 1984, Haila 1986). (2) As all natural environments are fragmented to variable degrees, and are
subjected to continuous change due to natural reasons,
no straightforward standard is available for assessing
human-caused fragmentation. (3) Different organisms
and ecological systems ‘‘experience’’ the degree of
fragmentation of a particular environment in variable,
even contradictory, ways. Hence, a multiplicity of both
spatial and temporal scales need to be considered, and
the relevant scales probably vary across species, geographic regions, and types of environment. In conser-
321
322
INVITED FEATURE
vation thinking, however, fragmentation is often presented as a unitary phenomenon, always happening in
the same way and having similar consequences. My
aim in this paper is to explore the implications of this
dilemma both for research and for practical conservation.
The accumulating empirical data on the ecological
effects of habitat fragmentation hardly support a uniform conceptual framework, either. As an example, I
analyze a recent fragmentation article (Stratford and
Stouffer 1999). In the opening paragraph of the article,
the authors present a list of ‘‘effects of habitat fragmentation on avian communities . . . well documented
in various parts of the temperate zone.’’ This includes
the following: reduction in abundance and richness of
forest interior species; landscape-level effects such as
reduced dispersal among fragments, weakened philopatry, and changes in demography; reduced genetic
variation in isolated populations; fitness effects in populations in isolated fragments through increased nest
predation and brood parasitism, and reduced pairing
success; species exclusion by other species associated
with disturbance; and direct effects on conditions in
fragments, for instance, on the microclimate.
Stratford and Stouffer (1999) support their list with
12 references to specific, well-documented cases. The
list invites two comments. First, several of the cases
are connected indirectly with fragmentation; in other
words, the actual causal relationships are ambiguous.
Second, it is very unlikely that all of these effects could
be demonstrated simultaneously in any single fragmented landscape. In fact, some of the cases are manifestly specific to particular regions. For example, increased brood parasitism in agricultural forest remnants
by the Brown-headed Cowbird (Molothrus ater) is an
issue in the United States, but there is no analogous
species in similar landscapes in the Old World. Similarly, the Noisy Miner (Manorina melanocephala), an
Australian honeyeater, is an aggressive colonial breeder
that chases away other species from forest remnants in
which the colonies are located, but this species has no
analogy in Eurasia or North America. Effects like these
are contingent upon natural history on the regional or
continental scale.
Another type of contingency arises from historical
variation in the ways humans have changed habitat
configurations. The characterization by John T. Curtis
(1956) of the mid-20th century landscape in Green
County, Wisconsin, USA, cited in the epigram, fits
some agricultural landscapes elsewhere quite well, but
does not fit other landscapes at all. Does such variation
matter? How certain can we be that harmful ecological
effects observed in a particular human-modified landscape actually are due to changes in the spatial configuration of habitat instead of other types of ecological
disturbance brought about by agriculture? The effects
vary across regions, depending on ecological condi-
Ecological Applications
Vol. 12, No. 2
tions and the history of human-induced environmental
change.
These empirical problems really follow from the
conceptual ambiguity of the term. The problem is simple: as fragmentation is expected to give rise to multiple, even contradictory, processes depending on the
context, context-specific auxiliary factors are always
needed to build up a connection between fragmentation
and particular effects. Hence, the following dilemma
concerning the concept of habitat fragmentation: Is a
conceptually ambiguous and empirically multifaceted
term fruitful as a generic description of human effects
on landscapes?
I am not claiming that habitat fragmentation has no
consequence in human-modified environments. Rather,
I challenge the framing of the issue as always being of
one and the same kind. Fragmentation gives rise to
context-specific effects. The question is, how to integrate fragmentation effects into a general view of human-induced ecological change without lapsing into an
overtly typological conceptual framework?
As a similar contradiction between general rules and
local, context-specific peculiarities is potentially undermining all applications of conservation biology into
practice, the story of habitat fragmentation has considerable general interest for conservation thinking. To
set the stage for an analysis, I construct a genealogical
narrative of the stabilization of a schematic view of
fragmentation by the early 1980s, extracting the genealogy from a close reading of fragmentation literature, with a particular focus on critical background assumptions. I introduce the idea of genealogy in the next
section. In the third section, I condense the narrative
into five claims about how the schematic view became
constituted. This is followed by a summarizing section
in which I suggest a model of the dynamics of thinking
that supported the schematic view. The schematic view
has been challenged, and fragmentation research has
greatly diversified: since the mid-1980s, habitat fragments are increasingly viewed as elements in dynamic
landscapes. This shift opens space for understanding
historical contingency and contextual specificity in the
humanity–nature relationships, and directs research
and management to specific problems and solutions. In
conservation thinking, however, a schematic image of
fragments as isolated islands in a hostile ‘‘sea’’ of human-modified environments is also still alive. I elaborate the tension between schematic theory and context-specific practical needs in the last sections.
RESEARCH PROBLEM AND METHODOLOGY:
CONCEPTS, GENEALOGY, AND SCIENCE
The idea of habitat fragmentation was a conceptual
innovation adopted in ecology in the 1970s. Why did
this happen in the way that it did? My basic hypothesis
is that the idea came about as a response to a new
problem space that was opened by the ‘‘environmental
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
awakening’’ during the 1960s, that is, the amazingly
sudden rise into public consciousness that human activity may seriously damage the global environment.
New administrative bodies, in charge of environmental
policy, were established in industrialized countries during the first years of the 1970s (Weale 1992). The ‘‘environment’’ became a common denominator of a large
variety of specific problems, ultimately, a source of a
moral protest (Haila 1999b). It is not easy to explain
why the awakening happened the way it did (e.g., Hannigan 1995), but I take it simply as an historical fact.
The new problem space opened by the rise of environmentalism posed a particular challenge to ecological research. An essential aspect of the new environmentalism was a fear that vital ecological processes
would collapse on the global scale. Somebody was expected to specify what, precisely, such a collapse consists of. Ecologists originally took up the challenge
with some enthusiasm, but this was soon muted by the
sobering realization that basic ecology does not provide
recipes for specifying environmental problems, let
alone means of solving them. McIntosh (1985) describes the early responses; see Boucher (1998) for a
personal narrative.
I specify the connection between fragmentation research and the rise of environmentalism in more detail
in the next section. First, however, we have to ask how
conceptual innovations arise in science.
For science to maintain conceptual coherence, novel
ideas must be related to concepts and ways of thinking
adopted previously. ‘‘Genealogy’’ is a systematic method for analyzing conceptual change that unfolds within
a particular field of thinking. Genealogical analysis of
a particular research field aims at more than merely
documenting the historical development of the field.
The aim is to analyze the dynamics of thinking within
the field, positioned in a rich context of research practices and societal connections. New concepts bring
about new perspectives on recognizing research problems, new ways of addressing the problems, and new
criteria through which truth claims within the field are
assessed. These ideas are applied to ecology in Haila
(1998).
In a sense, an established concept turns into a ‘‘subject’’ that has the power to define research priorities,
resolve disputes, and justify conclusions. Thus, as we
shall see, the concept of habitat fragmentation was first
presented as a hypothetical suggestion as to specified
effects of human-induced change in natural landscapes.
However, quite quickly the concept came to stand for
a particular, island-biogeographic interpretation of species impoverishment in small habitat patches, and this
interpretation began to lead its own life, offering justification for particular research projects and conservation recommendations. Through this process, fragmentation came to appear as a unitary phenomenon.
As data for the genealogical analysis, I used articles
323
in the fragmentation literature that have been often cited; I introduce specific sources in context in the next
sections. The analysis drew upon two specific theoretical ideas. The first one is Ian Hacking’s (1992) suggestion that ‘‘stabilization’’ of research is a precondition for the establishment of a successful research
tradition. Such stabilization takes place through a selfvindicating structure, created by a circular process of
trial-and-success within the field. This comprises all
stages of research, from observations and experiments,
through analysis of data, to drawing conclusions and
matching them with existing theory, which may be
modified as a result. Every particular success in this
endeavor, at any stage, supports the credibility of the
whole tradition and triggers further research that again
supports the tradition, and so on. Generally accepted
exemplars such as successful experiments play an essential role here: they form a resource on which further
experimentation can draw.
Second, a precondition for the stabilization of research is explanatory closure. In the most elementary
case, explanatory closure is achieved within a ‘‘contrast
space’’ that defines alternative explanatory possibilities
relative to the phenomenon to be explained (Garfinkel
1981, Dyke 1988). Within a fixed contrast space, credible alternatives are articulated and juxtaposed against
each other: the exclusion of one supports the other(s).
The credibility of any explanation, however, depends
on how the corresponding contrast space is constructed.
STABILIZATION OF A SCHEMATIC VIEW
FRAGMENTATION
OF
The origin: theory of island biogeography
The origin of the idea of habitat fragmentation in
ecological theory is easily identified: it was first articulated in MacArthur and Wilson’s (1967) monograph
on island biogeography. However, the idea was mentioned only in one brief paragraph in the short introductory chapter of the monograph:
Insularity is moreover a universal feature of biogeography. Many of the principles graphically displayed in the Galapagos Islands and other remote
archipelagos apply in lesser or greater degree to all
natural habitats. Consider, for example, the insular
nature of streams, caves, gallery forests, tide pools,
taiga as it breaks up in tundra, and tundra as it breaks
up in taiga. The same principles apply, and will apply
to an accelerating extent in the future, to formerly
continuous natural habitats now being broken up by
the encroachment of civilization, a process graphically illustrated by Curtis’ maps of the changing
woodland of Wisconsin . . . .
It is really remarkable, with hindsight, that this paragraph is all there is on fragmentation in MacArthur
and Wilson, as the book was an essential reference in
fragmentation literature in the early stages. In the par-
324
INVITED FEATURE
agraph, MacArthur and Wilson offered a metaphoric
vision: increasing insularity is a dominant theme in
environments under human influence, and the Galapagos is an appropriate model for insular environments.
This may be the most beautiful example of a metaphoric meaning transfer that can be found in the ecological literature.
The irony is that, in fact, what is really ecologically
peculiar about Galapagos and other oceanic archipelagoes has no relevance to habitat fragments created by
human activity, such as the remnant forests in Wisconsin studied by Curtis (1956). The peculiarities of
the fauna and flora of oceanic archipelagos result from
their peculiar geological and evolutionary histories
that, furthermore, are extremely variable from case to
case. By contrast, forest remnants, as remnants, have
no evolutionary history at all. Moreover, ecological
processes in forest remnants are in various ways critically influenced by direct interactions with remnant
surroundings, whereas such interactions are negligible
on oceanic islands.
When fragmentation research started in the 1970s,
it was firmly tied to the metaphoric analogy between
fragments and islands (see also Wiens 1995a). As a
specific example, I investigate the ornithological research conducted by Robert Whitcomb and his coworkers in the eastern deciduous forests of the United
States in the 1970s (summarized in R. F. Whitcomb et
al. 1981). The starting point for the research project
was defined by Whitcomb (1977) as follows: ‘‘Since
1974 we have been examining oak–hickory woodlots
in the Maryland Piedmont and Coastal Plains from the
perspective of island biogeography.’’ The adoption of
the island analogy was justified as follows:
‘‘(i)ntuitively, one would imagine that at least some of
the principles regulating species richness on oceanic
islands would operate on these mainland ‘habitat islands’ as well.’’
The island-biogeographic framing of the research
task fixed a contrast space within which questions were
formulated: the focus was either on species number or
on the presence or absence of individual species in
fragments as a function of fragment area. This, however, is misleading because a small woodlot can sustain
only a small number of individuals, and thus, of species, and a majority of these belong necessarily to the
common species. Consequently, species assemblages
in small fragments always appear ‘‘impoverished’’
compared with large fragments. This result is inevitable, given an appropriate size distribution of fragments censused. For instance, the study design in Whitcomb et al. (1977) included 25 fragments with an area
range of 1–14 ha, and five fragments .70 ha used as
a reference. A 10-ha woodlot cannot possibly host the
same assemblage of species as a 70-ha one.
This inevitable species–area pattern came to dominate the imagination of the researchers: a small patch
Ecological Applications
Vol. 12, No. 2
of whatever kind of environment implies impoverishment and extinctions. The conclusion is not amenable
to empirical refutation because the pattern of apparent
impoverishment is always observed in any particular
fragmented landscape. Helliwell (1976) raised this
point as a critique of a contemporaneous British study,
but it has been all but neglected in the fragmentation
literature (Haila 1988, 1999a).
Intermediary conclusion.—The metaphoric application of island biogeography to ‘‘all natural habitats’’
created a vision that species impoverishment in humanmodified landscapes is directly comparable to the species impoverishment observed on distant oceanic islands; i.e., it is due to biogeographic processes. Recurring documentation of the expected pattern, reduced
species number and absence of scarce species in small
fragments, was accepted as evidence for such impoverishment, which helped in stabilizing the approach.
Pioneering works interpreted in new light:
John T. Curtis and Wisconsin woodlands
As their example of habitat changes in human-modified environments, MacArthur and Wilson (1967) referred to John T. Curtis’ article entitled ‘‘The modification of mid-latitude grasslands and forests by man,’’
published in a famous symposium volume Man’s Role
in Changing the Face of the Earth (Curtis 1956).
The context probably influenced the style of Curtis’
paper: it was a descriptive and somewhat speculative
natural history commentary on human modification of
mid-latitude habitats. He covered both the Old and the
New World nominally in his chapter, but for specific
examples, he drew mainly on his own data from Wisconsin. I summarize what Curtis wrote on fragmentation.
Curtis pointed out that, in agricultural landscapes,
original forest cover is often extensively destroyed so
that only small, remnant forests are left, surrounded by
modified environmental types (see the epigram). He
called this the ‘‘physiognomic result’’ of agricultural
occupation of forested land. The matrix between the
remnants was alien for practically all members of the
forest biota. On the other hand, many members of the
animal community greatly benefited from the increase
in the total amount of ‘‘edge.’’ Second, Curtis took up
biogeographic dynamics: as a consequence of changes
in habitat proportions, some species of open country
had extended their range ‘‘well into the original forest
country of the eastern United States.’’ Next, he noted
that species tend to vanish from small forest remnants,
and he identified a possible mechanism: local populations that die out as a consequence of natural disturbances and catastrophes might not be able to recolonize
from adjacent unaffected areas. Finally, Curtis suggested that the genetic composition of populations isolated in the remnants might change. He concluded by
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
predicting that such microevolution would provide a
fertile ground for future study.
To sum up, Curtis’ short article drew the conclusion
that human-induced modification of landscape structure helps species to spread to new regions, causes local
extinctions, and creates new selection regimes for species populations isolated in small woodlots. This was
a novel idea. Curtis did not have any theoretical framework to draw upon. Rather, he made specific suggestions on types of biotic change that the ‘‘physiognomic
result’’ of agricultural land clearance might bring
about. These were not restricted to impoverishment:
edge effect opens living space for a set of species, and
range shifts bring new species into new regions.
When Curtis’ work was incorporated into mainstream fragmentation research, his broad interest was
truncated and cast into a biogeographic mold. The statistics and scheme compiled by Curtis on forest change
in Cadiz Township, Green County, Wisconsin, USA,
has become an emblem of forest fragmentation. This
is ironic, because the most dramatic feature of the case
is forest loss: according to Curtis, forest cover by 1950
had been reduced to 3.6% of the 1831 situation. In fact,
the use of Curtis’ figure as an emblem of fragmentation
has made it difficult to distinguish habitat loss and
habitat fragmentation from each other.
Intermediary conclusion.—The use of Curtis’ work
as an emblem of habitat fragmentation, interpreted in
the light of the island-biogeographic theory, strengthened the inclination of ecologists to view human-modified environments as ‘‘archipelagoes’’ and to confuse
habitat fragmentation and habitat loss.
Fragments vs. surroundings: a total contrast
postulated
The island-biogeographic perspective on habitat
fragmentation was first applied to reserve design; this
created a series of influential articles from the mid1970s onward, as well as criticisms somewhat later (see
Simberloff 1997). The application was based on an
analogy between reserves and islands, an analogy that
was explicitly drawn by Frank Preston (1962), as was
duly acknowledged later.
A critical background assumption of the ‘‘reserves
equal islands’’ analogy is that areas surrounding insular
preserves are completely hostile to species living in
the preserves. I cite an example: ‘‘The fundamental
assumption is that nature reserves act as habitat islands
in an inhospitable sea of environment that has been
modified by man, and thus that the empirical findings
of island biogeography are pertinent’’ (Gilpin and Diamond 1980).
Evaluating this claim is, of course, an empirical matter, but it is untenable as a broad generalization. Even
in agricultural areas, however intensive the cultivation
methods, some species can forage outside forest remnants and thus include some of the surrounding area
325
in their home range. Curtis realized this in Wisconsin.
In forested environments, the situation is much more
complex; I will return to this issue. Big predators have
often been used as an example of species that supposedly require large, uniform, pristine habitats. For instance, Whitcomb et al. (1981:187) wrote: ‘‘A preserve
large enough to maintain a viable population of timber
wolves, for example, would be likely to preserve primary producers, herbivores, and insectivores as well.’’
It is more likely, however, that the big predators do not
require large areas of pristine habitat. Big predators are
excluded from human-modified environments by persecution. The recent recovery of the timber wolf (Canis
lupus) in the western Great Lakes region in the United
States is due to discontinuation of persecution, not to
any particular features of forest structure (Mladenoff
et al. 1995).
In fact, the provision of Gilpin and Diamond (1980)
can be read the other way around: whenever the preserves or fragments are not surrounded by an ‘‘inhospitable sea,’’ the ‘‘empirical findings of island biogeographic theory’’ are not pertinent. This conclusion has
seldom been drawn, however.
Intermediary conclusion.—Through the seductive
analogy between fragments and oceanic islands, a strict
contrast between fragments or preserves and the surrounding matrix was accepted as a fundamental assumption. Consequently, the mechanisms that supposedly make fragments ‘‘insular’’ were not scrutinized.
Assumed uniformity of unfragmented ‘‘reference’’
environment
Robert Whitcomb later characterized the background
of his ornithological work conducted in Maryland in
the 1970s as follows (R. F. Whitcomb 1987:165): ‘‘I
believe that the eastern deciduous forest was essentially
homogeneous from the point of view of forest interior
bird species.’’
This belief seems erroneous on two counts: (1) major
climatic gradients cut through the area, and (2) there
is ecologically significant variation in forest types on
various spatial scales, from local to regional, according
to topography, soil, disturbance history, and so on.
Birds often respond to both types of environmental
heterogeneity, as attested by their range variation and
habitat distributions.
The uniformity assumption was tacitly accepted in
early fragmentation studies: fragment area was regarded as an independent variable sufficient to explain statistically the variation in species numbers across a
range of fragments of different size. Ambuel and Temple (1983) conducted the first ornithological study in
which the effect of habitat variation was statistically
controlled. The uniformity assumption is also implicit
in how observations were extrapolated from small
patches to the regional scale. Consider the following
citation (Blake and Karr 1987:1724; references omit-
326
INVITED FEATURE
ted): ‘‘Many species do not occur in isolated woodlots
even when area of the woodlot exceeds minimum territory requirements and large forest tracts are necessary
for their survival.’’ The conclusion is logically invalid
because the absence of a particular species from particular small woodlots can be due to the lack of suitable
habitat in these particular woodlots. It can also be due
to a ‘‘sampling’’ effect: a forest bird with an average
density of, say, two pairs/km2 and minimum territory
requirements of 5 ha, not unreasonable figures at all,
would be absent from four out of five 10-ha woodlots,
on average, for statistical reasons alone.
As time and space are tied together in ecology (Preston 1960, Wiens 1981, Haila and Levins 1992), an assumption of spatial homogeneity implies temporal constancy. In the early fragmentation literature, the possibility that natural population fluctuations might influence the distribution of populations across
fragmented landscapes was often neglected. B. L.
Whitcomb et al. (1977) offer an example. They studied
changes in bird species composition in one woodlot
censused first in 1947 and a second time in 1975, i.e.,
with an interval of 28 yr. They noticed that, for instance, the Worm-eating Warbler (Helmitheros vermivorus) was present in 1947 with one pair, but absent
in 1975, and considered this an interesting case of
‘‘turnover,’’ again raising the analogy with oceanic islands: ‘‘(t)urnovers of this type might be compared with
those which occur on oceanic islands, which, for one
reason or another, carry more species than appropriate.’’ This interpretation is not legitimate, however: in
a period of 28 yr, pairs of breeding birds will shift their
locations many times over, e.g., from one woodlot to
the next. McCoy (1982) criticized too loose a notion
of ‘‘turnover’’ on similar grounds. In fact, birds are
known to shift their territory locations from year to
year at the population level (Preston 1960, Haila 1983,
Haila and Hanski 1987, Haila et al. 1993, 1996).
That spatial and temporal variability matters in ecology was, of course, common knowledge in the 1970s;
hence, their neglect seems strange with hindsight. C.
B. Williams (1943, 1964) and Frank Preston (1948,
1960, 1962), in particular, had demonstrated with databased quantitative studies the influence of environmental heterogeneity on species richness and composition in local multispecies collections. Preston drew
an explicit analogy between temporal and spatial variation in a sampling process, and presented small patches of land metaphorically as ‘‘drawing samples’’ from
a surrounding universe. This extensive background
knowledge was, however, pushed aside.
The uniformity assumption provides another example of the significance of the contrast space within
which explanatory closure is reached. A contrast space
focused on species number as a function of fragment
area does not allow the articulation of spatial and temporal variability as significant factors. By contrast,
Ecological Applications
Vol. 12, No. 2
Preston’s (1948, 1960) idea that local assemblages
should be considered ‘‘samples’’ drawn from a surrounding ‘‘universe’’ offers another contrast space in
which both spatial and temporal variability in the ‘‘universe’’ become focal questions (Connor and McCoy
1979, Haila 1983).
Intermediary conclusion.—Fragmented environments were projected against an idealized, spatially
homogeneous and temporally constant natural background. As a consequence, well-established ecological
knowledge on how environmental variation affects the
composition of multispecies collections was ignored.
Fragmentation and the rise of environmentalism
I previously raised the hypothesis that the environmental awakening created a novel problem space for
ecological research, which may explain the fascination
of ecologists with habitat fragmentation in the 1970s.
A number of ecological research traditions began to
explore this problem space, and biogeographically inspired fragmentation research was one among them.
The analogy of preserves and oceanic islands was
thought to be fruitful.
I will not review the criticism raised against the analogy (see, e.g., Simberloff 1997), but will pay attention
to a conceptual issue. Jared Diamond (1976) supported
the island analogy with the following argument: ‘‘Recently several workers have recognized that a predictive understanding of extinction might be obtained
from island biogeography, since refuges of natural habitat in a sea of human-altered environment behave as
islands for species dependent on natural habitat.’’
How should one understand the term ‘‘predictive understanding’’ in this argument? There are two alternatives. Successful predictions can be based either on
a statistical regularity (reliable correlation) or on a correspondence with a causal relationship. The understanding offered by island biogeography is of the correlation type: area does not cause extinctions, although
area does, by and large, correlate negatively with the
risk of local population extinction. A critical question
is whether this correlation is reliable enough to really
allow predictions. Diamond’s argument is explicitly
based on the ‘‘refuges equal islands’’ analogy. This is
not allowable without empirical support, as I have already noted. Furthermore, there are both theoretical and
empirical grounds to believe that critical ecological
processes in isolated vs. non-isolated situations are different (Haila 1990); if this is the case, the analogy
between islands and reserves fails (Saunders et al.
1991). The causes of local population extinctions vary
across circumstances; hence, the ecology of the situation has to be known in each case.
Nevertheless, firm conclusions on conservation were
drawn from research on small fragments from the very
beginning. In fact, Robert Whitcomb and his colleagues
presented conclusions from their ornithological work
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
327
promise was a rhetorical strategy that helped to stabilize fragmentation research in the short run.
Intermediary conclusion.—The environmental awakening created a social demand for ecologically grounded theoretical perspectives on ecological problems.
Fragmentation research was thought to provide insight
on human-induced extinctions through the analogy between species impoverishment on islands and habitat
fragments. This was supported with a rhetorical strategy of disciplinary promise.
SCHEMATIC VIEW OF FRAGMENTATION
INTELLECTUAL ATTRACTOR
FIG. 1. A conceptual phase space that constrains the possible trajectories of fragmentation research; the axes stand
for issues on which background assumptions are made behind
any particular research project. The attractor shows where the
dominant assumptions of the schematic view of fragmentation
meet.
for the first time in an exchange on the application of
island biogeography to conservation in Science (R. F.
Whitcomb et al. 1976); the exchange was started by
Simberloff and Abele (1976). The central Maryland–
Washington, D.C. database that Whitcomb et al. drew
upon was not described in the article, but a footnote
indicated that it was developed by many workers over
three decades and included: ‘‘155 breeding bird censuses of 49 forest plots, 93 point surveys in deciduous
forest fragments, three countywide bird atlas mapping
projects, and 111 breeding bird surveys of 55 routes.’’
From this database, the authors drew the following conclusion:
Thus the available data for the Eastern forest, far
from demonstrating that large faunal preserves are
unnecessary, describe a troubled system in which
local and regional extinctions of forest interior specialists are commonplace and in which large series
of existing small, isolated forest areas have failed to
preserve, even in contemporary time, many of the
small avian species that once dominated the forest.
It is quite remarkable that this firm, purportedly databacked position was printed in no less a forum than
Science before any of the primary data had actually
been analyzed and published. The data came out four
years later in Whitcomb et al. (1981). I think that this
incident provides an example of a ‘‘disciplinary promise’’ (Haila 1998), i.e., a promise that island-biogeographic conservation theory helps in addressing and
solving conservation problems. Giving a disciplinary
AS AN
A schematic view of habitat fragmentation was stabilized by the early 1980s. Its main elements were (1)
the island analogy; (2) an assumed stark contrast between fragments and their surroundings; and (3) the
assumed uniformity of natural background conditions.
An essential part of the stabilizing loop was the reduction of empirical research to species–area curve fitting, which always produced expected results. The perceived need to provide scientific advice on conservation issues provided the ethos.
I next give a dynamic interpretation of the genealogical narrative. This is based on an analogy between
the contrast space in which research questions are
framed and a physical phase space. As Alan Garfinkel
(1981:40) wrote, defending such an analogy, a phase
space is ‘‘a geometric representation of the possibilities
of a system.’’ Accordingly, one can envisage ideas in
a particular research field ‘‘moving’’ in a phase space
that gives a representation of possible trajectories; my
use of the analogy was originally inspired by Dyke
(1993).
How is a phase space of ideas constituted? Intellectual trajectories are constrained by critical background
assumptions; that background assumptions matter is a
familiar idea by now, applied to ecology by Haila and
Levins (1992). Hence, a phase space of ideas can be
defined by identifying the issues on which critical background assumptions take a stand, and by defining main
axes by these issues. Such axes, of course, are not
linearly scaled; rather, possible positions along the axes
are often separated into mutually opposing alternatives.
In other words, each axis stands for a contrast space
within which ideas are located.
I identified critical background assumptions of fragmentation research from the genealogical narrative.
They are positioned as the axes of a two-dimensional
phase space in Fig. 1. The first axis is constituted by
assumptions on the relationship between human-influenced areas and natural areas. The contrast space behind this axis is between a dualistic opposition vs. context-specific interaction and interpenetration. The second axis is constituted by assumptions on the nature
of spatiotemporal variability in ecological space. The
contrast space is between a uniform, Cartesian equi-
328
INVITED FEATURE
librium space vs. a complex and heterogeneous nonequilibrium space.
The assumptions tacitly accepted about these two
issues in the early fragmentation literature can be represented as follows. Along the first axis, fragmentation
research was constrained by an assumed total contrast
between human-influenced areas and natural areas.
This was articulated by Gilpin and Diamond (1980) as
cited previously. Along the second axis, fragmentation
research was constrained by an assumption of uniform
and unchanging natural background. This was articulated by Robert Whitcomb and his colleagues (B. L.
Whitcomb et al. 1977, R. F. Whitcomb 1987) as cited
previously.
These background assumptions meet in the upper
right-hand corner of Fig. 1. This region constitutes an
‘‘intellectual attractor’’ in this phase space: the trajectory of thinking on fragmentation is directed toward
this attractor, given the background assumptions. An
intellectual attractor is effective because it seems difficult, even preposterous, to think in any other way,
i.e., to challenge commonly accepted background assumptions.
Both of the issues that I used to delineate the phase
space in Fig. 1 have been relevant in much of modern
ecological thinking. The first, the contrast between human-influenced and natural areas, accords with a dualistic view of the relationship between culture and
nature. Culture–nature dualism has ancient roots in the
Western tradition and it still has considerable influence
in environmentalism (Passmore 1980, Plumwood 1993,
Haila 1999b, 2000). The second issue draws upon an
equilibrium view of ecology (for analysis and criticism,
see Wiens 1984, 1989a). The equilibrium view implies
that every habitat comprises a uniform ecological space
that is divided among competing species, and that different habitats are distinctly recognizable. In island
biogeography, this view of habitat was implied through
the concept of ‘‘the closed community’’ (MacArthur
and Wilson 1967:94–104).
Environmental concerns make understandable the
dynamics depicted in Fig. 1. The environmental crisis
created a demand for ecologists to give scientific advice
on conservation issues. The advice had to be backed
by theories that conform to an ideal of modern ecology
as an exact and predictive science, based on general
principles. Kingsland (1985) traced the intellectual history of this ideal from the early 20th century through
the influence of G. E. Hutchinson up to MacArthur and
Wilson and Island Biogeography.
Daniel Simberloff (1997:275) characterized recently
this urge as follows: ‘‘Until the mid-1970s, conservation biology was not a distinct science. Rather, it was
part of ecology and wildlife biology and largely consisted of research on the autecology and habitats of
species of particular concern. In 1974–1975, the application of the dynamic equilibrium theory of island
Ecological Applications
Vol. 12, No. 2
biogeography to the design of nature preserves helped
inaugurate a new conservation biology as well as a new
kind of biogeography.’’ Simberloff’s characterization
has paradoxical implications—a misleading metaphor
inaugurating conservation biology as ‘‘a distinct science’’—but it certainly supports my hypothesis on the
importance of conservation issues.
THE SCHEMATIC VIEW CHALLENGED
Fragmentation research has changed enormously
since the early 1980s, and the questions asked have
become much more variable than they were in the early
stages. In this section, I first summarize results of a
computer search from Biological Abstracts of 1998 using the key words ‘‘fragmentation and habitat’’ and
‘‘fragmentation and forest.’’ The aim of the search was
to identify the range of study organisms and specified
problems covered in recent fragmentation research. The
search gave 93 and 75 relevant titles, respectively; 47
of the articles were on both lists, so the total was 121.
The number of articles dealing with a single species
vs. more than one species was almost equal, 62 and
57, respectively; in addition, two addressed landscape
structures. The broad range of study objects is summarized as follows:
1) Taxonomically specified study organisms.—(A)
Animals (total 71): mammals, 18; birds, 22; reptiles,
3; amphibians, 4; fishes, 4; insects, 14; crustaceans, 3;
mollusks, 1; Monotremata (Platypus), 1. (B) Plants (total 24): flowering plants, 6; trees, 10; seaweeds, 1; algae, 1; unspecified, 6.
2) Taxonomically heterogeneous study organisms.—
Various collections of species, 12; functionally defined
groups of species, 4; species interactions, 2.
3) Models studied.—Eight models.
Research problems are more difficult to classify than
study objects, but the following figures indicate approximate distributions of the types of problems addressed in the papers, as interpreted from their abstracts: population viability 47; species richness/impoverishment 24; ecosystem effects 23; specified processes such as dispersal, foraging, movements, etc. 11;
descriptive studies 9. To summarize, a majority of the
fragmentation studies conducted in the late 1990s addressed ecological processes influencing population viability, but descriptive research on ecosystem and community characteristics was almost equally popular. Taxonomically, the studies covered a broad range of organisms, but vertebrates were the dominant group.
McGarigal and Cushman (2002) drew similar conclusions in their review of experimental approaches to
fragmentation in the late 1990s: vertebrates, particularly songbirds, dominated as study organisms, and
most studies were focused on patches instead of landscapes.
In what follows, I give a condensed interpretation of
how the schematic view got to be challenged. This had
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
both theoretical and empirical sources. The theoretical
criticism had several roots that essentially add up to
challenging the assumption depicted on the y-axis in
Fig. 1. An important root was the increasing criticism
and, finally, almost unanimous rejection of equilibrium
assumptions in ecology (see Wiens 1984, 1989a). This
went together with a suspicion and rejection of the
concept of a ‘‘closed community’’ maintained by interspecific competition (which, as we saw, was an explicit element in the theory of island biogeography).
The autecology of individual species was emphasized
instead. This emphasis found support from an older,
individualistic view of ecological assemblages that has
its roots extending to the early 20th century (McIntosh
1985). The rejection of equilibrium community ecology undermined the relevance of species number as a
characteristic of species assemblages in islands and
habitat fragments, i.e., the island-biogeographic explanatory closure was put into doubt.
Once the equilibrium assumptions were relaxed, new
questions were opened for empirical investigation, e.g.,
whether the composition and diversity of local assemblages are regulated by local or regional processes
(Ricklefs 1987). This also had implications for fragmentation studies: if regional processes are important,
then what is observed locally may be an epiphenomenon of regional abundance variations. Issues of ecological scaling were brought to the foreground in ecological thinking (Wiens 1981, Delcourt et al. 1983). As
Wiens (1989b) noted, ‘‘scaling’’ became almost a buzzword by the late 1980s. The problem of ecological
scaling has two aspects: methodological, the design of
sampling relative to the processes that are studied, and
ontological, the organization of different ecological
processes in space and time vis-à-vis each other. We
can never be certain about particular ontological assumptions, but some assumptions are more plausible
than others in the light of particular theoretical commitments and empirical results. Scaling problems have
direct implications for the ecology of fragmented environments: different processes may be important depending on the scale of the fragments studied (Haila
1990, Lord and Norton 1990).
Landscape ecology originated as a new research program in the 1980s, in a way as an amalgamation of all
these theoretical developments (Wiens 1995 b): a uniform Cartesian space was replaced with environmental
heterogeneity; equilibrium assumptions were replaced
with spatial and temporal variability driven by nonequilibrium processes in various scales; and an unquestioned search for universal regularities was replaced with the acknowledgment of spatial and temporal specificity of ecological processes. New hypotheses were suggested about non-equilibrium processes
determining landscape dynamics (Shugart 1998). Many
of the basic ideas originated in previous decades, but
these were put together in a novel way. For fragmen-
329
tation research, particularly important emphasis was
that relevant features of landscapes must be defined
from an organismal perspective, in other words, the
same physical landscape may be experienced very differently by different organisms (Haila 1991). Fragments were increasingly viewed not as analogues to
‘‘islands,’’ but rather as habitat patches embedded in
a particular type of surroundings (Wiens 1995a).
Another challenge to the island-biogeographic view
of fragmentation arose from experience accumulating
in field research showing that the island-biogeographic
approach did not allow the right questions to be asked
about fragmentation effects. Effectively, the ‘‘area vs.
species number’’ contrast space of framing questions
faded into the background, and the focus shifted to
population-level studies. This occurred in particular research programs as a time-bound transition: problematic species were identified by conducting quantitative
surveys across a particular fragmented environment,
and in the next stage these species were studied in detail
(Haila and Hanski 1984). A good example of such a
transition is Raimo Virkkala’s work on the avifauna in
managed vs. old-growth forests in the Finnish Lapland,
which started with surveys (Virkkala 1987) and then
focused on a specialist of old growth, the Siberian Tit,
Parus cinctus (Virkkala 1990).
Focused studies on ecological processes in fragmented landscapes have produced a great variety of
results, as pointed out in the Introduction (see also
Haila 1999a, Harrison and Bruna 1999). The perspective on landscape structure has inspired interesting
modeling work (Andrén 1996, Fahrig 1998), but it is
a difficult question to study empirically because of logistic problems. Research on the dynamics of single
species on a landscape scale borders on metapopulation
studies (Hanski and Gilpin 1997).
Similar theoretical developments inspired a ‘‘new
paradigm’’ in conservation biology (Pickett et al.
1992). The determinants of population viability in heterogeneous and variable, non-equilibrium environments have become an important research focus. However, the island-biogeographic perspective is the dominant approach to evaluating extinction risks in conservation contexts; recent examples are several articles
in Whitmore and Sayer (1992), Wilson (1992), and May
et al. (1995). This is paradoxical because detailed reviews of the empirical relevance of the relationship
tend to conclude that it is vacuous; the arguments to
support this conclusion are summarized in Heywood
and Stuart (1992) and Whitmore (1997). I believe that
this paradox reflects a genuine contradiction in conservation thinking. On the one hand, the relevance of
the species–area relationship depends on the validity
of the ‘‘refuges equal islands’’ analogy, but this analogy is highly questionable. Hence, doubts arise whenever it is accurately assessed. On the other hand, conservation biologists feel an urge to use the relationship
330
Ecological Applications
Vol. 12, No. 2
INVITED FEATURE
PLATE 1. A typical example of fragmentation in managed boreal forest in southern Finland. The edge on the side of a recent cutover
is sharp, but the fragment merges together with
a nearby continuous forest along a depression
by a small creek (toward the right in the picture). The degree of isolation of such fragments
is, of course, an empirical question. Photograph
by Yrjö Haila.
in a predictive fashion as a first approximation. This
was indicated in my citation in this paper of Jared
Diamond; another justification is provided by Wilson
(1992:263): ‘‘Models of this kind are used routinely in
science when direct measurements cannot be made.
They yield first approximations that can be improved
stepwise as better models are devised and more data
added.’’
The only possible solution to this paradox is to assess
seriously and realistically whether the area surrounding
particular habitat fragments is, indeed, an ‘‘ecological
desert,’’ but this task has seldom been undertaken. A
sharp distinction is certainly a good first guess in many
agricultural regions (see Plate 1). By contrast, a detailed assessment is always needed in forested environments. In southern Finnish boreal forests, for instance, the percentage of species restricted to mature
forests is ;5–10% in birds and ground arthropods (Haila et al. 1994). Whether a majority of the species in
these taxa can persist in modified forests depends on
fine-scaled details of forest structure (Haila et al. 1989,
1996, Niemelä et al. 1996). The percentage of obligatory old-growth species is certainly higher in other
taxa, such as insects living in decaying wood (Väisänen
et al. 1993), but this statement opens up a new problem:
do these insects require old growth ‘‘per se,’’ or rather
some microhabitat usually available in old growth? The
second alternative may be true: several specialized beetles seem to persist in clearcuts if the right kind of dead
wood is left behind (Kaila et al. 1997).
I think that the justification for the use of the species–
area relationship given by E. O. Wilson in the previous
citation is not valid. The logic of the model is misleading in a concrete way: it directs attention to a contrast space that does not allow one to ask the specific
questions about extinction risks that need to be asked.
‘‘Area’’ is a poor indicator of the specific factors that
cause extinction threats of particular species. In the
southern Finnish taiga, for instance, it is not possible
to define such criteria for ‘‘adequate habitat’’ that could
be easily bounded on specific spots on the map and
would be valid for any large collection of species. Specific criteria are valid for only a few species at a time,
and ‘‘habitats’’ merge into one another in many more
than two dimensions (Haila et al. 1996). ‘‘Valuable
habitat’’ does not stand out as separate from a surrounding inhospitable ‘‘sea’’ of human-modified forests. On the other hand, this does not mean that we
have to fall back to purely autecological, species-byspecies consideration. Generalizations are possible as
to what kind of species are vulnerable in a particular
type of environment (Haila 1999a). Furthermore, such
vulnerabilities can be identified through empirical surveys (for examples, see Margules and Austin 1995).
BROADENING
FRAMEWORK: FRAGMENTATION
HABITAT DEGRADATION
THE
AS
To return to Fig. 1, the theoretical criticism voiced
against the ideal of a uniform, Cartesian equilibrium
world has pushed fragmentation research downward
along the y-axis. However, my interpretation is that the
region in the lower right-hand side of the figure does
not have stable equilibria. If a total contrast between
human-modified and natural environments is assumed,
equilibrium thinking is smuggled in through the back
door, as an idealization of a harmonious, pre-human,
natural environment (Haila 1999b).
To get away from the basin of the dominant attractor,
we need to move toward the left along the x-axis, i.e.,
to challenge the assumed universal contrast between
human-modified and natural environments. This implies contextualization of environmental problems
(Haila 1997, 1999b). To achieve this aim, I suggest that
the perspective be broadened and fragmentation be regarded as a specific form of habitat degradation (for a
similar suggestion, see Harrison and Bruna 1999).
Then, habitat fragmentation merges together with other
types of harmful changes induced by human activity.
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
This raises contextualized questions and also brings
human actors into the picture in efforts to prevent or
mitigate the harmful effects.
The urgency of this move depends on what kind of
precision is thought to be necessary when conservation
biologists characterize human-induced destruction of
natural habitats. I believe that great precision is needed
on the type of destruction in each particular case. Environmental problems can only be solved concretely,
in specific places at specific times, and people have to
be involved: managers, landowners, or local inhabitants, depending on the case. Hence, when conservationists address a particular problem, they should know
the type of remedy that might work. When conservationists grope in the right direction, a process of social
learning can start and strengthen the commitment of
the people involved. If, instead, conservationists resort
to an abstract model that says nothing specific about
the situation at hand, such commitment may fade. Furthermore, too sweeping generalizations about humancaused destruction of nature support a metaphysical
humanity–nature dualism that is counterproductive in
environmental thinking (Haila 1999b, 2000). The issue
also has an ethical dimension. Conservation recommendations are often intrusive upon ordinary people’s
lives and subsistence; consequently, the recommendations must be well founded and specific in order to
gain—or deserve—popular support.
The framing of questions should be sensitive to the
historical context: the particular type of human-induced
habitat degradation is variable from case to case (e.g.,
Landa et al. 1997, McIntyre and Hobbs 1999). Conservation biology seeks solutions to problems caused
by human activity, and it simply cannot be assumed at
the outset that different types of problems can be separated into well-defined ‘‘scientific kinds’’ (Hacking
1991) that would allow unambiguous explanatory closures. This is an empirical matter, not to be decided ex
cathedra. To address the issue, ecologists should take
human-modified environments seriously, as the rich
complexes of human-induced and naturally occurring
processes that they really are (Haila 1999b). This also
means that assessing the variability of fragmentation
effects in any particular situation is a research problem
in its own right. Description of particular cases through
long-term observational or experimental work is required; several examples are reviewed in the other articles in this Invited Feature. Other good examples include the long-term experiment in Wog Wog, New
South Wales, by Chris Margules and coworkers (Margules 1992), and the detailed documentation of distribution patterns of a wide range of taxa in the western
Australian Wheatbelt by Richard Hobbs, Denis Saunders, and others (Hobbs and Saunders 1993).
A good research strategy is to use specific, wellunderstood cases as ‘‘analog models’’ to identify important qualitative features that may have more general
331
relevance. By the term, analog model, I refer to a material system that is known in detail and that thus constitutes a real-life model of an abstract idea. A particular analog model is a first test of whether a particular
abstract notion is realistic at all. For instance, a real
pendulum is an analog model of a harmonic oscillator.
In ecological research, experimental model systems
function as analog models in this sense, both in the
laboratory (Griesemer and Wade 1988) and in the field
(Ims et al. 1993). Analog models cannot be directly
‘‘fitted’’ from one situation to another; rather, they suggest qualitatively important aspects and processes that
deserve particular attention in studying the other situation. The relationship between a pendulum and a
mass-and-springs oscillator, for instance, is similar:
they are both models of a harmonic oscillator, but
knowing the mechanical principle of a pendulum does
not help a bit if one has to construct a mass-and-springs
oscillator.
Sound reason can and should be used whenever habitat fragmentation has to be addressed as a practical
conservation issue. In the following, I list some obvious rules of thumb (a somewhat similar list is in Haila
1999a): (1) Avoid clearing native habitats, particularly
in regions where this has already happened to a large
degree. (2) Restore native vegetation by active measures such as planting, particularly in heavily cleared
country such as western Australia. This helps the survival of the native biota and is also economically sound
by rescuing some of the degraded lands back to agriculture (Saunders and Briggs 2002). (3) Cherish corridors and other connecting habitat routes in heavily
cleared landscapes. Corridors are generally valuable in
agricultural landscapes (Saunders et al. 1991), although
statistically rigorous experimental testing of their significance may be logistically impossible. In forested
environments, the role of corridors is more questionable. (4) Maintain environmental heterogeneity both on
the landscape level and within landscape elements such
as habitat patches. (5) Identify particularly important
microhabitats (‘‘key biotopes’’) and preserve these
both within and outside preserves. (6) Use umbrella
species to assess the quality of the landscape: if they
can persist, some other organisms can persist as well
(Martikainen et al. 1998).
An important practical question is to get both managers in various practical fields and ordinary people
involved. Conservation has to be integrated into human
lived experience and daily practices. This is only possible if both the conservationists and the public understand what the specific issues are in each specific
situation.
AN AFTERTHOUGHT: LEGACY OF ISLANDBIOGEOGRAPHIC FRAGMENTATION RESEARCH
In hindsight, the island-biogeographic view of habitat fragmentation that took shape some two decades
INVITED FEATURE
332
ago can be criticized on several grounds. However, this
does not entail a judgment that the view had no positive
role in the development of conservation biology. I
agree with Simberloff (1997; see the previous citation)
that the framework paved the way for new approaches
in conservation research. Haila and Järvinen (1982)
made a similar assessment of the role of island biogeography in ecological theory. In particular, fragmentation research inspired by island biogeography represented an early articulation of the spatial dimension
of human-induced, large-scale habitat degradation.
On the other hand, one should not be led to think
that this episode in the development of ecological
thinking was inevitable. Indeed, pre-island-biogeographic literature includes interesting studies of nonisolated real islands that are, after all, much more similar to habitat fragments that oceanic islands (Haila
1990). These present us with alternative trajectories
that were available for research. For instance, in his
inventory of the bird faunas of small British islands,
David Lack (1942) specified very carefully both human-induced and natural factors, including habitat
changes, that had affected the fluctuations in bird species composition and population numbers on different
islands. Another example is the work of Edward Beals
(1960) on the bird communities in the Apostle Islands
of Lake Superior, Wisconsin, USA. Beals was primarily
interested in the distribution of individual bird species
across different types of forest found on different islands. His background assumption was that species are
distributed across the total habitat space according to
their habitat needs in an individualistic fashion. According to Beals, island area affected birds primarily
indirectly, through the positive correlation between
area and habitat diversity.
Conservation research based on natural history also
reached important results in the era before island biogeography, but this is too large a topic to be covered
here. Plant ecologists, in general, have been less susceptible than animal ecologists to grand and unifying
explanatory schemes, as Mike Austin argues (e.g., Austin 1999). Such studies could have opened the way for
a different style of research on ecological isolates and
on conservation issues.
The history of research is readily ‘‘Whiggish’’: alternatives that did not take off are buried and forgotten,
and only the lineage of the mainstream remains. In the
case of fragmentation research, however, pre-islandbiogeographic alternative perspectives are still identifiable. They can be picked up and carried further.
ACKNOWLEDGMENTS
The genealogical study was made possible by a sabbatical
year financed by the Finnish Academy. Mirkka Jones helped
with the computer search on Biological Abstracts. Marc-André Villard and two anonymous referees provided fruitful
comments.
Ecological Applications
Vol. 12, No. 2
LITERATURE CITED
Ambuel, B., and S. A. Temple. 1983. Area-dependent changes in the bird communities and vegetation of southern Wisconsin forests. Ecology 64:1057–1068.
Andrén, H. 1996. Population responses to habitat fragmentation: statistical power and the random sample hypothesis.
Oikos 76:235–242.
Austin, M. P. 1999. The potential contribution of vegetation
ecology to biodiversity resarch. Ecography 22:465–484.
Beals, E. 1960. Forest bird communities in the Apostle Islands of Wisconsin. Wilson Bulletin 72:156–181.
Blake, J. G., and J. R. Karr. 1987. Breeding birds of isolated
woodlots: area and habitat relationships. Ecology 68:1724–
1734.
Boucher, D. 1998. Newtonian ecology and beyond. Science
as Culture 7:493–517.
Connor, E. F., and E. D. McCoy. 1979. The statistics and
biology of the species–area relationship. American Naturalist 113:791–833.
Curtis, J. T. 1956. The modification of mid-latitude grasslands and forests by man. Pages 721–736 in W. L. J. Thomas, editor. Man’s role in changing the face of the earth.
University of Chicago Press, Chicago, Illinois, USA.
Delcourt, R. H., P. A. Delcourt, and T. Webb. 1983. Dynamic
plant ecology: the spectrum of vegetational change in space
and time. Quaternary Science Reviews 1:153–175.
Diamond, J. M. 1976. Island biogeography and conservation:
strategy and limitations. Science 193:1027–1029.
Dyke, C. 1988. The evolutionary dynamics of complex systems. A study in biosocial complexity. Oxford University
Press, Oxford, UK.
Dyke, C. 1993. Extralogical excavations. Philosophy in the
age of shovelry. Pages 102–126 in J. Caputo and M. Yount,
editors. Foucault and the critique of institutions. Pennsylvania State University Press, Philadelphia, Pennsylvania,
USA.
Fahrig, L. 1998. When does fragmentation of breeding habitat affect population survival? Ecological Modelling 105:
273–292.
Garfinkel, A. 1981. Forms of explanation. Rethinking the
questions in social theory. Yale University Press, New Haven, Connecticut, USA.
Gilpin, M. E., and J. M. Diamond. 1980. Subdivision of
nature reserves and the maintenance of species diversity.
Nature, London 285:267–268.
Griesemer, J. R., and M. J. Wade. 1988. Laboratory models,
causal explanation, and group selection. Biology and Philosophy 3:67–96.
Hacking, I. 1991. A tradition of natural kinds. Philosophical
Studies 61:109–126.
Hacking, I. 1992. The self-vindication of the laboratory sciences. Pages 29–64 in A. Pickering, editor. Science as practice and culture. University of Chicago Press, Chicago,
Illinois, USA.
Haila, Y. 1983. Land birds on northern islands: a sampling
metaphor for insular colonization. Oikos 41:334–351.
Haila, Y. 1986. North European land birds in forest fragments: evidence for area effects? Pages 315–319 in J. Verner, M. L. Morrison, and C. J. Ralph, editors. Wildlife
habitat 2000: modeling habitat relationships of terrestrial
vertebrates. University of Wisconsin Press, Madison, Wisconsin, USA.
Haila, Y. 1988. Calculating and miscalculating density: the
role of habitat geometry. Ornis Scandinavica 19:88–92.
Haila, Y. 1990. Toward an ecological definition of an island:
a northwest European perspective. Journal of Biogeography
17:561–568.
Haila, Y. 1991. Implications of landscape heterogeneity for
April 2002
HABITAT FRAGMENTATION: A REASSESSMENT
bird conservation. International Ornithological Congress
20:2286–2291.
Haila, Y. 1997. ‘Wilderness’ and the multiple layers of environmental thought. Environment and History 3:129–147.
Haila, Y. 1998. Political undercurrents of modern ecology.
Science as Culture 7:465–491.
Haila, Y. 1999a. Islands and fragments. Pages 234–264 in
M. L. J. Hunter, editor. Maintaining biodiversity in forest
ecosystems. Cambridge University Press, Cambridge, UK.
Haila, Y. 1999b. Socioecologies. Ecography 22:337–348.
Haila, Y. 2000. Beyond the nature–culture dualism. Biology
and Philosophy 15:155–175.
Haila, Y., and I. K. Hanski. 1984. Methodology for studying
the effect of habitat fragmentation on land birds. Annales
Zoologici Fennici 21:393–397.
Haila, Y., and I. K. Hanski. 1987. Habitat and territory overlap of breeding passerines in the mosaic environment of
small, northern islands. Ornis Fennica 64:37–49.
Haila, Y., I. K. Hanski, J. Niemelä, P. Punttila, S. Raivio, and
H. Tukia. 1994. Forestry and the boreal fauna: matching
management with natural forest dynamics. Annales Zoologici Fennici 31:187–202.
Haila, Y., I. K. Hanski, and S. Raivio. 1989. Methodology
for studying minimum habitat requirements of forest birds.
Annales Zoologici Fennici 26:173–180.
Haila, Y., I. K. Hanski, and S. Raivio. 1993. Turnover of
breeding birds in small forest fragments: the ‘‘sampling’’
colonization hypothesis corroborated. Ecology 74:714–
725.
Haila, Y., and O. Järvinen. 1982. The role of theoretical
concepts in understanding the ecological theatre: a case
study on island biogeography. Pages 261–278 in E. Saarinen, editor. Conceptual issues in ecology. D. Reidel, Dordrecht, The Netherlands.
Haila, Y., and R. Levins. 1992. Humanity and nature. Ecology, science and politics. Pluto Press, London, UK.
Haila, Y., A. O. Nicholls, I. K. Hanski, and S. Raivio. 1996.
Stochasticity in bird habitat selection: year-to-year changes
in territory locations in a boreal forest bird assemblage.
Oikos 76:536–552.
Hannigan, J. A. 1995. Environmental sociology. A social
constructionist perspective. Routledge, London, UK.
Hanski, I. A., and M. E. Gilpin, editors. 1997. Metapopulation biology. Ecology, genetics, and evolution. Academic
Press, New York, New York, USA.
Harrison, S., and E. Bruna. 1999. Habitat fragmentation and
large-scale conservation: what do we know for sure? Ecography 22:225–232.
Helliwell, D. R. 1976. The effects of size and isolation on
the conservation value of woodland sites in Britain. Journal
of Biogeography 3:407–416.
Heywood, V. H., and S. N. Stuart. 1992. Species extinctions
in tropical forests. Pages 91–117 in T. C. Whitmore and J.
A. Sayer, editors. Tropical deforestation and species extinction. Chapman and Hall, London, UK.
Hobbs, R. J., and D. A. Saunders, editors. 1993. Reintegrating fragmented landscapes. Towards sustainable production and nature conservation. Springer-Verlag, New
York, New York, USA.
Ims, R. A., J. Rolstad, and P. Wegge. 1993. Predicting space
use responses to habitat fragmentation: can voles Microtus
oeconomus serve as an experimental model system (EMS)
for capercaillie grouse in boreal forest? Biological Conservation 63:261–268.
Kaila, L., P. Martikainen, and P. Punttila. 1997. Dead trees
left in clear-cuts benefit saproxylic Coleoptera adapted to
natural disturbances in boreal forest. Biodiversity and Conservation 6:1–18.
Kingsland, S. E. 1985. Modeling nature. Episodes in the
333
history of population ecology. Princeton University Press,
Princeton, New Jersey, USA.
Lack, D. 1942. Ecological features of the bird faunas of British small islands. Journal of Animal Ecology 11:9–36.
Landa, R., J. Meave, and J. Carabias. 1997. Environmental
deterioration in rural Mexico: an examination of the concept. Ecological Applications 7:316–329.
Lord, J. M., and D. A. Norton. 1990. Scale and the spatial
concept of fragmentation. Conservation Biology 4:197–
202.
Lubchenko, J., et al. 1991. The sustainable biosphere initiative: an ecological research agenda. A report from the Ecological Society of America. Ecology 72:371–412.
MacArthur, R. H., and E. O. Wilson. 1967. The theory of
island biogeography. Princeton University Press, Princeton,
New Jersey, USA.
Margules, C. R. 1992. The Wog Wog fragmentation experiment. Environmental Conservation 19:316–325.
Margules, C. R., and M. P. Austin. 1995. Biological models
for monitoring species decline: the construction and use of
data bases. Pages 183–196 in J. H. Lawton and R. M. May,
editors. Extinction rates. Oxford University Press, Oxford,
UK.
Martikainen, P., L. Kaila, and Y. Haila. 1998. Threatened
beetles in white-backed woodpecker habitats. Conservation
Biology 12:293–301.
May, R. M., J. H. Lawton, and N. E. Stork. 1995. Assessing
extinction rates. Pages 1–24 in J. H. Lawton and R. M.
May, editors. Extinction rates. Oxford University Press,
Oxford, UK.
McCoy, E. D. 1982. The application of island-biogeographic
theory to forest tracts: problems in the determination of
turnover rates. Biological Conservation 22:217–227.
McGarigal, K., and S. Cushman. 2002. Comparative evaluation of experimental approaches to the study of habitat
fragmentation effects. Ecological Applications 12:335–
345.
McIntosh, R. P. 1985. The background of ecology. Concept
and theory. Cambridge University Press, Cambridge, UK.
McIntyre, S., and R. Hobbs. 1999. A framework for conceptualizing human effects on landscapes and its relevance
to management and research models. Conservation Biology
13:1282–1292.
Mladenoff, D. J., T. A. Sickley, R. G. Haight, and A. P. Wydeven. 1995. A regional landscape analysis and a prediction of favourable grey wolf habitat in the northern Great
Lakes region. Conservation Biology 9:279–294.
Niemelä, J., Y. Haila, and P. Punttila. 1996. The importance
of small-scale heterogeneity in boreal forests: variation in
diversity in forest-floor invertebrates across the succession
gradient. Ecography 19:352–368.
Passmore, J. 1980. Man’s responsibility for nature. Ecological problems and Western traditions. Duckworth, London,
UK.
Pickett, S. T. A., V. T. Parker, and P. L. Fiedler. 1992. The
new paradigm in ecology: implications for conservation
biology above the species level. Pages 65–88 in P. L. Fiedler
and S. K. Jain, editors. Conservation biology. The theory
and pracice of nature conservation, preservation and management. Chapman and Hall, London, UK.
Plumwood, V. 1993. Feminism and the mastery of nature.
Routledge, London, UK.
Preston, F. W. 1948. The commonness and rarity of species.
Ecology 29:254–283.
Preston, F. W. 1960. Time and space and the variation of
species. Ecology 41:611–627.
Preston, F. W. 1962. The canonical distribution of commonness and rarity. Ecology 43:185–215,410–432.
Ricklefs, R. E. 1987. Community diversity: relative roles of
local and regional processes. Science 235:167–171.
334
INVITED FEATURE
Saunders, D. A., and S. V. Briggs. 2002. Nature grows in
straight lines—or does she? What are the consequences of
the mismatch between human-imposed linear boundaries
and ecosystem boundaries? An Australian example. Landscape and Urban Planning, in press.
Saunders, D. A., and R. J. Hobbs, editors. 1991. Nature
conservation 2: the role of corridors. Surrey Beatty and
Sons, Chipping Norton, NSW, Australia.
Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991.
Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5:18–32.
Shugart, H. H. 1998. Terrestrial ecosystems in changing Environments. Cambridge University Press, Cambridge, UK.
Simberloff, D. 1997. Biogeographic approaches and the new
conservation biology. Pages 274–284 in S. T. A. Pickett,
R. S. Ostfeld, M. Shachak, and G. E. Likens, editors. The
ecological basis of conservation. Heterogeneity, ecosystems, and biodiversity. Chapman and Hall, New York, New
York, USA.
Simberloff, D., and L. G. Abele. 1976. Island biogeography
theory and conservation practice. Science 191:285–286.
Soulé, M. E., editor. 1986. Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland, Massachusetts, USA.
Stratford, J. A., and P. C. Stouffer. 1999. Local extinctions
of terrestrial insectivorous birds in a fragmented landscape
near Manaus, Brazil. Conservation Biology 13:1416–1423.
Väisänen, R., O. Biström, and K. Heliövaara. 1993. Subcortical Coleoptera in dead pines and spruces: is primeval
species composition maintained in managed forests? Biodiversity and Conservation 2:95–113.
Virkkala, R. 1987. Effects of forest management on birds
breeding in northern Finland. Annales Zoologici Fennici
24:281–294.
Virkkala, R. 1990. Ecology of the Siberian Tit Parus cinctus
in relation to habitat quality: effects of forest management.
Ornis Scandinavica 21:139–146.
Weale, A. 1992. The new politics of pollution. Manchester
University Press, Manchester, UK.
Whitcomb, B. L., R. F. Whitcomb, and D. Bystrak. 1977.
Island biogeography and ‘‘habitat islands’’ of eastern forest. III. Long-term turnover and effects of selective logging
on the avifauna of forest fragments. American Birds 31:
17–23.
Whitcomb, R. F. 1977. Island biogeography and ‘‘habitat
islands’’ of eastern forest. I. Introduction. American Birds
31:3–5.
Whitcomb, R. F. 1987. North American forests and grass-
Ecological Applications
Vol. 12, No. 2
lands: biotic conservation. Pages 163–176 in D. A. Saunders, G. W. Arnold, A. A. Burbridge, and A. J. M. Hopkins,
editors. Nature conservation. The role of remnants of native
vegetation. Surrey Beatty and Sons, Chipping Norton,
NSW, Australia.
Whitcomb, R. F., J. F. Lynch, P. A. Opler, and C. S. Robbins.
1976. Island biogeography and conservation: strategy and
limitations. Science 193:1030–1032.
Whitcomb, R. F., C. S. Robbins, J. F. Lynch, B. L. Whitcomb,
M. K. Klimkiexicz, and D. Bystrak. 1981. Effects of forest
fragmentation on avifauna of the eastern deciduous forest.
Pages 123–205 in R. L. Burgess and D. M. Sharpe, editors.
Forest island dynamics in man-dominated landscapes.
Springer-Verlag, New York, New York, USA.
Whitmore, T. C. 1997. Tropical forest disturbance, disappearance, and species loss. Pages 3–12 in W. F. Laurance
and R. O. J. Bierregaard, editors. Tropical forest remnants.
Ecology, management, and conservation of fragmented
communities. University of Chicago Press, Chicago, Illinois, USA.
Whitmore, T. C., and J. A. Sayer, editors. 1992. Tropical
deforestation and species extinction. Chapman and Hall,
London, UK.
Wiens, J. A. 1981. Scale problems in avian censusing. Studies
in Avian Biology 6:513–521.
Wiens, J. A. 1984. On understanding a non-equilibrium
world: Myths and reality in community patterns and processes. Pages 439–457 in D. R. J. Strong, D. Simberloff,
L. G. Abele, and A. B. Thistle, editors. Ecological communities. Conceptual issues and the evidence. Princeton
University Press, Princeton, New Jersey, USA.
Wiens, J. A. 1989a. The ecology of bird communities. Volume 1. Foundations and patterns. Cambridge University
Press, Cambridge, UK.
Wiens, J. A. 1989b. Spatial scaling in ecology. Functional
Ecology 3:385–397.
Wiens, J. A. 1995a. Habitat fragmentation: island v landscape
perspective on bird conservation. Ibis 137:S97–S104.
Wiens, J. A. 1995b. Landscape mosaics and ecological theory. Pages 1–26 in L. Hansson, L. Fahrig, and G. Merriam,
editors. Mosaic landscapes and ecological processes. Chapman and Hall, London, UK.
Williams, C. B. 1943. Area and number of species. Nature,
London 152:264–267.
Williams, C. B. 1964. Patterns in the balance of nature. Academic Press, London, UK.
Wilson, E. O. 1992. The diversity of life. Harvard University
Press, Cambridge, Massachusetts, USA.