REVIEWS REVIEWS REVIEWS
Gulliver travels to the fragmented tropics:
geographic variation in mechanisms of
avian extinction
Jeffrey A Stratford1 and W Douglas Robinson2
Irrespective of geography, forest destruction and fragmentation lead to lower avian species richness. The
underlying mechanisms causing local extirpations have been studied most thoroughly in northern temperate
landscapes, where higher levels of brood parasitism, nest predation, and possibly decreased food availability
are responsible for the loss of some species. Tropical landscapes are being similarly altered, but studies of
responses by tropical birds remain relatively scarce. Predicting how tropical birds respond to habitat loss and
fragmentation should not be extended directly from the results of temperate investigations. Tropical birds
possess different evolutionary and life histories, which make them vulnerable to a different suite of threats
than those normally considered for birds from temperate regions. These same traits, including greater physiological and sensory specialization, reduced dispersal capabilities, and much lower local and regional population densities, indicate that strategies for conserving bird diversity will be different in tropical landscapes than
those for temperate regions.
Front Ecol Environ 2005; 3(2): 91–98
are some laws and customs in this empire very
“There
peculiar, and if they were not so directly contrary
to those of my own dear country, I should be tempted to
say little in their justification” (Swift 1726). In Jonathan
Swift’s classic tale, Gulliver’s Travels, Gulliver finds himself in unfamiliar and extraordinary lands. What makes
his experiences so intriguing and comical are his transplanted notions of order and society colliding with the
very different notions held by his exotic hosts. In a similar way, ecological inferences derived from one geographic area may not transfer smoothly to others and, in
fact, may impede further understanding. Such may be the
case when comparing the ecology and evolution of tropical and temperate avifaunas, particularly neotropical and
northern temperate assemblages.
In a nutshell:
• Bird diversity declines in landscapes altered by habitat loss and
fragmentation, but most studies have been done in northern
temperate areas
• Tropical birds are also declining in disturbed landscapes, but as
a result of different threats
• Lower tolerance of microclimatic changes or light environments in fragmented habitats and lower dispersal abilities are
factors generally not considered for temperate birds
• Preservation of tropical diversity will improve if these traits are
considered during reserve design
1
Department of Biological Sciences, Auburn University, Auburn, AL;
Oak Creek Lab of Biology, Department of Fisheries and Wildlife,
Oregon State University, Corvallis, OR ([email protected])
2
© The Ecological Society of America
Differences in life histories and behaviors between
temperate and tropical birds have been recognized for a
long time (Skutch 1949). Alexander Skutch, during a
famous series of debates with David Lack about the latitudinal gradient in avian clutch sizes (Lack 1947; Skutch
1949), remarked that if most biologists had been raised
in the tropics they would ask why temperate birds lay
unusually large clutches, rather than asking, as did temperate biologists, why tropical clutches are so small.
Similarly, Ricklefs (2002) noted that his time in the
neotropics was pivotal in cementing the idea that many
neotropical bird lineages are distinct from North
American ones in both their evolutionary and life histories. The degree to which such differences influence the
conservation of bird populations in human-dominated
landscapes remains unappreciated.
Nevertheless, two present-day commonalities exist
between tropical and temperate birds. First, species in
both realms are confronted with destruction and conversion of forests (Faaborg et al. 1995; Laurance et al. 2000).
Areas of high endemism, such as the Atlantic rainforests
of Brazil, the northern and central Andes, and Amazonia,
are particularly at risk (Balmford and Long 1994).
Secondly, many species respond negatively to such disturbances and their populations decline or disappear in fragmented landscapes. Local extinction of species from fragmented forests is well documented across northern
temperate areas (Whitcomb et al. 1981) as well as in the
neotropics (Stouffer and Bierregaard 1995; Robinson
1999, 2001). Thus, temperate and tropical birds face similar challenges. Yet the mechanisms producing the
observed declines in species richness may differ between
www.frontiersinecology.org
91
Geography of bird extinctions
92
JA Stratford and WD Robinson
In contrast, brood parasitism
appears to be a minor risk to forest
birds in most mainland neotropical
landscapes. Although two generalist
brood parasites, shiny (M bonariensis)
and bronzed cowbirds (M aeneus),
occur in Central and South American landscapes, current levels of
brood parasitism in fragmented
forests are generally low (Figure 1).
The risks of brood parasitism can be
greater in more subtropical landscapes, such as Mexico, Argentina,
and the West Indies, where cowbirds
tend to be more numerous, and have
caused major declines in some
species, such as the yellow-shouldered
blackbird (Agelaius xanthomus)
Figure 1. A pasture with cattle surrounding a forest fragment in central Brazil. In North (Wiley et al. 1991). Furthermore,
American landscapes, songbirds breeding in the remnant forest would suffer from high where brood parasites do occur regubrood parasitism, but in many neotropical landscapes proliferation of brood parasitic larly in equatorial areas, they tend to
cowbirds is not yet a problem.
parasitize only a few specific host
species with whom they may be
the two regions. Conservation strategies applied in tem- tightly coevolved, such as the giant cowbird (Scaphidura
perate North America may therefore be less effective in oryzivora) and its hosts, oropendolas and caciques (Payne
the neotropics.
1997). Thus, one of the primary threats to forest birds living
Here, we briefly review currently accepted hypotheses for in temperate fragments is currently of little or no consebird extirpations as a result of forest fragmentation. We also quence to most tropical species. How neotropical cowbird
identify several hypotheses that should be considered and distributions might change as landscapes become deforested
studied in detail for neotropical birds. Our goal is not to pro- has not, to our knowledge, been evaluated.
vide an exhaustive review of the hypotheses proposed to
explain declines or local extinctions of birds, but rather to Nest predation
highlight the salient differences in responses to environmental change by tropical and temperate birds. We focus largely Reproductive failure in forest fragments may also be
on comparing results from northern temperate areas with caused by abnormally elevated nest predation (Robinson
those from tropical ones. Further study is needed to under- et al. 1995). Some edge-loving nest predators, such as
stand the degree to which factors operating in northern tem- skunks, raccoons, and snakes, are more abundant in
human-altered landscapes (Chalfoun et al. 2002). Nests
perate landscapes also operate in southern temperate areas.
in forest fragments may be easier for predators to detect,
especially those located along edges, where predator
Northern temperate extinction mechanisms
activity is greater (Chalfoun et al. 2002). Other characteristics of the habitat remnant, such as vegetation strucBrood parasitism
ture or food supply, may nevertheless remain attractive to
Of the primary factors responsible for population declines in birds. In such instances, forest birds may be common, but
the temperate zone, brood parasitism (the laying of eggs in fail at reproduction because of brood parasitism or nest
nests of other bird species and subsequent abandonment of predation, a situation sometimes referred to as an ecologoffspring to be raised by foster parents) by cowbirds has ical trap (Misenhelter and Rotenberry 2000).
received perhaps the most attention, at least in eastern and
Nest predation has yet to be studied rigorously in tropimidwestern North America (Smith et al. 2000). Brown- cal landscapes, fragmented or not, largely because of the
headed cowbirds (Molothrus ater) increase in abundance in difficulty of finding sufficient numbers of nests of compardisturbed landscapes, particularly fragmented forests in an atively rare species (Robinson et al. 2000). Extinctions of
agricultural matrix (Robinson et al. 1995). High levels of birds from Barro Colorado Island, Panama, a former hillcowbird parasitism can cause population declines, which top of lowland rainforest isolated during flooding of the
increases the risk of local extirpations (Hoover 2003), and Panama Canal, have been explained by elevated levels of
even extinctions for a few extremely vulnerable species, nest predation (Terborgh 1974). The island, at 1562 ha,
such as Kirtland’s warbler (Dendroica kirtlandii) (Kepler et al. is too small to maintain populations of top carnivores,
1996).
such as big cats and raptors, which may have allowed
www.frontiersinecology.org
© The Ecological Society of America
JA Stratford and WD Robinson
Geography of bird extinctions
numbers of medium-sized mammalian
nest predators to increase unchecked.
Studies using artificial bird nests suggested that predation was greater on
Barro Colorado Island than nearby
mainland sites (Sieving 1992).
However, the accuracy with which
artificial nests measure predation of
real bird nests is questionable (Moore
and Robinson 2004). Furthermore,
mammalian nest predators appear not
to be abundant on Barro Colorado
compared with nearby mainland forest where big cats and eagles are
extant (Wright et al. 1994). Finally,
video evidence of predators attacking
understory bird nests on the island
indicated that 80% of attacks were by
snakes, not mammals (Figure 2;
Robinson et al. in press).
Despite such concerns, can results Figure 2. A neotropical bird-eating snake (Pseustes poecilinota) consumes a great
from Barro Colorado be generalized tinamou (Tinamus major) egg in central Panama. Snakes are important predators in the
to other tropical landscapes? Because lowland tropics and middle temperate latitudes, but are relatively uncommon or absent in
they are surrounded by water and are far northern temperate areas.
therefore relatively protected from
invasions that threaten fragments surrounded by agricul- reduced prey abundance hypothesis is currently in doubt
tural and suburban habitats (such as spread of edge-loving because many species of potential arthropod prey can
nest predators), islands are good models for understanding become more abundant after fragmentation (Didham
ecosystem disintegration (Leigh et al. 2002). Yet, most 1997). Indeed, in Costa Rica, Sekercioglu et al. (2002)
fragments occur in areas where access by species from sur- did not find any differences in prey biomass between
rounding habitats is much easier. As discussed above, birds forested and fragmented sites.
in temperate regions suffer because of this. In tropical
landscapes, however, many of which are inhabited by Suggested additional mechanisms for neotropical
small farmers, ease of access is also greater for humans,
bird extinctions
who commonly persecute snakes and larger mammals
(Redford 1992). In a sense, then, top-down control of nest Nest predation, brood parasitism, and prey abundance are
predators could be greater in tropical fragmented land- the three hypotheses commonly proposed to explain local
scapes, allowing bird populations greater opportunity to extinctions in temperate and neotropical landscapes
breed successfully. Apart from various artificial nest stud- (Figure 3). Yet there is either no data to support these
ies, no other data from tropical landscapes allow an evalu- hypotheses in the tropics, or the evidence suggests that
fragmentation has the opposite effect there. How, then,
ation of the effects of fragmentation on predation risk.
can we explain why neotropical birds in forest fragments
become extinct? Instead of explanations that focus on
Prey abundance
biotic interactions, we suggest that the evolutionary hisForest fragmentation has been associated with decreased tories of neotropical birds have resulted in physiologies
prey abundance for near-ground or ground-foraging insec- and behaviors that make them less tolerant of environtivorous birds (Burke and Nol 1998). Declines in food mental variation.
availability may be a consequence of microclimatic
changes such as increased temperature and decreased Physiological constraints
humidity near fragment edges (Williams-Linera 1990;
Saunders et al. 1999). In the neotropics, arthropod popu- Microclimatic changes in fragments may affect birds
lation responses to fragmentation are complex and poorly directly if altered temperature and humidity levels in forunderstood. As might be expected, the effects of habitat est remnants expose them to evapotranspiration rates
fragmentation differ across arthropod taxa and through that cause physiological stress. The idea of greater habitat
time as the edge changes in structure (Didham 1997). specialization in the tropics is an old one (Janzen 1967;
Rainforest fragmentation does alter the composition of Karr and Freemark 1983), but has received less attention
arthropod communities, but the applicability of the than it deserves (Marra and Remsen 1997). It is under© The Ecological Society of America
www.frontiersinecology.org
93
Geography of bird extinctions
Courtesy NK Strycker
94
JA Stratford and WD Robinson
temperatures varying from winter to summer by an order
of magnitude more than a resident rainforest species
would experience in its entire lifetime, and more than
some tropical species’ ancestors might have experienced
in thousands of generations. Unlike species from temperate areas, many neotropical species may be much less
physiologically tolerant of climatic variation, particularly
if those species inhabit the relatively stable understories
of closed-canopy mature forest.
Evaluations of this hypothesis are lacking, because of
the rarity of measurements of metabolic rates, particularly
with respect to ranges of thermal tolerance in neotropical
birds. However, some indirect evidence is available.
Resident insectivorous birds in Panama move to sites with
higher humidity during the heat of dry season days (Karr
and Freemark 1983). Two common understory species
near Manaus, Brazil, exhibit slower feather growth rates in
forest fragments compared to the feather growth rates of
Figure 3. Some specialized foragers, such as this bicolored birds in continuous forest (Stratford and Stouffer 2001).
antbird (Gymnopithys leucaspis) in Panama, gather all their In the Afrotropics, feather asymmetries of birds captured
food by catching arthropods fleeing from raiding swarms of in small fragments were more severe than those of birds
predatory army ants. When the ants disappear from forest birds in larger fragments, suggesting developmental probfragments, the birds do too.
lems in the birds from small fragments (Lens et al. 2002).
Phylogenetically controlled comparisons of temperate
standable that ornithologists studying birds in temperate and tropical bird physiological tolerances are needed. If
regions tend to look elsewhere for useful hypotheses. the metabolic constraint hypothesis is valid, many tropiMany of these species are probably more tolerant, given cal species should have narrower thermoneutral zones and
the wide range of conditions they encounter during a cal- be less tolerant of small changes in temperature and
endar year; migratory species, for example, must endure humidity. Predicting the size of those small changes could
highly variable climatic conditions as they journey thou- be accomplished by measuring annual ranges of temperasands of kilometers. Non-migratory species must survive tures and humidities experienced in the interior of unfragmented forests (Figure 4). We also predict
that the fraction of bird communities disappearing in fragmented landscapes will be
negatively correlated with the degree of
annual and daily climatic variation. For
example, bird communities in more seasonal tropical environments, such as seasonally dry forests, some deserts, and higher
altitude habitats, should be more tolerant
to fragmentation, whereas relatively aseasonal lowland forest bird communities are
likely to contain a greater proportion of
fragmentation-sensitive species.
Bird species whose habitat use is
restricted to the most unvarying microhabitats, such as the floor of old-growth
forests, should be more sensitive to habitat
fragmentation than birds inhabiting forest
microhabitats with wider climatic fluctuations, such as canopy- and edge-dwelling
Figure 4. Magnitude of daily temperature fluctuations within continuous forest species (Figure 4). Species distribution
increases from the ground to the canopy and from the forest interior towards the data are consistent with this prediction,
forest edge and into grassland or pasture surrounding fragments (Didham and revealing that gap specialists and edge-lovLawton 1999, and unpublished data). We hypothesize that species using the least ing species (eg many types of hummingvariable microhabitats will be the most likely to disappear from landscapes lacking birds) are tolerant of fragmentation
such relatively stable microclimatic refugia.
(Stouffer and Bierregaard 1995). In addiwww.frontiersinecology.org
© The Ecological Society of America
JA Stratford and WD Robinson
tion, the area available to a sensitive understory species
should be lower than the actual physical size of the fragments. Since edges are often exposed to sun and drying
winds, they will tend to experience more microclimatic
variation than interior sites, which are protected from
such forces. Sensitive species would, therefore, drop out of
fragments earlier than would be predicted based solely on
habitat patch area. The role of physiological tolerances or
metabolic constraints as mechanistic explanations for
tropical bird extinctions is ripe for further evaluation.
Geography of bird extinctions
components and cell sensitivities between fragmentation-sensitive and fragmentation-tolerant species.
Comparisons could be done in a phylogenetically controlled manner, given the occurrence within some families (eg Thamnophilidae) of both forest-interior and gapspecialist species. Additional experiments could include
habitat choice tests in chambers where light environments are manipulated and bird use of those environments is quantified.
Dispersal ability
Visual constraints
Another hypothesis that was previously dismissed as basiJust as deforestation and forest fragmentation may pre- cally irrelevant to temperate breeding birds is that some
sent abiotic conditions that stress thermoregulatory species could experience difficulties colonizing isolated
processes, visual constraints may also limit habitat use habitat remnants. Given that birds are generally very
(Canaday 1996). We know that some neotropical birds, mobile, the inability or unwillingness of neotropical birds
such as manakins and cotingas, are highly sensitive to the to traverse gaps seems surprising to many temperate-zone
light environments they select for sexual displays (Endler biologists (Harris and Reed 2002). This attitude might be
and Théry 1996). Likewise, many other species may be changing, since more recent evidence suggests that some
sensitive to changes in light levels. In mature tropical for- species from temperate regions may be more inhibited
est, less than 5% of the light striking the canopy pene- from crossing gaps than others (Bélisle et al. 2000; Gobeil
trates to the ground (Shuttleworth et al. 1984). Bird and Villard 2002). Considering that many temperatespecies foraging exclusively in the lower understory or on zone species fly thousands of kilometers during migration,
the ground of evergreen tropical forests must therefore see it is particularly interesting that we are discovering how
well in low-light conditions. To do so, some species may some species settle for a breeding season and then appear
have unusually large eyes to enhance light gathering or to restrict what types of landscapes they will move
have internal structural modifications that otherwise through (Bélisle and St. Clair 2001).
enhance visual acuity in poorly lit situations.
In contrast, dispersal limitation is a longstanding
A potential cost of such specialization is an inability to hypothesis among tropical ornithologists (Wallace 1853;
use brightly lit habitats such as large forest gaps, edges, or Willis 1974; Mayr and Diamond 2001). Within the tropcanopies. Although such constraints will probably not lead ics, there may be hundreds of forest-dwelling species that
directly to higher extinction risk, they could explain low will not move across non-forested habitats (Figure 5). For
rates of recolonization after local extirpation by retarding instance, Develey and Stouffer (2001) found that many
the progress of light-sensitive species across brightly lit mixed-flock species were unlikely to cross a road of only
habitat gaps. It is also possible that
species sensitive to bright light may
be unable to detect the presence of
distant habitat patches in a landscape. That is, their perceptual range
could be limited compared to species
that forage regularly in gaps, edges,
and other brighter habitats.
Given the general absence of
such light-limited habitats for diurnal birds in temperate areas, this
degree of specialization is unlikely
to be an important component of
extinction risk for these species.
Particular groups of neotropical
birds that might be more sensitive
to light conditions are the ground
antbirds (Formicariidae) and some
ovenbirds (Sclerurus leaftossers).
Tests of the light-sensitivity Figure 5. A river in central Brazil. Alfred Russell Wallace first noticed that few forest
hypothesis should include histologi- birds fly across water in the lowland neotropics. Genetic studies have also demonstrated
cal comparisons of ocular structural reduced avian gene flow across tropical rivers.
© The Ecological Society of America
www.frontiersinecology.org
95
Geography of bird extinctions
JA Stratford and WD Robinson
may not be as difficult as flights across
open water if some species are willing
to leave forest but need stopping
points along the way to ensure a successful journey.
96
Low population densities and large
home ranges
The average abundance of most forestdwelling, neotropical species is an
order of magnitude lower than their
north temperate ecological counterparts (Terborgh et al. 1990; Robinson
et al. 2000). In Amazonian Peru, 35%
of species occur in densities of <1
pair/100 ha and median abundance is
2.5 pairs/100 ha (Terborgh et al. 1990).
Figure 6. Diagrammatic representation of recolonization patterns in a temperate In other Amazonian sites, median
fragment, Trelease Woods, Illinois, US, and a tropical fragment, Barro Colorado abundance may be even lower
Island, Panama. Stippled bars indicate years in which survey data were available. (Stouffer and Bierregaard 1995). Low
Solid bars indicate dates each species was present. Species are year-round residents, so abundance goes hand in hand with
migratory behavior does not account for recolonizations. Data from Barro Colorado much larger home ranges in the tropare for six year-round residents with loud, distinctive songs, which are therefore ics, establishing a greater degree of
unlikely to be missed by ornithologists if the species were present. Recolonization of area-sensitivity than is normally seen
Trelease Woods after local extirpation was common, whereas none of the species in bird species from northern temperrecolonized Barro Colorado Island.
ate zones. The smallest territories of
passerines in Amazonia are 3 to 4 ha,
30 m in width. Consequently, when populations in iso- whereas many breeding season territories of birds in
lated forest remnants disappear, for whatever reason, the temperate areas are less than 1 ha (Terborgh et al. 1990).
probability of recolonization ever reconstituting a popu- When forests are fragmented, the rarity of most tropical
lation will be minimal, perhaps even zero, for some species will result in a low probability of occurrence.
species (Robinson 1999). Such dramatic restrictions on Likewise, large territory sizes and patchy distributions
immigration are highly unlikely in most temperate land- will ensure that few individuals of most species will be
scapes, perhaps because of the propensity of many species present even in 100-ha fragments, a size that many biolto exhibit long-distance movements during migration ogists studying northern temperate species would conand postnatal dispersal. Studies of genetic differentiation sider to be large. The probability of long-term persissupport the idea that many neotropical species are quite tence in all but fragments of hundreds to thousands of
sedentary (Capparella 1988; Bates 2000), whereas most hectares is expected to be small unless recolonizations
species from northern temperate areas show little genetic are frequent.
differentiation across their geographic ranges (Lovette et
al. 2003).
Conclusions
Long-term datasets evaluating extinction and recolonization phenomena are rare for any landscape. A simple The unusually specialized lives of tropical birds, relative
comparison between Kendeigh’s (1980) long-term sur- to temperate species, subjects them to greater risk of
veys of Trelease Woods in central Illinois, USA, and extinction in altered landscapes. The most sensitive
long-term data from Barro Colorado Island, Panama species are highly adapted to environments of the under(Willis and Eisenman 1979; Robinson 1999), indicates story and forest floor, where variations in humidity, temregular recolonization in the temperate fragment but not perature, and light are very small. Furthermore, at least
in the tropical one (Figure 6). Acquiring additional over recent evolutionary time, the continuous distribulong-term survey data from fragmented landscapes is tion of mature forest means that year-round resident
needed to better understand this possible latitudinal dif- species rarely, if ever, had to cross non-forest habitats,
ference in immigration tendencies. Experimental resulting in morphological specializations for walking
approaches are also needed to evaluate the abilities of and reduced abilities for sustained flight. Although many
tropical species to cross habitat gaps of various distances of these hypotheses still require intensive evaluation by
and across various habitat types. For example, dispersal experimental and observational studies, the pace at
across agricultural habitats such as pastures or row crops which tropical landscapes are being fragmented demands
www.frontiersinecology.org
© The Ecological Society of America
JA Stratford and WD Robinson
Geography of bird extinctions
that we move forward and design
tropical reserve systems with these
specialized lives of tropical birds in
mind. We suggest the following
approaches.
First, given the very large home
ranges and the general rarity of
most species, every effort must be
made to preserve the largest habitat
remnants still extant. In many
cases, this should include the
preservation of many small forest
patches in a region, followed by the
reforestation of surrounding disturbed habitats. Some dispersallimited species may not move
across pastures, but will move
through young second-growth forest, which will facilitate passive
recolonization of smaller fragments. Figure 7. An abandoned pasture surrounding a forest fragment in Brazil. Revegetation
Second, when protection of the along margins of fragments provides some insulation against effects of drying winds and solar
matrix surrounding fragments is penetration. As regrowth of trees in the pasture continues, more fragmentation-sensitive bird
impossible, corridors connecting species will disperse across the landscape and recolonization of fragments will increase.
isolated remnants are essential.
Reducing the isolation of patches by reconnecting them References
is absolutely critical to promote dispersal in the tropics, Balmford A and Long A. 1994. Avian endemism and forest loss.
Nature 372: 623–24.
although it may not be necessary in many temperate
Bates JM. 2000. Allozymic genetic structure and natural habitat
landscapes (Hannon and Schmiegelow 2002).
fragmentation: data for five species of Amazonian forest birds.
Third, the effective area of remnants will be increased
Condor 102: 770–83.
if microclimatic variability is reduced and the light envi- Bélisle M and St Clair CC. 2001. Cumulative effects of barriers on
the movements of forest birds. Conserv Ecol 5: 9. www.
ronment is carefully regulated (Figure 7). Three strateecologyandsociety.org/vol5/iss2/art9/. Viewed 30 October
gies will help here. One is to construct vegetative buffers
2004.
around edges of new fragments to reduce wind and light Bélisle M, Desrochers A, and Fortin M-J. 2001. Influence of forest
penetration. In many neotropical locations, the creation
cover on movements of forest birds: a homing experiment.
Ecology 82:1893–1904.
of buffer zones can be accomplished relatively quickly
and passively by allowing second-growth trees to grow Burke DM and Nol E. 1998. Influence of food abundance, nest site
habitat, and forest fragmentation on breeding Ovenbirds. Auk
around the margins of forest fragments. Additional con115: 96–104.
siderations are to exclude cattle and other grazers which Canaday C. 1996. Loss of insectivorous birds along a gradient of
defoliate the ground and understories of forests and to
human impact in Amazonia. Biol Conserv 77: 63–77.
exclude, or carefully limit, timber extraction, as the Capparella AP. 1988. Genetic variation in neotropical birds: implications for the speciation process. Proceedings of the
removal of timber permits increased solar penetration by
International Ornithological Congress 19: 1658–64.
opening up canopies.
Chalfoun AD, Thompson III FR, and Ratnaswamy MJ. 2002. Nest
These approaches will also benefit conservation of
predators and fragmentation: a review and meta-analysis.
birds in temperate regions, but many tropical birds lack
Conserv Biol 16: 306–18.
Develey PF and Stouffer PC. 2001. Effects of roads on movements
the same levels of resilience to disturbance because of
by understory birds in mixed-species flocks in central
their evolutionary and life histories. Gulliver would also
Amazonian Brazil. Conserv Biol 15: 1416–22.
have found the lives of tropical birds to be very peculiar.
Didham RK. 1997. An overview of invertebrate responses to forest
Understanding these peculiarities will lead to effective
fragmentation. In: Watt A, Stork NE, and Hunter M (Eds).
conservation.
Forests and insects. London, UK: Chapman and Hall.
Acknowledgements
We thank the Brazilian and Panamanian natural resource
agencies for allowing us to study birds in their countries.
The work was supported by the Auburn University
Center for Forest Sustainability and National Science
Foundation grants 0212587, 0408186, and 0422233.
© The Ecological Society of America
Didham RK and Lawton JH. 1999. Edge structure determines the
magnitude of changes in microclimate and vegetation structure
in tropical forest fragments. Biotropica 31: 17–30.
Endler JA and Théry M. 1996. Interacting effects of lek placement,
display behavior, ambient light, and color patterns in three
neotropical forest-dwelling birds. Am Nat 148: 421–52.
Faaborg J, Brittingham M, Donovan T, and Blake J. 1995. Habitat
fragmentation in the temperate zone. In: Martin TE and Finch
DM (Eds). Ecology and management of neotropical migratory
www.frontiersinecology.org
97
Geography of bird extinctions
98
birds. New York, NY: Oxford University Press.
Gobeil J-F and Villard M-A. 2002. Permeability of three boreal forest landscape types to bird movements as determined from
experimental translocations. Oikos 98: 447–58.
Hannon SJ and Schmiegelow FKA. 2002. Corridors may not
improve the conservation value of small reserves for most
boreal birds. Ecol Appl 12: 1457–68.
Harris RJ and Reed JM. 2002. Behavioral barriers to non-migratory
movements of birds. Ann Zool Fenn 39: 275–90.
Hoover JP. 2003. Multiple effects of brood parasitism reduce the
reproductive success of prothonotary warblers, Protonotaria citrea. Anim Behav 65: 923–34.
Janzen DH. 1967. Why mountain passes are higher in the tropics.
Am Nat 101: 233–50.
Jullien M and Thiollay J-M. 1996. Effects of rain forest disturbance
and fragmentation: comparative changes of the raptor community along natural and human-made gradients in French
Guiana. J Trop Ecol 23: 7–25.
Karr JR and Freemark KE. 1983. Habitat selection and environmental gradients: dynamics in the “stable” tropics. Ecol Monogr
64: 1481–94.
Kendeigh SC. 1980. Bird populations in east central Illinois: fluctuations, variations, and development over a half-century.
Urbana, IL: University of Illinois Press.
Kepler CB, Irvine GW, Decapita ME, and Weinrich J. 1996. The
conservation management of Kirtland’s Warbler. Bird Conserv
Int 6: 11–22.
Lack D. 1947. The significance of clutch-size. Ibis 89: 302–52.
Laurance WF, Cochrance MA, Bergen S, et al. 2000. The future of
the Brazilian Amazon. Science 291: 438.
Leigh Jr EG, Cosson J-F, Pons J-M, and Forget P-M. 2002. En quoi
l’étude des ilots forestiers permet-elle de mieux connaitre le
fonctionnement de la foret tropicale? Revue d’Ecologie 57:
181–94.
Lens L, Van Dongen S, Norris K, et al. 2002. Avian persistence in
fragmented rainforest. Science 298: 1236–38.
Lovette IJ, Clegg SM, and Smith TB. 2003. Limited utility of
mtDNA markers for determining connectivity among breeding
and overwintering locations in three neotropical migrant birds.
Conserv Biol 18: 156–66.
Marra PP and Remsen JV. 1997. Insights into the maintenance of
high species diversity in the neotropics: habitat selection and
foraging behavior in understory birds of tropical and temperate
forests. Ornithological Monographs 48: 445–83.
Mayr E and Diamond J. 2001. The birds of northern Melanesia:
speciation, ecology & biogeography. New York, NY: Oxford
University Press.
Misenhelter MD and Rotenberry JT. 2000. Choices and consequences of habitat occupancy and nest site selection in sage
sparrows. Ecology 81: 2892–2901.
Moore R and Robinson WD. 2004. Artificial bird nests, external
validity, and bias in ecological field studies. Ecology 85:
1562–67.
Payne RB. 1997. Avian brood parasitism. In: Clayton DH and
Moore J (Eds). Host–parasite evolution: general principles and
avian models. Oxford, UK: Oxford University Press.
Redford KH. 1992. The empty forest. Bioscience 42: 412–22.
Ricklefs RE. 2002. Splendid isolation: historical ecology of the
South American passerine fauna. J Avian Biol 33: 207–11.
Robinson SK, Thompson III FR, Donovon TM, et al. 1995.
Regional forest fragmentation and the nesting success of migratory birds. Science 267: 1987–90.
Robinson WD. 1999. Long-term changes in the avifauna of Barro
www.frontiersinecology.org
JA Stratford and WD Robinson
Colorado Island, Panama, a tropical forest isolate. Conserv Biol
13: 85–97.
Robinson WD. 2001. Changes in abundances of birds in a neotropical forest fragment over 25 years. Anim Biodivers Conserv 24:
51–65.
Robinson WD, Brawn JD, and Robinson SK. 2000: Forest bird
community structure in Central Panama: influence of spatial
scale and biogeography. Ecol Monogr 70: 209–35.
Robinson WD, Rompré GR, and Robinson TR. Videography of
Panama bird nests shows snakes are principal predators.
Ornitologia Neotropical. In press.
Saunders SC, Chen J, Drummer TD, and Crow TR. 1999.
Modeling temperature gradients across edges over time in a
managed landscape. Forest Ecol Manag 117: 17–31.
Sekercioglu CH, Ehrlich PR, Daily GC, et al. 2002. Disappearance
of insectivorous birds from tropical forest fragments. P Natl
Acad Sci 99: 263–67.
Sieving KE. 1992. Nest predation and differential insular extinction among selected forest birds of central Panama. Ecology 73:
2310–28.
Shuttleworth WJ, Gash JHC, Moore J, et al. 1984. Observations of
radiation exchange above and below Amazonian forest. Q J
Roy Meteor Soc 110: 1163–69.
Skutch AF. 1949. Do tropical birds rear as many young as they can
nourish? Ibis 91: 430–55.
Smith JMN, Cook TL, Rothstein SI, Robinson SK, and Sealy SG.
2000. Ecology and management of cowbirds and their hosts.
Austin, TX: University of Texas Press.
Stouffer PC and Bierregaard Jr RO. 1995. Use of Amazonian forest
fragments by understory insectivorous birds. Ecology 76:
2429–45.
Stratford JA and Stouffer PC. 2001. Reduced feather growth rates
of two common birds inhabiting central Amazonian forest fragments. Conserv Biol 15: 721–28.
Swift J. 1726. Travels into several remote nations of the world.
London, UK: Benjamin Motte.
Terborgh J. 1974. Preservation of natural diversity: the problem of
extinction prone species. BioScience 24: 715–22.
Terborgh J, Robinson SK, Parker III TA, et al. 1990. Structure and
organization of an Amazonian forest bird community. Ecol
Monogr 60: 213–38.
Wallace AR. 1853. A narrative of travels on the Amazon and Rio
Negro with an account of the native tribes, and observations
on the climate, geology, and natural history of the Amazon
Valley. London, UK: Reeve & Co.
Whitcomb RF, Robbins CS, Lynch JF, et al. 1981. Effects of forest
fragmentation on avifauna of the eastern deciduous forest. In:
Burgess RL and Sharpe DM (Eds). Forest island dynamics in
man-dominated landscapes. New York, NY: Springer-Verlag.
Wiley JW, Post W, and Cruz A. 1991. Conservation of the yellowshouldered blackbird (Agelaius xanthomus), an endangered
West Indian species. Biol Conserv 55: 119–38.
Williams-Linera G. 1990. Vegetation structure and environmental
conditions of forest edges in Panama. J Ecol 78: 356–73.
Willis EO. 1974. Populations and local extinctions of birds on
Barro Colorado Island, Panama. Ecol Monogr 44: 153–69.
Willis EO and Eisenmann E. 1979. A revised list of birds of Barro
Colorado Island, Panama. Smithsonian Contributions to Zoology
291.
Wright SJ, Gompper ME and DeLeon B. 1994. Are large predators
keystone species in neotropical forests? The evidence from
Barro Colorado Island. Oikos 71: 279–94.
© The Ecological Society of America