Ecology of Bird Introduction

30 Sep 2003 14:24 AR AR200-ES34-04.tex AR200-ES34-04.
sgm LaTeX2e(2002/01/18) P1: GJB

10.1146/annurev.ecolsys.34.011802.132353
Annu. Rev. Ecol. Evol. Syst. 2003. 34:71–98

doi: 10.1146/annurev.ecolsys.34.011802.132353
Copyright ° c 2003 by Annual Reviews. All rights reserved
First published online as a Review in Advance on July 8, 2003
THE ECOLOGY OF BIRD INTRODUCTIONS

Richard P. Duncan,1 Tim M. Blackburn,2 and Daniel Sol3
1
Ecology and Entomology Group, Soil, Plant and Ecological Sciences Division,
Access provided by Eastern Kentucky University - Crabbe Library on 01/28/19. For personal use only.
P.O. Box 84, Lincoln University, Canterbury, New Zealand; email: [email protected]
2
School of Biosciences, University of Birmingham, Edgbaston, Birmingham B15 2TT,
Annu. Rev. Ecol. Evol. Syst. 2003.34:71-98. Downloaded from www.annualreviews.org
United Kingdom; email: [email protected]

3
Department of Biology, McGill University, 1205 avenue Docteur Penfield, Montréal,
Québec, H3A 1B1 Canada; email: [email protected]
Key Words establishment success, exotic species, introduced birds, biological

invasion, biotic resistance, invasive species
■ Abstract A growing number of species have been transported and introduced by
humans to new locations and have established self-sustaining wild populations beyond
their natural range limits. Many of these species go on to have significant environ-
mental or economic impacts. However, not all species transported and introduced to
new locations succeed in establishing wild populations, and of the established species
only some become widespread and abundant. What factors underlie this variation in
invasion success? Here, we review progress that has been made in identifying factors
underpinning invasion success from studies of bird introductions. We review what is
known about the introduction, establishment, and spread of introduced bird species,
focusing on comparative studies that use historical records to test hypotheses about
what factors determine success at different stages in the invasion process. We close
with suggestions for future research.
INTRODUCTION
As humans have spread around the globe, they have deliberately or accidentally
transported a huge variety of plant and animal species to locations beyond their
natural ranges (Elton 1958; Lodge 1993; Williamson 1996, 1999). By design or
accident many of these introduced species have established self-sustaining wild
populations, which have subsequently increased and spread to varying degrees.
Many of these species have profoundly affected the ecosystems they have in-
vaded, and some have imposed substantial economic and health costs on human
societies (Elton 1958, Ebenhard 1988, Lever 1994, Simberloff 1995, Williamson
1996, Vitousek et al. 1997, Parker et al. 1999, Dalmazzone 2000, Mack et al.
2000, Perrings et al. 2000, McNeely 2001). Thus, the establishment and spread
of introduced species is currently recognized as a major risk worldwide. This risk
is likely to increase as greater volumes of transport and trade increase the rate at
which novel species are introduced to new locations (Ricciardi et al. 2000).
1543-592X/03/1215-0071$14.00 71
30 Sep 2003 14:24 AR AR200-ES34-04.tex AR200-ES34-04.sgm LaTeX2e(2002/01/18) P1: GJB
72 DUNCAN ¥ BLACKBURN ¥ SOL
Nevertheless, only a small proportion of the species introduced to a new lo-

cation establish wild populations, and of these only a small proportion become
abundant or widespread and have significant impact (Lodge 1993, Williamson
1996, Williamson & Fitter 1996). Given that once established the eradication or
control of introduced species is costly, the most effective way to minimize their im-
pact is to prevent establishment or spread in the first place (Ricciardi & Rassmusen
1998, Mack et al. 2000). This approach requires that we understand the factors
underlying success at different stages in the invasion process (Figure 1) so that we
can identify situations where invasion risk is high. Specifically, what are the fac-
tors that allow certain species to establish and spread when introduced to locations
outside their natural range?

Despite concerted effort, the field of invasion ecology has been criticized for
its lack of success in answering this question (Ehrlich 1989, Vermeij 1996, Mack
et al. 2000). In this review we hope to show that considerable progress has been
made in identifying the factors underpinning invasion success using historical
data on bird introductions. We discuss what is known about the introduction,
establishment, and spread of introduced birds, focusing on comparative studies
that use the historical record of bird introductions to test hypotheses about the
factors determining invasion success.
Why Study Bird Introductions?

Data on historical bird introductions provide a rare opportunity to test hypotheses
about invasion success for at least three reasons. First, there is an excellent record
of the bird species introduced to locations around the world (Long 1981, Lever
1987). Most bird introductions occurred in the eighteenth and nineteenth centuries
during the major period of European expansion and settlement. Introductions were
often associated with the formation of acclimatization societies that aimed to estab-
lish beneficial or desirable species in the new settlements (see especially Thomson
1922) or with private individuals who sometimes attempted to naturalize a variety
of species (e.g., Eastham Guild on Tahiti; Guild 1938, 1940). Birds were prominent
among the species introduced by settlers, especially for hunting, biocontrol, and
aesthetic reasons. Many societies or individuals kept records of the birds they in-
troduced, sometimes including details such as the numbers of individuals released,
the exact location of release, and the origin of the birds introduced (Thomson 1922,
Long 1981, Lever 1987). Because most of these introductions occurred decades
to centuries ago, and because birds are conspicuous and well studied, introduction
outcomes can be determined with reasonable certainty. We can therefore compile
comprehensive lists of the bird species introduced to new locations, whether or
not those introduced species established wild populations, and the extent to which
the established species have spread.
Second, because many attempts were made to introduce birds, there are data
on a large and taxonomically diverse set of species introduced to a wide range
of locations with which to test hypotheses about invasion success. Worldwide,
there have been recorded more than 1400 attempts to introduce about 400 species
THE ECOLOGY OF BIRD INTRODUCTIONS 73

Figure 1 Schematic representation of stages in the process of a human-caused inva-

sion that we recognize in this review. The stages through which a species must pass
are shown in bold. The names of species that do or do not pass through each stage are
shown in italics.
of birds (Long 1981, Lever 1987, Lockwood 1999, Blackburn & Duncan 2001b,
Cassey 2002).
Third, birds in general are a well-studied group. Thus, we can supplement
data on bird introductions with published information on the phylogeny, life his-
tory, ecology, and natural distribution of these species (e.g., Sibley & Ahlquist
1990; Sibley & Monroe 1990, 1993; Dunning 1993; Bennett & Owens 2002). The
availability of these data is central to testing some key hypotheses about factors
underlying invasion success.
Despite these advantages, there are two difficulties associated with using the
historical record of bird introductions in quantitative studies. First, the historical
record is incomplete, and the quality of the data varies from location to location. In
New Zealand, which has perhaps the most complete and detailed historical record,
at least seven birds were introduced that cannot be identified to species (Thomson
1922) and have therefore been excluded from analyses of introduction outcomes.
The record of failed introductions will invariably be less complete than the record
of successful introductions because species that failed to establish leave no further
trace of their presence. We do not know the extent of this bias in historical data.
Second, the choice of which birds to introduce to which locations was not made at
random but was the result of historical circumstances. The resulting pattern of intro-
ductions is itself informative but raises important issues for statistical analysis and
interpretation. We consider these issues and ways to deal with them in this review.
Definitions and Framework

Terms such as invasive and introduction have been used in different ways in the
literature (e.g., Williamson 1996, Davis & Thomson 2000, Richardson et al. 2000,
Daehler 2001, Kolar & Lodge 2001). In this section, we clarify our terminology
and define stages that we recognize in the invasion process (Figure 1), which serves
as a framework for our review.
We consider only introductions of alien, exotic, or nonindigenous bird species,
all terms used to describe species transported and introduced to a new location by
humans (see also Davis & Thomson 2000, Richardson et al. 2000, Daehler 2001,
Kolar & Lodge 2001). We do not consider natural invasions or range expansions
nor, for the most part, re-introductions of locally extinct species for conservation
purposes. The term invasion here refers specifically to exotic species that have
been deliberately or accidentally transported by humans to a new location beyond
their normal geographic range limits.
We recognize four transitions in the process of invasion by an exotic species
(Figure 1) (see Williamson 1996, Richardson et al. 2000, Davis & Thomson 2000,
Daehler 2001, Kolar & Lodge 2001). First, the exotic species must be transported
from its native geographic range to a new location. Second, the species must be re-
leased or escape into the environment. Third, the species must succeed in establish-
ing a self-sustaining wild population following release. Finally, species that estab-
lish successfully may increase in abundance and spread beyond the release point;
the extent of this spread defines their geographic range in the new environment. We
term these stages transport, introduction, establishment, and spread (see also Lock-
wood 1999, Kolar & Lodge 2001, Sakai et al. 2001). For a species to have reached
a given stage in the invasion process, it must have passed through all prior stages.
An introduced species is one that passes through the first two stages and is
released (or escapes) into a new environment. An established species is one that
establishes a self-sustaining wild population following introduction (also termed
naturalized). An established species that succeeds in spreading beyond the site
of introduction is termed invasive. The introduction of a species to a location
is termed an introduction event. When we use the term introduction, we refer

to a species that has been transported to, and released or escaped into, an alien
environment, regardless of whether it establishes or spreads.

Not all stages in the invasion process have been equally well studied, a fact
reflected in the attention we give them in this review. Most effort has gone into
identifying factors that influence the successful transition from introduction to
establishment. No studies have attempted to quantify the first stage in the process:
which species of birds have been selected for transport to new locations, regardless
of whether they have been introduced or not (but see Guix et al. 1997 for an
analysis of the pet trade of Neotropical parrots in Barcelona). We start, therefore,
by addressing two questions associated with the introduction stage: (a) Which
species were selected for transport followed by introduction to new locations; and
(b) to which locations were birds introduced? We then examine factors associated
with success in establishment and spread.
TRANSPORT WITH INTRODUCTION
Which Species Were Selected for Introduction?

Fewer than 5% of all bird species have been transported and introduced to a new
location (Blackburn & Duncan 2001b). It is important to understand why some
species and not others were chosen for introduction because biases in the type of
species chosen, termed selectivity, may limit the generality of conclusions we draw
regarding the factors affecting success in later stages of the invasion process. If only
certain types of birds were chosen for introduction, then the factors that determine
invasion success in this group may not necessarily apply to other groups of birds.
Selectivity also has important implications for the statistical analysis of success at
later stages in the invasion process (see below). Selectivity in the birds chosen for
introduction has been examined with regard to the taxonomic affiliations of species,
the geographic origin of species, and specific characteristics of the taxa chosen.
TAXONOMIC SELECTIVITY Taxonomic selectivity in global bird introductions has

been examined for all introduced species (Blackburn & Duncan 2001b) and for
the subset of species that were both introduced and established (Lockwood 1999,
Lockwood et al. 2000). The results of both analyses are similar in that families over-
represented at the introduction stage are also the families over-represented among
established species. This fact implies that a major cause of taxonomic selectivity
among established species is taxonomic selectivity in the species chosen for intro-
duction. Two thirds of all species chosen for introduction belong to just 6 of the
145 bird families: Anatidae, Columbidae, Fringillidae, Passeridae, Phasianidae,
and Psittacidae (Blackburn & Duncan 2001b). The over-representation of species
in these families almost certainly reflects two of the major motivations behind
introducing birds. Species in the families Anatidae and Phasianidae were primar-
ily introduced for hunting, while species in the families Fringillidae, Passeridae,
and Psittacidae are frequently kept as cage birds, and many were introduced for
aesthetic reasons (see Long 1981, Lever 1987). Nevertheless, between 69% and
94% of the species in these families have never been the subject of an introduction
event (Blackburn & Duncan 2001b).
GEOGRAPHICAL SELECTIVITY As for many other taxa, bird species richness is

higher nearer the Equator. Nevertheless, most birds chosen for introduction have
their native geographic ranges centered in temperate regions, between 30◦ and
50◦ in both hemispheres (Blackburn & Duncan 2001b). Only around 8% of intro-
duction events involve species from the Neotropics, although the Neotropics are
home to around 30% of all bird species (Rahbek 1997). Conversely, about 16%
of introduction events involve species from the Palaearctic, despite the fact that
only about 10% of bird species have geographic ranges that include this region
(Evans 1997; Sibley & Monroe 1990, 1993). Although there has been no formal
test, species from tropical regions appear to be greatly under-represented in lists
of introduced species.
Much of this geographical selectivity can be attributed to introductions car-
ried out by European settlers, who mostly colonized other temperate parts of the
world. Thus, opportunities for the transport and trade of birds during the eighteenth
and nineteenth centuries would have been greatest between Western Europe and
the predominantly temperate locations where Europeans settled, including North
America, South Africa, Australia, and New Zealand. Around 60% of the bird
species introduced to New Zealand originate from the Palaearctic and Australasian
regions (Duncan et al. 2003), the two places with which trade and transport were
most frequent during early European settlement of New Zealand.
CHARACTER SELECTIVITY These taxonomic and geographical patterns of selec-

tivity imply that the birds chosen for introduction will be nonrandom with respect
to their life history and ecological traits, because these traits are not randomly dis-
tributed across taxa or regions (see, e.g., Bergmann 1847; Lack 1947, 1948; Owens
& Bennett 1995; Gaston & Blackburn 2000; Cardillo 2002). Thus, we would ex-
pect introduced birds predominantly to possess the characteristics of temperate
game birds (Anatidae and Phasianidae) or cage birds (Psittacidae, Fringillidae,
and Passeridae). There have been few tests of character selectivity for introduced
birds, although for those introduced to Australia, a high proportion of species are
ground nesters that use grassland, cultivated, or suburban habitats and have largely
vegetarian diets (seeds, fruit, vegetation) (Newsome & Noble 1986). Introduced
species are also larger-bodied, on average, than would be expected if they were a
random sample of bird species (116.6 g versus 50.5 g for introduced and all bird
species, respectively; Blackburn & Gaston 1994, Cassey 2001a).
Character selectivity also occurs independently of taxonomic and geographical
selectivity. For both a geographic (species in the British avifauna) and a taxonomic
(order Anseriformes; wildfowl) subset of the world’s birds, species with larger
population sizes were significantly more likely to have been chosen for introduction
(Blackburn & Duncan 2001b). In addition, larger-bodied wildfowl and resident
British birds that were widely distributed had a higher probability of selection.
Species with large population size, wide distribution, and resident status would
have been the most readily available for capture and transport to new locations.
While most species were released at very few locations, a few species, typically
widely distributed and abundant in their native ranges, were widely introduced
(Long 1981, Lever 1987, Cassey 2002).
Together, these results suggest a general hierarchy of causes that have con-
tributed to the selective introduction of certain birds. First, the species chosen
for introduction were concentrated in geographic regions that in large part re-
flect the origin of European settlers and their subsequent patterns of settlement
and trade. Second, from within geographic regions, people chose certain kinds of
bird for introduction for a variety of reasons. An emphasis on birds for hunting
and aesthetic purposes has resulted in birds from five families being significantly
over-represented among those chosen. Finally, given that birds were chosen for
particular purposes and from certain regions, the species that were finally caught,
transported, and introduced tend to be those that were common in the source lo-
cations. People preferentially selected abundant, widely distributed species, either
because they were most readily available for capture or because they were species
with which they were most familiar, and therefore most desired to introduce.
To Which Locations Were Birds Introduced?

People mostly introduced birds to islands. Although islands make up about 3% of
ice-free land area (Mielke 1989), around 70% of introduction events have been to
islands (Blackburn & Duncan 2001b). Slightly over one half of all introductions
were to Pacific islands (especially the Hawaiian islands) and Australasia. The
greatest number of mainland introductions was to the Nearctic, followed by the
Palaearctic, and continental Australia. Birds have been introduced to all major
regions of the world and to most latitudes with ice-free land. However, relatively
few introductions have been to equatorial regions; most were to latitudes between
10◦ and 40◦ on both sides of the equator (Blackburn & Duncan 2001b).
ESTABLISHMENT
A species introduced to a new location will either succeed or fail to establish a self-
sustaining population. There are numerous reasons why introductions could fail,
making it unlikely that we will ever be able to predict with certainty the outcome
of any one introduction event (Ehrlich 1989, Gilpin 1990). Nevertheless, a major
goal of invasion ecology is to identify factors that have some ability to explain the
outcome of introduction events, to rank factors in terms of their importance, and to
determine whether different factors are important under different circumstances,
and why (Williamson 1999).
Factors hypothesized to influence establishment can be grouped into three cat-
egories: (a) characteristics of the species introduced, such as their population
growth rate or migratory strategy; (b) features of the introduction location, such as
its environment or insularity; and (c) factors associated with, and often unique to,
each introduction event, such as the number of individuals released. We term these
species-, location-, and event-level factors, respectively (Blackburn & Duncan

2001a; see Williamson 1999, 2001).
Two crucial consequences of the selectivity in historical introductions (see
above) are that factors at different levels are likely to be confounded, and that
introduction events are unlikely to represent statistically independent data points
when analyses include introduction data from more than one location. We first
consider these issues because they are central to how we analyze and interpret
comparative data on establishment success in any taxonomic group.
Statistical Analysis of Historical Introduction Outcomes

Species-, location-, and event-level factors are likely to be confounded because
selectivity means that introductions are nonrandom with regard to the taxonomy
and characteristics of the species introduced and their locations of origin and in-
troduction. For example, species that are more common in Britain were introduced
in greater numbers to New Zealand (Blackburn & Duncan 2001b). Species that
are common in Britain also possess traits that distinguish them from less common
species (e.g., smaller body mass and typical niche position; Nee et al. 1991, Gregory
& Gaston 2000). Hence, for introductions of British birds to New Zealand, fac-
tors at the species-level (certain life-history traits) and event-level (the number of
individuals introduced) are confounded. The solution in this case is to identify
confounding factors and to control for them statistically using techniques such as
multiple regression.
The second issue is that, when we consider data from several introduction
locations, each introduction event is unlikely to represent a statistically independent
data point (Blackburn & Duncan 2001a,b). This is because species were often
introduced to multiple locations, and multiple species were introduced to each
location. If some species are better at establishing than others, or if it is easier to
establish at some locations than others, then the outcome of introductions of the
same species, or of introductions to the same location, will be correlated. In these
circumstances, the introduction of 5 species to each of 10 different locations does
not provide 50 independent pieces of evidence to assess how species- or location-
level factors affect establishment (see also Hurlbert 1984, McArdle 1996).
Introduction events are unlikely to be independent if they can be clustered by
both species and location. Furthermore, species and locations may themselves
cluster into higher-level units. In particular, closely related species are more likely
to share life-history, morphological, and behavioral traits than are more distantly
related species. If these shared traits affect the likelihood of establishment, then
introduction outcomes will be clustered by phylogenetic or taxonomic relatedness.
Clustering, leading to correlated responses in groups of observations, typically
violates a core assumption of standard statistical models: that the error terms in the
model are independent (i.e., uncorrelated). This leads to standard error estimates
that are smaller than the true values, resulting in overestimates of the significance
of factors included in the model (a greater frequency of Type I errors). In plain

terms, we are more likely to obtain models containing spuriously significant factors
if we fail correctly to account for the clustering of observations in our data.

There are at least two approaches to dealing with this problem when analyzing
historical introduction data. First, it is less of an issue if we consider introductions
of multiple species to a single location (for example, the outcome of bird introduc-
tions to Tahiti), because these data are not clustered by species and location—we
have only one event per species at a single location (likewise for introductions
of a single species to multiple locations). While this tactic largely overcomes the
problem, it limits the questions we can address, and we still have to consider non-
independence owing to phylogenetic relatedness. Alternatively, there has been re-
newed interest in developing statistical models that deal directly with the problem
of nonindependence in clustered data. These approaches include the use of gen-
eralized least squares, linear mixed models, generalized linear mixed models and
generalized estimating equations (Diggle et al. 1994, Goldstein 1995). These ap-
proaches are appropriate for analyzing spatially, temporally, and phylogenetically
clustered data that are regularly encountered in ecological studies. Generalized lin-
ear mixed models have been applied to a global analysis of establishment success
in birds, to account for the clustering of introduction events by species, taxonomic
groups, and geographic locations (Blackburn & Duncan 2001a). Clearly, prob-
lems of clustering and nonindependence in introduction data are not specific to
birds, and comparative studies of invasion success in any taxonomic group should
carefully consider these issues.
Event-Level Influences on Establishment Success

INTRODUCTION EFFORT A feature of introduced species is that they are typically
released in low numbers and so start with small founding populations. Smaller
populations are at greater risk of extinction from stochastic fluctuations owing to
demographic, environmental, or genetic stochasticity (Pimm 1991). In addition,
small populations may suffer from Allee effects (for example, an inability to find
mates at low population densities) and are more prone to extinction from natural
catastrophes (e.g., several populations of introduced birds appear to have died out
as a result of hurricanes; see Long 1981). Given that they start out with small
founding populations, it is not surprising that most introductions fail. For birds,
744 of 1466 (51%) introduction events in the global data set analyzed by Blackburn
& Duncan (2001a) did not result in establishment, and this success rate is high
relative to other taxa (Williamson 1996, Williamson & Fitter 1996).
A straightforward explanation for the success of some introductions is that they
involved the release of a greater number of individuals and so were more likely
to escape the threats facing small founding populations. That is, establishment
success increases with greater introduction effort, or propagule pressure. Because
information on release sizes has been recorded in many cases, it is possible to use
historical bird introductions to test this hypothesis. All studies that have addressed
this question so far have shown that species introduced with greater effort (typically
measured as the total number of individuals released) have a higher probability
of establishment [Dawson 1984 (cited in Williamson 1996); Newsome & Noble

1986; Griffith et al. 1989; Pimm 1991; Veltman et al. 1996; Duncan 1997; Green
1997; Sol & Lefebvre 2000; Duncan et al. 2001, 2003; Cassey 2001b; Duncan
& Blackburn 2002], although all but two of these studies use data from just two lo-
cations, New Zealand and Australia. Nevertheless, observational and experimental
evidence in other animals has also shown that the size of the founder population
increases the chance of establishment (e.g., Hopper & Roush 1993, Berggren
2001, Forsyth & Duncan 2001). These results are consistent with the hypothesis
that small populations are less likely to establish owing to stochastic events and
that these play a major role in invasions (Green 1997). Nevertheless, there are
exceptions. Despite the introduction of over half a million common quail, Co-
turnix coturnix, to more than 30 states in North America between 1875 and 1958
(Bump 1970), no introduction succeeded (Lever 1987). Clearly, factors other than
propagule pressure can be critical in determining the fate of introductions (see
below).
The role of demographic stochasticity in avian introductions has been explicitly
modeled by Legendre et al. (1999). Because demographic stochasticity can gen-
erate random fluctuations in sex ratio leading to difficulty in finding a mate, they
constructed a two-sex model with an explicit mating system and calculated result-
ing extinction probabilities via Monte Carlo simulations. The results showed that
demographic uncertainty imposes high extinction risk on monogamous as com-
pared to polygynous birds, and on short-lived as compared to long-lived birds. The
model fit reasonably well to the extinction probabilities calculated for passerine
birds introduced to New Zealand.
Nevertheless, stochasticity is not the only mechanism that could explain the link
between introduction effort and establishment success. A greater number of release
attempts may improve establishment by increasing the chance that at least one
release population encounters favorable conditions for population growth (Crawley
1986). Griffith et al. (1989) found that introductions involving releases over a
greater number of years had a greater probability of establishing, independently of
the number of individuals released, although Green (1997) failed to find a similar
relationship for New Zealand bird introductions. In a more direct test, Veltman et al.
(1996) reported that species successfully introduced to New Zealand were released
at more localities than unsuccessful species, although the number of localities and
the total number of individuals released are correlated in these data.
Green (1997) also suggested that the link between introduction effort and estab-
lishment success could arise indirectly if, for example, people directed more effort
toward releasing species that seemed pre-adapted to the conditions they would
encounter at their location of introduction. Similarly, introduction effort could be
related to establishment success indirectly through relationships with abundance
in the source location and associated species traits. Nevertheless, studies that have
attempted to control for these indirect relationships by including species traits, en-
vironmental tolerances, and overseas range size in multiple regression models have
consistently found that introduction effort remains by far the strongest independent
explanation for establishment success in birds (e.g., Duncan et al. 2001).
CLIMATE/ENVIRONMENTAL MATCHING One of the most frequently stated hypothe-

ses in the biological invasion literature is that species should have a better chance
of establishing if the climate and physical environment at the location of in-
troduction and in the species’ natural range are closely matched (Brown 1989,
Williamson 1996). Data from bird introductions provide some of the very few
quantitative tests of the hypothesis that a close climate/environmental match be-
tween a target and source location enhances establishment success. We include
climate/environmental matching under event-level effects because this is a feature
unique to each introduction event.
Avian establishment success is significantly greater when the difference be-
tween a species’ latitude of origin and its latitude of introduction is small
(Blackburn & Duncan 2001a, Cassey 2001b), when climatic conditions in the
locations of origin and introduction are more similar (Duncan et al. 2001), and
when species are introduced to locations within their native biogeographical re-
gion (Blackburn & Duncan 2001a, Cassey 2003). Likewise, for re-introductions,
Wolf et al. (1998) found that species re-introduced in the core of their original
range had greater establishment success than species released in the periphery.
Regions at similar latitudes or within the same biogeographic region are likely to
be similar in climatic and habitat conditions. Hence, these results support the hy-
pothesis that introduction success is enhanced if species are matched with suitable
environments (Brown 1989).
Also related to the idea that climate/environmental matching is important is the
finding in some studies that bird species with larger geographic ranges are more
likely to establish following introduction (Moulton & Pimm 1986, Blackburn &
Duncan 2001a, Duncan et al. 2001). Species may have large geographic ranges be-
cause they can exploit a broad range of conditions (they have large niche breadth),
or use conditions that are themselves widespread (they have a typical niche posi-
tion; Gaston 1994). In a similar vein, it has been suggested that being native to a
relatively variable abiotic environment could also enhance establishment success
(Ehrlich 1989), because species that survive in such environments can generally
cope with a wide range of environmental conditions (see also Stevens 1989). Any
of these situations would increase the chance that a species introduced to a new
location would encounter conditions favorable to its survival, although only the
range size effect has received empirical support.
Location-Level Influences on Establishment Success

Three characteristics of a location are likely to influence establishment: enemies,
resources, and the physical environment (Shea & Chesson 2002). Establishment
should be favored at locations where there are fewer enemies (predators, parasites,
or diseases), more available resources (owing to higher resource availability or an
absence of competitors), or a more benign environment. It is difficult to identify
precisely which enemies, resources, or aspects of the physical environment influ-
ence establishment, so studies of location-level effects often consider features of
the environment that ought to be correlated with these characteristics. For exam-
ple, species richness is commonly used as a surrogate for the number of enemies,
competitors or vacant niches likely to be present at a location (Shea & Chesson
2002). Species richness itself tends to co-vary with latitude and whether or not the
location is an island (Elton 1958).
In fact, there is little evidence from bird introductions to support the hypothesis
(Elton 1958) that species-poor locations are easier to invade than species-rich lo-
cations. Case (1996) showed that the number of established bird species does not
vary significantly with the richness of the native avifauna nor the variety of poten-
tial mammalian predators. Whereas New Zealand has fewer native but a greater
number of exotic bird species than Australia, there is no difference in establish-
ment probability for birds introduced to these locations (Sol 2000a). Likewise, a
global analysis of bird introductions failed to find a relationship between estab-
lishment success and latitude of introduction, implying no significant effect of
resident species richness on establishment (Blackburn & Duncan 2001a).
Islands are often viewed as more vulnerable to invasion because of lower species
richness, but also owing to factors such as an absence of certain functional groups,
reduced competitive ability of island species, intensive human exploitation, or re-
duced habitat diversity (e.g., Vitousek 1988, Dalmazzone 2000). In support of this
notion, Newsome & Noble (1986) reported a higher failure rate for bird species
introduced to mainland Australia compared with bird species introduced to Aus-
tralia’s offshore islands. However, this result could be an artifact of differences in
invasive ability between the species introduced to mainlands and islands. A species-
by-species examination of introduced birds in two independent island-mainland
comparisons (New Zealand versus Australia, and Hawaiian Islands versus U.S.
mainland) found no evidence that islands were easier to invade (Sol 2000a). This
result has been generalized in global analyses of bird introductions that failed to
find a relationship between establishment success and whether the introduction
was to a mainland or island location (Blackburn & Duncan 2001a, Cassey 2003).
The high proportions of exotic bird species found on islands appears primarily to
be a consequence of the many attempts to introduce birds to islands (see above
To Which Locations Were Birds Introduced?) rather than any inherent feature of
islands that make them easy to invade (Sol 2000a, Blackburn & Duncan 2001a).
Correlative studies testing for a relationship between species richness and in-
vasion success have been criticized because variation in species richness may be
confounded with variation in other factors that affect establishment, making it

difficult to isolate species richness effects (Shea & Chesson 2002). For birds, the
studies described above suggest that factors other than resident species richness are
overwhelmingly important in determining the outcome of introductions, even after
statistically controlling for confounding factors. This lack of an effect is probably
real. Competition between native and introduced birds does not appear to regulate
establishment because most exotic birds establish in habitats highly modified by
humans, such as farmland and urban areas, which are little used by native species
(Diamond & Veitch 1981, Simberloff 1992, Smallwood 1994, Case 1996). Such
modified habitats may favor pre-adapted invaders through release of resources
and enemy reduction (Mack et al. 2000, Shea & Chesson 2002). Thus, many in-
troduced birds may succeed because they exploit either vacant niches created by
human activities or existing niches that humans have expanded.
Some authors have suggested that establishment is more likely for species
that can exploit niche opportunities left by extinct or declining native species
(Herbold & Moyle 1986, Case 1996, Mack et al. 2000). This is less likely if native
and exotic species show strong habitat segregation. Nevertheless, Cassey (2001b)
showed that, whereas extinct bird species in New Zealand tend to be larger-bodied
than extant species, they do not differ in size from established exotic species. He
suggested that some established invaders could be occupying the niches of extinct
species. Case (1996) likewise reported a positive relationship between the numbers
of established and recently extinct bird species at locations around the world. This
relationship does not arise because introduced birds cause the extinction of native
species: Most extinctions occurred prior to bird introductions. More likely, high
extinction rates are associated with high levels of human disturbance, which in turn
creates habitat favorable for the establishment of introduced birds (Case 1996).
Habitat segregation suggests that competition between native and exotic species
has little role in determining whether or not introduced birds establish. However,
it leaves open the question as to whether interactions with native species can
prevent introduced species from spreading into more pristine habitats. Diamond
& Veitch (1981) reported that, in New Zealand, exotic birds that are abundant on
the main islands are virtually absent from unmodified forests on offshore islands,
despite having ample opportunities to colonize these. There is no evidence that this
segregation results from competitive interactions. More likely, introduced species
establish in habitats to which they are pre-adapted but to which many native species
are not, and vice versa (Case 1996, Sax & Brown 2000).
Whereas competition between exotic and native bird species does not appear
to influence establishment, it is possible that competition between the introduced
species may be important at locations with many introductions. In a series of
studies, Moulton and coworkers (e.g., Moulton & Pimm 1983; Moulton 1985,
1993; Lockwood et al. 1993; Lockwood & Moulton 1994; Brooke et al. 1995;
Moulton & Sanderson 1999; Moulton et al. 2001b) have used two lines of evidence
to argue that this is the case. First, for passerine introductions to Hawaii and
Saint Helena, the probability that a species will establish declines through time
as the number of established species accumulates. Moulton and coworkers argue

that it is harder for progressively later introductions to establish because they
face increasing competition from a greater number of already established species.
Second, at several locations successfully established species overlap less in their
morphological characteristics than would be expected if a random selection of
the introduced species had established. Moulton and coworkers argue that this
pattern arises because species are more likely to fail due to competition with other
introduced species of similar morphology, leading to a pattern of morphological
overdispersion among the established species.

However, these conclusions have been questioned by studies showing that the
same patterns could arise through various artifacts (Simberloff & Boecklen 1991,
Duncan 1997, Duncan & Blackburn 2002). Whereas later introductions of passer-
ine birds to New Zealand faced a greater number of already established species
and were less likely to succeed, this relationship was confounded by introduction
effort (Duncan 1997). Later introductions tended to be of fewer individuals and
so were less likely to succeed for that reason. Similarly, Duncan & Blackburn
(2002) showed that significant morphological overdispersion among gamebirds
established in New Zealand (Moulton et al. 2001b) could not have resulted from
competition, but could be explained by greater effort being applied to the introduc-
tion of species that were morphologically different from one another. In addition
to facing a greater diversity of already established birds, later introductions to New
Zealand, and probably elsewhere, would also have faced a greater diversity and
abundance of introduced predators, making it impossible to isolate the effect of
competition (Duncan et al. 2003). These results for New Zealand do not rule out
the importance of inter-specific competition in other introduced bird assemblages,
but they do question the evidence presented in support of this mechanism. Overall,
there is little unequivocal support for the hypothesis that competition affects the
outcome of bird introductions.
The idea that establishment is affected by competitive and predatory interac-
tions has suggested at least three further hypotheses. First, social foraging species
should be better invaders than solitary species (Mayr 1965, Ehrlich 1989), because
social foraging can increase the probability of detecting predators, locating food,
and learning about new food sources. This hypothesis has not received empirical
support (Sol 2000b, Duncan et al. 2001). Second, ground-nesting birds should have
lower probabilities of establishing than canopy-, shrub-, or hole-nesters, because
nest predation is generally higher in ground-nesters (see Reed 1999). There is some
support for this hypothesis (Newsome & Noble 1986; McLain et al. 1999; but see
Sol et al. 2002). Third, herbivores have been predicted to invade new habitats more
easily than carnivores (Hairston et al. 1960, Crawley 1986), because competition
is thought to be less intense among herbivores than among carnivores. The hy-
pothesis has not received empirical support for birds (e.g., Veltman et al. 1996),
and its assumptions are questionable (Crawley 1986).
A further location-level factor proposed to influence establishment is the area
of suitable habitat (Smallwood 1994, Case 1996). Larger areas can support more
individuals and are thus more likely to maintain self-sustaining populations in the
long-term. However, this fact may be of little relevance to establishment when pop-
ulations generally start out small. Indeed larger areas may result in more dispersed
founding populations and a lower probability of success owing, for example, to
greater difficulty finding mates. Case (1996) found a weak, but significant, posi-
tive correlation between island area and the number of species introduced, but no
correlation with establishment success. Site invasibility was not found to decrease
with area in California’s nature reserves (Smallwood 1994).
Finally, in a global analysis of bird introductions, Blackburn & Duncan (2001a)

found significant variation in establishment rate among biogeographic regions after
having controlled for other confounding factors such as latitude. It is unclear what
underlies such large-scale geographic variation in establishment probability.
Species-Level Influences on Establishment Success

The widespread success of certain introduced birds, such as the starling (Sturnus
vulgaris), house sparrow (Passer domesticus), and rock dove (Columba livia; Long
1981), suggests that some species may simply be good at establishing in novel
environments. Whether species differ inherently in their probability of establishing,
and if so, what characteristics regulate this difference, are two questions that have
long interested ecologists (Mayr 1965, Ehrlich 1989).
DO SPECIES DIFFER IN THEIR PROBABILITY OF ESTABLISHING? Simberloff &

Boecklen (1991) noted what they termed an all-or-none (AON) pattern for bird
introductions to the Hawaiian islands. Bird species tended to establish success-
fully or to fail repeatedly on all islands to which they were introduced, with mixed
outcomes, where species established on some but not other islands, being rare (but
see Moulton 1993; Brooke et al. 1995; Moulton & Sanderson 1996, 1999; Duncan
1997; Duncan & Young 1999). Such a pattern would suggest that species-level
attributes are a key determinant of establishment. Some species are particularly
good invaders, and so succeed everywhere they are introduced, while other species
are poor invaders and fail regardless of the location.
A brief perusal of any compendium of avian introductions (e.g., Long 1981,
Lever 1987) is enough to show that birds do not follow a strict AON pattern. This is
expected even if species-level attributes strongly influence establishment success
(Duncan & Young 1999). Species are unlikely to be either absolutely good or poor
invaders because event- and location-level factors will also affect establishment.
Hence, even an exceptionally good invader might occasionally fail because of
unfavorable circumstances, and vice versa. The key question is whether there is
significant variation in establishment probability among introduced species, the
alternative being that all species establish with equal probability. In testing this
hypothesis, we need to control for confounding factors—an inherently poor invader
might appear better than it really is if introduced to locations that are easy to invade.
Fitting a mixed model to data on global bird introductions, and having controlled
for location- and event-level factors likely to affect establishment, Blackburn &
Duncan (2001a) showed that there are highly significant differences among species
in establishment probability. Given this variation, do certain attributes of the species
themselves distinguish the good from the poor invaders?
WHICH TRAITS INFLUENCE ESTABLISHMENT SUCCESS? The idea that certain spe-
cies attributes influence establishment success is supported by comparative studies
linking the outcome of historical bird introductions to behavioral, life history, and
morphological traits (e.g., Newsome & Noble 1986, McLain et al. 1995, Veltman
et al. 1996, Green 1997, Cassey 2001a, Sol 2003). However, two problems make
it difficult to draw firm general conclusions from the studies to date. First, some
studies present results from univariate statistical models while other studies present
results from multivariate models. Multivariate models are more powerful in cor-
relative studies because they reduce the likelihood that a significant relationship
between two variables results from their correlation with a third variable, assuming
that important confounding variables have been included in the analysis. Second,
some studies suffer from the problems of non-independence we discussed above.
In particular, the importance of traits in explaining establishment is often assessed
by fitting a generalized linear model with establishment included as either a binary
outcome (the result of each introduction event coded as success or failure) or a
binomial outcome (the number of successful events / the total number of intro-
duction events for each species). In either model, a species is represented in the
analysis as many times as it has been introduced. However, when the data include
multiple introductions of the same species to different locations, then each event
will not represent an independent data point because the outcomes of introductions
of the same species to different locations are almost certainly correlated. In these
circumstances, it is important to consider extensions of generalized linear models
(such as generalized linear mixed models or generalized estimating equations) that
take into account the clustering of events and minimize the problem of inflated
type I error rates.
The traits identified as influencing establishment success are of three types:
(a) traits that preadapt species to the new environment, (b) traits that favor popula-
tion increase from a low level, and (c) traits that constrain establishment success.
Preadaptations to establishment One reason for failing to establish is that a

species fails to find a suitable niche in the new environment, either because the
niche does not exist or because it is occupied by other species. The chance of
finding a favorable niche will depend on the presence of preadaptations to exploit
the resources and escape enemies in the new environment.
A generalist species with a wide niche breadth should have, on average, a better
chance of finding appropriate resources and environmental conditions wherever
it is introduced than a more specialized species (Ehrlich 1989, Forys & Allen
1999). A problem with testing this hypothesis is the difficulty in quantifying niche
breadth (Gaston 1994). Using a coarse measure of dietary breadth, McLain et al.
(1999) found that, for 132 passerine species introduced to nine islands, those with
a broader dietary range were more likely to establish (see also Wolf et al. 1998,
Cassey 2001b). Similarly, Brooks (2001) found that introduced birds categorized
as habitat specialists were less likely to establish successfully. However, other
studies have failed to find such relationships (Veltman et al. 1996, Sol et al. 2002),
and it is difficult to determine if these discrepancies result from a weak influence of
niche breadth or reflect the noise associated with using coarse measures to quantify
breadth.
A species released into a new environment will face a variety of novel chal-
lenges for which it may not be well adapted (Sol 2003). Animals may partly
compensate for this lack of adaptive fit by inventing new behaviors or adjusting
established behaviors to the novel conditions (Klopfer 1962, Plotkin & Odling-
Smee 1979, Morse 1980, Arcese et al. 1997, Brooker et al. 1998, Berger et al. 2001,
Lefebvre et al. 2003, Sol 2003). This behavioral flexibility may aid establishment
through, for example, the ready adoption of unexploited new food resources, the
adjustment of breeding to the prevailing breeding conditions, or rapid behavioral
changes to avoid novel enemies (Sol 2003). Because behaviorally flexible species
are believed to be more exploratory (Greenberg & Mettke-Hofmann 2001) and
ecologically generalist (Sol 2003), they may also have higher chances of discov-
ering and adopting new habitats or new resources that may enhance survival and
reproduction in the new environment. The hypothesis that behavioral flexibility
enhances establishment (Mayr 1965) is supported by the finding that established
birds tend to have a larger brain size per unit body mass and to show more innova-
tive behaviors in their region of origin than failed species (Sol & Lefebvre 2000,
Sol et al. 2002).
Certain traits associated with generalist behavior may enhance establishment
for some species, whereas other traits may preadapt species to specific habitats.
The latter may explain why birds that inhabit human-modified habitats in their
native range and those with a history of close association with humans tend to
be successful invaders (Sol et al. 2002; see also Mayr 1965, Brown 1989, Sax &
Brown 2000).
Securing the bridgehead Most introduced populations start off small and face the
threat of stochastic extinction. One argument is that species with higher rates of
population growth should have a higher probability of establishing because they
can more quickly escape the risks associated with remaining at small population
size (Moulton & Pimm 1986, Pimm et al. 1988, Pimm 1991). Because population
growth rates are difficult to measure, most studies that test this hypothesis use life-
history traits known to be correlated with population growth rate as surrogates.
Thus, species with small body mass, short development times, multiple broods per
season, and large clutch sizes are expected to have higher establishment. However,
small-bodied species with high rates of population growth also tend to have more
variable population sizes, which could increase their risk of extinction because
population densities will drop to low levels more often (Pimm 1991). In these
circumstances, larger-bodied, longer-lived, species with slower rates of population

growth that are relatively unaffected by environmental fluctuations are expected
to have higher establishment (Legendre et al. 1999, Forsyth & Duncan 2001).
These contradictory theoretical predictions are mirrored in empirical findings.
Whereas some studies report a positive relationship between clutch size and es-
tablishment success (Green 1997, Cassey 2001b), others have reported negative
relationships or no relationship at all (Veltman et al. 1996, Duncan et al. 2001).
Cassey (2001b) found that birds with longer generation times were more likely to
establish following introduction to New Zealand, but Griffith et al. (1989) reported
higher establishment success for re-introduced species classified as “early breed-
ers” than for those considered “late breeders.” Results for body mass are similarly
equivocal (e.g., Veltman et al. 1996, Green 1997, Sol & Lefebvre 2000, Blackburn
& Duncan 2001b, Sol et al. 2002). Moreover, the relationship between body mass
and establishment appears to depend on the taxonomic level considered. Cassey
(2001a) found that global introduction success is significantly negatively related
to body size across species, families, and higher family nodes. However, within
taxa, larger-bodied species are more likely to establish.
Population growth rate has also been invoked in the suggestion that colonial
species should establish less readily than noncolonial. Colonial species may have,
at low population densities, a reduced per capita growth rate owing to the Allee
effect, which would increase the probability of stochastic extinction (Reed 1999).
The only study that has examined this prediction failed to find support for it (Sol
2000b).
Constraints on success Some traits may lower establishment probability regard-

less of other traits the species possess. The two most commonly cited such traits
are migratory behavior and sexual color dimorphism. In New Zealand, migratory
species are less likely to establish than nonmigratory species (Veltman et al. 1996,
Duncan et al. 2003; but see Duncan et al. 2001, Sol et al. 2002). The reasons for
this result are unclear, but they could be related to physiological costs associated
with preparing for migration, or because suitable habitats to migrate to are not
available or cannot be found from the new location (Veltman et al. 1996). In birds,
sexual dimorphism in plumage color appears to be the result of sexual selection,
and in some studies sexually dimorphic species are less likely to establish follow-
ing introduction than sexually monomorphic species (McLain et al. 1995, 1999;
Sorci et al. 1998; Sol et al. 2002; but see Sol & Lefebvre 2000; Duncan et al. 2001,
2003). The reasons for this may be associated with the cost to males of producing
and maintaining secondary sexual characters (A.P. Møller, M.C. Gontard-Danek,
unpublished manuscript).
TAXONOMIC VARIATION IN ESTABLISHMENT PROBABILITY Given that species dif-

fer significantly in establishment probability, how is this variation partitioned
among different levels in the taxonomic hierarchy? If significant variation resides
at high levels in the hierarchy, say among families, then it should be possible to
predict establishment from knowledge of a species’ taxonomic status alone (Daehler

& Strong 1993). In addition, knowing which families, or other higher taxa, have
a high establishment rate may suggest traits shared by members of those groups
that contribute to their success.
Applying methods such as nested ANOVA and mixed models to bird introduc-
tion data reveals that most variation in establishment probability resides at low
rather than high taxonomic levels (Blackburn & Duncan 2001a, Sol et al. 2002).
That is, closely related species are likely to differ substantially in their probability
of establishment. Nevertheless, the extent to which at least some variation in es-

tablishment resides at higher taxonomic levels remains unclear. Sol et al. (2002)
reported that 21% of variation in establishment rate was found at order level. Hav-
ing controlled for variation in location-level effects, Blackburn & Duncan (2001a)
found no similar significant effect in their global analysis of bird introductions. In
Australia, introduced gamebirds (order Galliformes) were significantly less likely
to establish than other species (Duncan et al. 2001), with similar low rates of estab-
lishment observed in New Zealand and the United States. This observation could,
in part, reflect hunting pressure, although in the United States long moratorium
periods were implemented before the hunting of gamebird species commenced
(Long 1981). Moulton et al. (2001a) also reported lower establishment success in
Galliformes and Columbiformes relative to Passeriformes.
The general pattern of variance partitioning suggests that interspecific variation
in establishment probability is driven primarily by traits that vary among closely
related species. It follows that traits shared by closely related species, including
phylogenetically conserved life history traits such as body mass and clutch size,
should poorly explain establishment. The equivocal results reported for many such
traits, along with their failure significantly to explain establishment for global bird
introductions (Blackburn & Duncan 2001a), bear this out. The general failure to
identify life history traits that consistently explain establishment may reflect the
overwhelming importance of factors such as introduction effort or environmental
matching. Equally, however, there may be no simple set of traits that favor es-
tablishment. Instead, different traits may enhance establishment under different
conditions, and at any one location there may be establishment opportunities for
species possessing a variety of trait combinations.
SPREAD
The European starling (Sturnus vulgaris) established following introduction to
New York state in the late nineteenth century and has subsequently spread to
become among the most widely distributed and abundant birds in North America.
At about the same time, the crested myna (Acridotheres cristatellus) was introduced
and became established in Vancouver. This close relative of the starling became
reasonably numerous but spread little, remaining largely confined to Vancouver and
its environs, with the population now having dwindled to low numbers. Relative to
other taxa, birds are generally good dispersers. Why then, following establishment,
do bird species differ so markedly in the rate at which they spread and the final
extent of their distribution?
Because birds are well studied, sufficient records exist quantitatively to model
range expansion in several introduced species (Hengeveld 1989, Van den Bosch
et al. 1992, Veit & Lewis 1996, Williamson 1996, Shigesada & Kawasaki 1997,
Lensink 1998, Gammon & Maurer 2002, Silva et al. 2002). Whereas these models
allow us to examine factors affecting the rate of expansion in a few typically
widespread species, no study has yet quantified or examined the reasons underlying
interspecific variation in rate of spread.
Three studies have nevertheless examined the determinants of current geo-

graphic range size in two introduced avifaunas: New Zealand (Duncan et al. 1999)
and Australia (Duncan et al. 2001, Williamson 2001). The extent of habitat or envi-
ronmental conditions suitable for an introduced species should ultimately constrain
its distribution in a new location. This hypothesis is supported by data from both
avifaunas. In New Zealand, the species with the largest geographic ranges were
those whose preferred habitat was most widespread, specifically species that use
extensive human-modified habitats such as farmland. On the larger and more cli-
matically varied continent of Australia, species with larger geographic ranges were
those with a greater area of more climatically suitable habitat available, with this
variable explaining 69% of the variation in range sizes.
All three studies show that species with a larger native geographic range size
achieved a larger range size in the location of introduction (Duncan et al. 1999,
2001; Williamson 2001). At least two explanations could underlie this relationship.
First, as we have discussed, species with larger native ranges may have broader
environmental tolerances or use more widespread resources, enhancing success
at both establishment and spread. We suspect this is the primary cause of this
relationship. The second explanation follows from the finding that species with
larger native range sizes tend to be introduced more often and in greater num-
bers (Blackburn & Duncan 2001b). In New Zealand, the species introduced with
greater effort were not only more likely to establish but were also more likely to
spread and achieve a larger geographic range size (Duncan et al. 1999). One ex-
planation for this finding is that species with large founding populations might be
able to capture a greater proportion of shared resources from species with smaller
founding populations. This initial advantage could have compounded itself—those
species initially able to capture a greater share of resources would have had faster
population growth and rate of spread, allowing them further to pre-empt resources
at newly colonized sites as their ranges expanded. As we would expect, this effect
is most pronounced among closely related species that are more likely to compete
for similar resources (Duncan et al. 1999). This explanation implies that compe-
tition may play a role in limiting the range sizes of established species but not
in influencing establishment success (see above). Nevertheless, it requires further
testing even to establish if the relationship between introduction effort and range
size holds more generally (it does not for Australia; Duncan et al. 2001).
Finally, several life history traits consistently relate to the geographic range
size of established species in New Zealand and Australia (Duncan et al. 1999,
2001; Cassey 2001b; but see Williamson 2001). More widespread species tend
to be those with life history traits associated with high population growth rates,
characteristically small-bodied, rapidly developing species with high fecundity.
It has been suggested that species with fast population growth rates have larger
ranges because they may be less vulnerable to local extinction when colonizing
unoccupied sites (Gaston 1988).
FUTURE DIRECTIONS
Throughout this review, we have emphasized how the quantity and quality of data
on historical bird introductions have provided opportunities to test key hypotheses
about the factors underpinning invasion success. We finish by highlighting five
promising research directions.
First, attempts to analyze data on bird introductions have highlighted issues
that are likely to be a feature of all historical introduction data: confounding of
explanatory variables and non-independence of observations. Given that historical
introduction data provide us with opportunities to test hypotheses that we cannot
test experimentally (because of cost or ethical considerations), a priority is to
identify and apply appropriate statistical methods to analyzing such data for all
taxonomic groups. Undoubtedly, the degree and type of non-independence will
depend on the circumstances surrounding the transport and introduction of different
taxa. Birds and plants selected for transport and introduction, for example, probably
show different patterns of taxonomic clustering. Analyses that do not consider
issues of non-independence should be interpreted with caution.
Second, while comparative studies of bird introductions have provided useful
tests of several invasion hypotheses, others remain for which bird introductions
appear ideally suited. For example, the “enemy release” hypothesis posits that some
invaders do better in their location of introduction than in their native range owing
to a lack of natural enemies (e.g., competitors, predators, and pathogens). Because
birds are well studied, a considerable amount of detailed information exists on
the factors regulating the populations of many bird species in their native ranges.
Similarly detailed studies identifying the factors regulating populations of the same
species that have established at new locations could test, for example, whether
release from predation or competition is the explanation underlying enhanced
success in the new location.
Third, measuring and explaining the impact of invaders remains a major un-
resolved issue in invasion biology (Ricciardi et al. 2000). The impacts of bird
invaders have generally received less attention than those associated with other
taxa (see, e.g., Ebenhard 1988). One possible reason is the perception that, be-
cause most avian invaders occur in human-modified habitats rather than in pristine
habitats (Case 1996), their ecological impact should be relatively less important
(e.g., Diamond & Veitch 1981). However, birds have also been reported to gener-
ate serious ecological impacts on recipient communities, including hybridization
and introgression with native species (Rhymer & Simberloff 1996), transmission
of diseases (van Riper et al. 1986), competition and predation (Penny 1974; but
see Koenig 2003), and habitat alteration (Ebenhard 1988). In addition, introduced
birds have major economic impacts in many locations (Lever 1994, Bomford &
Sinclair 2002). Defining and measuring precisely what we mean by impact re-
mains a challenging task, but one that is critical to setting priorities for managing
invasive species (Parker et al. 1999; Smith et al. 1999; Williamson 1999, 2001;
Ricciardi et al. 2000). Assuming that we can quantify impacts, we can, in principle,
apply comparative methods to identify why some species have greater impact than
others, as has been done for other invasion transitions, and use this information in
explanatory models. Here again it will be important to consider how attributes of
the invader and characteristics of the community interact to determine impact. For
example, islands are hypothesized to be more vulnerable to the impacts of exotic
invaders because island species have not been exposed to mainland selective pres-
sures (Loope et al. 1988, Simberloff 1995). Data on introduced birds may provide
opportunities to test these and related hypotheses.
Fourth, the field of invasion ecology has been criticized for drawing a distinction
between invasions resulting from human-caused introductions and natural inva-
sions (as we do in this study) when the underlying processes may be very similar
and mutually informative (Davis et al. 2001). Because many natural bird inva-
sions have been well documented (e.g., Hengeveld 1989, Clegg et al. 2002), birds
provide opportunities to compare these invasion pathways and to assess whether
an understanding of invasion dynamics gleaned from human-caused introductions
can reliably inform us of natural invasion processes, and vice versa. We caution
that this need not be the case. In particular, introduced birds are a distinctly non-
random subset of the world’s birds so that conclusions about how these species
behave may not transfer to other bird groups.
Finally, species introduced to novel habitats provide unique opportunities to
investigate the evolutionary process (Mooney & Cleland 2001). Established pop-
ulations of many introduced species are isolated from source populations and
may diverge rapidly from their ancestors through a combination of divergent
natural selection, genetic drift and divergence under uniform selection (e.g.,
Selander & Johnston 1967, Baker & Moeed 1987, Baker 1992, Reznick &
Ghalambor 2001, Badyaev et al. 2002). In their classic studies on geographic
variation in house sparrows introduced to North America, Selander & Johnston
(1967) found major geographic divergence in coloration and morphometric char-
acters within less than 100 years following establishment. More recent studies in
other birds have also demonstrated rapid genetic and life history differentiation in
introduced populations (e.g., Baker & Mooed 1987, Baker 1992, Badyaev et al.
2002). Given the well-documented history of many bird introductions, often de-
tailing dates and numbers of individuals released, and the range of locations at
which some species have established, populations of introduced birds provide a
remarkable and largely untapped resource for addressing questions about evolu-
tionary change.
ACKNOWLEDGMENTS
We thank Julie Lockwood and Mark Williamson for comments on the manuscript.
TMB especially thanks M. Bergman. DS was supported by a Québec Ministry
of Education Postdoctoral Fellowship and a NSERC (Canada) grant to Louis
Lefebvre.
The Annual Review of Ecology, Evolution, and Systematics is online at

http://ecolsys.annualreviews.org
LITERATURE CITED
Arcese P, Keller LF, Cary JR. 1997. Why hire Wärmeökonomie der Thiere zu ihrer Grösse.
a behaviorist into a conservation or man- Göttinger studien 3:595–708
agement team? In Behavioral Approaches Blackburn TM, Duncan RP. 2001a. Determi-
to Conservation in the Wild, ed. JR Clem- nants of establishment success in introduced
mons, R Buchholz, pp. 48–71. Cambridge, birds. Nature 414:195–97
UK: Cambridge Univ. Press Blackburn TM, Duncan RP. 2001b. Establish-
Badyaev AV, Hill GE, Beck ML, Dervan AE, ment patterns of exotic birds are constrained
Duckworth RA, et al. 2002. Sex-biased by non-random patterns in introduction. J.
hatching order and adaptive population diver- Biogeogr. 28:927–39
gence in a passerine bird. Science 295:316– Blackburn TM, Gaston KJ. 1994. The distribu-
18 tion of body sizes of the worlds bird species.
Baker AJ. 1992. Genetic and morphometric di- Oikos 70:127–30
vergence in ancestral European and descen- Bomford M, Sinclair R. 2002. Australian re-
dent New Zealand populations of chaffinches search on bird pests: impact, management
(Fringilla coelebs). Evolution 46:1784– and future directions. Emu 102:29–45
800 Brooke RK, Lockwood JL, Moulton MP. 1995.
Baker AJ, Moeed A. 1987. Rapid genetic dif- Patterns of success in passeriform bird
ferentiation and founder effects in colonizing introductions on Saint Helena. Oecologia
populations of common mynas (Acridotheres 103:337–42
tristis). Evolution 41:525–38 Brooker ML, Davies NB, Noble DG. 1998.
Bennett PM, Owens IPF. 2002. The Evolution- Rapid decline of host defences in response to
ary Ecology of Birds. Oxford: Oxford Univ. reduced cuckoo parasitism: behavioural flex-
Press ibility of reed warblers in a changing world.
Berger J, Swenson JE, Persson I-L. 2001. Re- Proc. R. Soc. London Ser. B 265:1277–82
colonizing carnivores and native prey: con- Brooks T. 2001. Are unsuccessful avian in-
servation lessons from Pleistocene extinc- vaders rarer in their native range than suc-
tions. Science 291:1036–39 cessful invaders? In Biotic Homogenization,
Berggren A. 2001. Colonization success in Roe- ed. JL Lockwood, ML McKinney, pp. 125–
sel’s bush-cricket Metrioptera roeseli: the 55. New York: Kluwer
effect of propagule size. Ecology 82:274– Brown JH. 1989. Patterns, modes and extents
80 of invasions by vertebrates. See Drake et al.
Bergmann C. 1847. Ueber die Verhältnisse der 1989, pp. 85–109
Bump G. 1970. The Coturnix or Old World sion ecology from the rest of ecology. Divers.
Quails. US Dep. Int. Fish, Wildl. Serv., Distrib. 7:97–102
Bur. Sport, Fish. Wildl, FGL-10. Washing- Dawson JC. 1984. A statistical analysis of
ton, DC: GPO species characteristics affecting the success
Cardillo M. 2002. The life-history basis of lat- of bird introductions. BSc. thesis. Univ. York,
itudinal diversity gradients: how do species UK. 47 pp.
traits vary from the poles to the equator? J. Diamond JM, Veitch CR. 1981. Extinctions and
Anim. Ecol. 71:79–87 introductions in the New Zealand avifauna:
Case TJ. 1996. Global patterns in the establish- cause and effect? Science 211:499–501
ment and distribution of exotic birds. Biol. Diggle P, Liang K, Zeger SL. 1994. Analysis of
Conserv. 78:69–96 Longitudinal Data. Oxford: Clarendon
Cassey P. 2001a. Are there body size implica- Drake JA, Mooney HA, di Castri F, Groves RH,
tions for the success of globally introduced Kruger FG, eds. 1989. Biological Invasions:
land birds? Ecography 24:413–20 a Global Perspective. Chichester: Wiley. 525
Cassey P. 2001b. Determining variation in the pp.
success of New Zealand land birds. Global Duncan RP. 1997. The role of competition and
Ecol. Biogeogr. 10:161–72 introduction effort in the success of passeri-
Cassey P. 2002. Comparative analysis of form birds introduced to New Zealand. Am.
successful establishment among introduced Nat. 149:903–15
landbirds. PhD thesis. Griffith Univ., Aust. Duncan RP, Blackburn TM. 2002. Morpho-
172 pp. logical over-dispersion in game birds (Aves:
Cassey P. 2003. A comparative analysis of the Galliformes) successfully introduced to New
relative success of introduced landbirds on Zealand was not caused by interspecific com-
islands. Evol. Ecol. Res. In press petition. Evol. Ecol. Res. 4:551–61
Clegg SM, Degnan SM, Kikkawa J, Moritz C, Duncan RP, Blackburn TM, Cassey P. 2003.
Estoup A, Owens IPF. 2002. Genetic conse- Factors affecting the release, establishment
quences of sequential founder events by an and spread of introduced birds in New
island-colonizing bird. Proc. Natl. Acad. Sci. Zealand. In Biological Invasions in New
USA 99:8127–32 Zealand, ed. RB Allen, WG Lee. Berlin:
Crawley MJ. 1986. The population biology of Springer-Verlag. In press
invaders. Philos. Trans. R. Soc. London B Duncan RP, Blackburn TM, Veltman CJ. 1999.
314:711–31 Determinants of geographical range sizes: a
Daehler CC. 2001. Two ways to be an invader, test using introduced New Zealand birds. J.
but one is more suitable for ecology. Bull. Anim. Ecol. 68:963–75
Ecol. Soc. Amer. 82:101–2 Duncan RP, Bomford M, Forsyth DM, Conibear
Daehler CC, Strong DR. 1993. Prediction L. 2001. High predictability in introduction
and biological invasions. Trends Ecol. Evol. outcomes and the geograpical range size of
8:380–81 introduced Australian birds: a role for cli-
Dalmazzone S. 2000. Economic factors affect- mate. J. Anim. Ecol. 70:621–32
ing vulnerability to biological invasions. See Duncan RP, Young JR. 1999. The fate of Passer-
Perrings et al. 2000, pp. 17–30 iform introductions on oceanic islands. Con-
Davis MA, Thompson K. 2000. Eight ways to serv. Biol. 13:934–36
be a colonizer; two ways to be an invader: Dunning JB. 1993. CRC Handbook of Avian
A proposed nomenclature scheme for inva- Body Masses. Boca Raton, FL: CRC
sion ecology. Bull. Ecol. Soc. Am. 81:226– Ebenhard T. 1988. Introduced birds and mam-
30 mals and their ecological effects. Swed.
Davis MA, Thompson K, Grime JP. 2001. Wildl. Res. 13:1–107
Charles S. Elton and the dissociation of inva- Ehrlich PH. 1989. Attributes of invaders and the
invading processes: vertebrates. See Drake of foreign birds. Avicultural Magazine 3:8–
et al. 1989, pp. 315–28 11
Elton CS. 1958. The Ecology of Invasions by Guild E. 1940. Western bluebirds in Tahiti. Avi-
Animals and Plants. London: Methuen cultural Magazine 5:284–85
Evans LGR. 1997. The Definitive Checklist of Guix JC, Jover L, Ruiz X. 1997. Muestreos
the Birds of the Western Palaearctic. Amer- del comercio de psitácidos neotropicales en
sham: LGRE la ciudad de Barcelona, España: 1991–1996.
Forsyth DM, Duncan RP. 2001. Propagule size Ararajuba 5:159–67
and the relative success of exotic ungulate Hairston NG, Smith FE, Slobodkin LB. 1960.
and bird introductions to New Zealand. Am. Community structure, population control,
Nat. 157:583–95 and competition. Am. Nat. 94:421–25
Forys JM, Allen CR. 1999. Biological invasions Hengeveld R. 1989. Dynamics of Biological In-
and deletions: community change in south vasions. London: Chapman & Hall
Florida. Biol. Conserv. 87:341–47 Herbold B, Moyle PB. 1986. Introduced species
Gammon DE, Maurer BA. 2002. Evidence for and vacant niches. Am. Nat. 128:751–
non-uniform dispersal in the biological in- 60
vasions of two naturalized North American Hopper KR, Roush RT. 1993. Mate finding, dis-
bird species. Global Ecol. Biogeogr. 11:155– persal, number released, and the success of
61 biological control introductions. Ecol. Ento-
Gaston KJ. 1988. Patterns in the local and re- mol. 18:321–31
gional dynamics of moth populations. Oikos Hurlbert SH. 1984. Pseudoreplication and the
53:49–57 design of ecological field experiments. Ecol.
Gaston KJ. 1994. Rarity. London: Chapman & Mon. 54: 187–211
Hall Klopfer PH. 1962. Behavioral Aspects of Ecol-
Gaston KJ, Blackburn TM. 2000. Pattern and ogy. London: Prentice-Hall
Process in Macroecology. Oxford: Blackwell Koenig WD. 2003. European starlings and their
Sci. effect on native cavity-nesting birds. Con-
Gilpin M. 1990. Ecological prediction. Science serv. Biol. In press
248:88–89 Kolar CK, Lodge DM. 2001. Progress in in-
Goldstein H. 1995. Multilevel Statistical Mod- vasion biology: predicting invaders. Trends
els. Oxford: Edward Arnold Ecol. Evol. 16:199–204
Green RE. 1997. The influence of numbers re- Lack D. 1947. The significance of clutch-size.
leased on the outcome of attempts to intro- Part I—Intraspecific variations. Ibis 89:302–
duce exotic bird species to New Zealand. J. 52
Anim. Ecol. 66:25–35 Lack D. 1948. The significance of clutch-size.
Greenberg RS, Mettke-Hofmann C. 2001. Eco- Part III—Some interspecific comparisons.
logical aspects of neophobia and neophilia in Ibis 90:25–45
birds. In Current Ornithology, ed. V Nolan Lefebvre L, Reader SM, Sol D. 2003. Brains,
Jr, ED Ketterson, CF Thompson, 16:119–78. innovations and evolution in birds and pri-
New York: Kluwer mates. Brain Behav. Evol. In press
Gregory RD, Gaston KJ. 2000. Explanations of Legendre S, Clobert J, Møller AP, Sorci G.
commonness and rarity in British breeding 1999. Demographic stochasticity and social
birds: separating resource use and resource mating system in the process of extinction of
availability. Oikos 88:515–26 small populations: the case of passerines in-
Griffith B, Scott JM, Carpenter JW, Reed C. troduced to New Zealand. Am. Nat. 153:449–
1989. Translocation as a species tool: status 63
and strategy. Science 245:477–80 Lensink R. 1998. Temporal and spatial ex-
Guild E. 1938. Tahitian aviculture: acclimation pansion of the Egyptian goose Alopochen
aegyptiacus in The Netherlands, 1967–94. J. 1999. Sexual selection and extinction: the
Biogeogr. 25:251–63 fate of plumage-dimorphic and plumage-
Lever C. 1987. Naturalized Birds of the World. monomorphic birds introduced onto islands.
New York: Longman Sci. Tech. Evol. Ecol. Res. 1:549–65
Lever C. 1994. Naturalized Animals: the Ecol- McNeely JA, ed. 2001. The Great Reshuf-
ogy of Successfully Introduced Species. Lon- fling: Human Dimensions of Invasive Alien
don: T & AD Poyser Species. Gland, Switz./Cambridge, UK:
Lockwood JL. 1999. Using taxonomy to predict IUCN
success among introduced avifauna: relative Mielke HW. 1989. Patterns of Life: Biogeog-
importance of transport and establishment. raphy of a Changing World. Boston: Unwin

Conserv. Biol. 13:560–67 Hyman
Lockwood JL, Brooks TM, McKinney ML. Mooney HA, Cleland EE. 2001. The evolution-
2000. Taxonomic homogeneization of the ary impact of invasive species. Proc. Natl.
global avifauna. Anim. Conserv. 3:27–35 Acad. Sci. USA 98:5446–51
Lockwood JL, Moulton MP. 1994. Ecomorpho- Morse DH. 1980. Behavioral Mechanisms in
logical pattern in Bermuda birds: the influ- Ecology. Cambridge, MA: Harvard Univ.
ence of competition and implications for na- Press
ture preserves. Evol. Ecol. 8:53–60 Moulton MP. 1985. Morphological similarity
Lockwood JL, Moulton MP, Anderson SK. and coexistence of congeners: an experimen-
1993. Morphological assortment and the as- tal test with introduced Hawaiian birds. Oikos
sembly of communities of introduced passer- 44:301–5
iforms on oceanic islands: Tahiti versus Moulton MP. 1993. The all-or-none pattern in
Oahu. Am. Nat. 141:398–408 introduced Hawaiian passeriformes: the role
Lodge DM. 1993. Biological invasions: lessons of competition sustained. Am. Nat. 141:105–
for ecology. Trends Ecol. Evol. 8:133–36 19
Long JL. 1981. Introduced Birds of the World. Moulton MP, Miller KE, Tillman EA. 2001a.
London: David & Charles Patterns of success among introduced birds
Loope LL, Hamman O, Stone CP. 1988. Com- in the Hawaiian islands. Stud. Avian Biol.
parative conservation biology of oceanic 22:31–46
archipelagos. Hawaii and the Galápagos. Moulton MP, Pimm SL. 1983. The introduced
BioScience 38:272–82 Hawaiian avifauna: biogeographic evidence
Mack RN, Simberloff D, Lonsdale WM, Evans for competition. Am. Nat. 121:669–90
H, Clout M, Bazzaz FA. 2000. Biotic inva- Moulton MP, Pimm SL. 1986. Species intro-
sions: Causes, epidemiology, global conse- ductions to Hawaii. In Ecology of Biological
quences, and control. Ecol. Appl. 10:689– Invasions in North America and Hawaii, ed.
710 HA Mooney, JA Drake, pp. 231–49. Berlin:
Mayr E. 1965. The nature of colonising birds. In Springer-Verlag
The Genetics of Colonizing Species, ed. HG Moulton MP, Sanderson JG. 1996. Predict-
Baker, GL Stebbins, pp. 29–43. New York: ing the fate of Passeriform introductions in
Academic oceanic islands. Conserv. Biol. 11:552–58
McArdle BH. 1996. Levels of evidence in stud- Moulton MP, Sanderson JG. 1999. Fate of
ies of competition, predation, and disease. NZ Passeriform introductions: reply to Duncan
J. Ecol. 20:7–15 and Young. Conserv. Biol. 13:937–38
McLain DK, Moulton MP, Redfearn TP. 1995. Moulton MP, Sanderson JG, Labisky RF.
Sexual selection and the risk of extinction of 2001b. Patterns of success in game bird
introduced birds on oceanic islands. Oikos (Aves: Galliformes) introductions to the
90:599–605 Hawaiian islands and New Zealand. Evol.
McLain DK, Moulton MP, Sanderson JG. Ecol. Res. 3:507–19
Nee S, Read AF, Greenwood JJD, Harvey invaders: a priority for aquatic resource man-
PH. 1991. The relationship between abun- agement. Can. J. Fish. Aquat. Sci. 55:1759–
dance and body size in British birds. Nature 65
351:312–13 Ricciardi A, Steiner WWM, Mack RN, Sim-
Newsome AE, Noble IR. 1986. Ecological and berloff D. 2000. Toward a global informa-
physiological characters of invading species. tion system for invasive species. BioScience
In Biological Invasions, ed. RH Groves, JJ 50:239–44
Burdon, pp. 1–20. Cambridge, UK: Cam- Richardson DM, Pysek P, Rejmanek M,
bridge Univ. Press Barbour MG, Panetta FD, West CJ. 2000.
Owens IPF, Bennett PM. 1995. Ancient ecolog- Naturalization and invasion of alien plants:
ical diversification explains life-history vari- concepts and definitions. Divers. Distrib.
ation among living birds. Proc. R. Soc. Lon- 6:93–107

don Ser. B 261:227–32 Sakai AK, Allendorf FW, Holt JS, Lodge DM,
Parker IM, Simberloff D, Lonsdale WM, Good- Molofsky J, et al. 2001. The population bi-
ell K, Whonham M, et al. 1999. Impact: to- ology of invasive species. Annu. Rev. Ecol.
ward a framework for understanding the eco- Syst. 32:305–32
logical effects of invaders. Biol. Invasions Sax DF, Brown JH. 2000. The paradox of inva-
1:3–19 sion. Global Ecol. Biogeogr. 9:363–71
Penny M. 1974. The Birds of Seychelles and the Selander RK, Johnston RF. 1967. Evolution in
Outlying Islands. London: Collins the house sparrow. Intrapopulation variation
Perrings C, Williamson M, Dalmazzone S, eds. in North America. Condor 69:217–58
2000. The Economics of Biological Inva- Shea K, Chesson P. 2002. Community ecology
sions. Cheltenham, UK: Edward Elgar. 249 theory as a framework for biological inva-
pp. sions. Trends Ecol. Evol. 17:170–76
Pimm SL. 1991. The Balance of Nature? Shigesada N, Kawasaki K. 1997. Biological In-
Chicago: Univ. Chicago Press vasions: Theory and Practice. Oxford: Ox-
Pimm SL, Jones HL, Diamond J. 1988. On the ford Univ. Press
risk of extinction. Am. Nat. 132:757–85 Sibley CG, Ahlquist JE. 1990. Phylogeny and
Plotkin HC, Odling-Smee FJ. 1979. Learning, Classification of Birds: a Study in Molecular
change and evolution: an enquiry into the te- Evolution. New Haven: Yale Univ. Press
lenomy of learning. Adv. Stud. Behav. 10:1– Sibley CG, Monroe BL. 1990. Distribution and
41 Taxonomy of Birds of the World. New Haven:
Rahbek C. 1997. The relationship among area, Yale Univ. Press
elevation, and regional species richness in Sibley CG, Monroe BL. 1993. Supplement to
Neotropical birds. Am. Nat. 149:875–902 the Distribution and Taxonomy of Birds of
Reed JM. 1999. The role of behavior in recent the World. New Haven: Yale Univ. Press
avian extinctions and endangerments. Con- Silva T, Reino LM, Borralho R. 2002. A model
serv. Biol. 13:232–41 for range expansion of an introduced species:
Reznick DN, Ghalambor CK. 2001. The popu- the common waxbill Estrilda astrild in Por-
lation ecology of contemporary adaptations: tugal. Divers. Distrib. 8:319–26
what empirical studies reveal about the con- Simberloff D. 1992. Extinction, survival, and
ditions that promote adaptive evolution. Ge- effects of birds introduced to the Mascarenes.
netica 112–13:183–98 Acta Oecol. 13:663–78
Rhymer JM, Simberloff D. 1996. Extinction by Simberloff D. 1995. Why do introduced species
hybridization and introgression. Annu. Rev. appear to devastate islands more than main-
Ecol. Syst. 27:83–109 land areas? Pac. Sci. 49:87–97
Ricciardi A, Rasmussen JB. 1998. Predicting Simberloff D, Boecklen W. 1991. Patterns
the identity and impact of future biological of extinction in the introduced Hawaiian
avifauna: a reexamination of the role of com- van Riper C III, van Riper CG, Goff ML, Laird
petition. Am. Nat. 138:300–27 M. 1986. The epizootiology and ecologi-
Smallwood KS. 1994. Site invasibility by exotic cal significance of malaria in Hawaiian land
birds and mammals. Biol. Conserv. 69:251– birds. Ecol. Monog. 56:327–44
59 Veit RR, Lewis MA. 1996. Dispersal, popula-
Smith CS, Lonsdale WM, Fortune J. 1999. tion growth, and the Allee effect: dynamics
When to ignore advice: invasion predictions of the house finch invasion of eastern North
and decision theory. Biol. Invasion 1:89–96 America. Am. Nat. 148:255–74
Sol D. 2000a. Are islands more susceptible to be Veltman CJ, Nee S, Crawley MJ. 1996. Corre-
invaded than continents? Birds say no. Ecog- lates of introduction success in exotic New
raphy 23:687–92 Zealand birds. Am. Nat. 147:542–57
Sol D. 2000b. Introduced species: a signifi- Vermeij GJ. 1996. An agenda for invasion biol-
cant component of the global environmental ogy. Biol. Conserv. 78:3–9
change. PhD thesis. Univ. Barcelona. 160 pp. Vitousek PM. 1988. Diversity and biological
Sol D. 2003. Behavioural flexibility: a ne- invasions of oceanic islands. In Biodiversity,
glected issue in the ecological and evolution- ed. EO Wilson, FM Peter, pp. 181–89. Wash-
ary literature? In Animal Innovation, ed. K ington, DC: Nat. Acad. Press
Laland, SM Reader. Oxford: Oxford Univ. Vitousek PM, D’Antonio CM, Loope LL,
Press. In press Rejmánek M, Westbrooks R. 1997. Intro-
Sol D, Lefebvre L. 2000. Behavioural flexibility duced species: a significant component of
predicts invasion success in birds introduced human-caused global change. NZ J. Ecol.
to New Zealand. Oikos 90:599–605 21:1–16
Sol D, Timmermans S, Lefebvre L. 2002. Be- Williamson M. 1996. Biological Invasions.
havioural flexibility and invasion success in London: Chapman & Hall
birds. Anim. Behav. 63:495–502 Williamson M. 1999. Invasions. Ecography
Sorci G, Møller AP, Clobert J. 1998. Plumage 22:5–12
dichromatism of birds predicts introduction Williamson M. 2001. Can the impacts of in-
success in New Zealand. J. Anim. Ecol. vasive species be predicted? In Weed Risk
67:263–69 Assessment, ed. RH Groves, FD Panetta,
Stevens GC. 1989. The latitudinal gradient in JG Virtue, pp. 20–33. Collingwood, Aust.:
geographical range: how so many species co- CSIRO
exist in the tropics. Am. Nat. 133:240–56 Williamson M, Fitter A. 1996. The vary-
Thomson GM. 1922. The Naturalization of ing success of invaders. Ecology 77:1661–
Plants and Animals in New Zealand. Cam- 66
bridge, UK: Cambridge Univ. Press Wolf CM, Griffith B, Reed C, Temple SA. 1998.
Van den Bosch F, Hengeveld R, Metz JAJ. 1992. Avian and mammalian translocations: update
Analysing the velocity of animal range ex- and reanalysis of 1987 survey data. Conserv.
pansion. J. Biogeogr. 19:135–50 Biol. 86:1142–54
P1: FDS
October 17, 2003 16:25 Annual Reviews AR200-FM
Annual Review of Ecology, Evolution, and Systematics

Volume 34, 2003
CONTENTS
EFFECTS OF INTRODUCED BEES ON NATIVE ECOSYSTEMS,

Dave Goulson 1
AVIAN SEXUAL DICHROMATISM IN RELATION TO PHYLOGENY AND

ECOLOGY, Alexander V. Badyaev and Geoffrey E. Hill 27
PALEOBIOGEOGRAPHY: THE RELEVANCE OF FOSSILS TO
BIOGEOGRAPHY, Bruce S. Lieberman 51
THE ECOLOGY OF BIRD INTRODUCTIONS, Richard P. Duncan,
Tim M. Blackburn, and Daniel Sol 71
THE EFFECTS OF GENETIC AND GEOGRAPHIC STRUCTURE ON
NEUTRAL VARIATION, Brian Charlesworth, Deborah Charlesworth, and
Nicholas H. Barton 99
DATA, MODELS, AND DECISIONS IN US MARINE FISHERIES
MANAGEMENT: LESSONS FOR ECOLOGISTS, Kenneth A. Rose and
James H. Cowan Jr. 127
PARTITIONING OF TIME AS AN ECOLOGICAL RESOURCE,
Noga Kronfeld-Schor and Tamar Dayan 153
PERFORMANCE COMPARISONS OF CO-OCCURING NATIVE AND ALIEN
INVASIVE PLANTS: IMPLICATIONS FOR CONSERVATION AND
RESTORATION, Curtis C. Daehler 183
GENETIC VARIATION IN RARE AND COMMON PLANTS,
Christopher T. Cole 213
THE ECOLOGY AND EVOLUTION OF INSECT BACULOVIRUSES,
Jenny S. Cory and Judith H. Myers 239
LATITUDINAL GRADIENTS OF BIODIVERSITY: PATTERN, PROCESS,
SCALE, AND SYNTHESIS, M.R. Willig, D.M. Kaufman,
and R.D. Stevens 273
RECENT ADVANCES IN THE (MOLECULAR) PHYLOGENY OF
VERTEBRATES, Axel Meyer and Rafael Zardoya 311
THE ROLE OF REINFORCEMENT IN SPECIATION: THEORY AND DATA,
Maria R. Servedio and Mohamed A.F. Noor 339
EXTRA-PAIR PATERNITY IN BIRDS: CAUSES, CORRELATES, AND
CONFLICT, David F. Westneat and Ian R.K. Stewart 365
v
P1: FDS
October 17, 2003 16:25 Annual Reviews AR200-FM
vi CONTENTS
SPECIES-LEVEL PARAPHYLY AND POLYPHYLY: FREQUENCY, CAUSES,

AND CONSEQUENCES, WITH INSIGHTS FROM ANIMAL
MITOCHONDRIAL DNA, Daniel J. Funk and Kevin E. Omland 397
PROTECTIVE ANT-PLANT INTERACTIONS AS MODEL SYSTEMS IN
ECOLOGICAL AND EVOLUTIONARY RESEARCH, Martin Heil
and Doyle McKey 425
FUNCTIONAL MATRIX: A CONCEPTUAL FRAMEWORK FOR PREDICTING
PLANT EFFECTS ON ECOSYSTEM PROCESSES, Valerie T. Eviner
and F. Stuart Chapin III 455

EFFECTS OF HABITAT FRAGMENTATION ON BIODIVERSITY,

Lenore Fahrig 487
SOCIAL ORGANIZATION AND PARASITE RISK IN MAMMALS:
INTEGRATING THEORY AND EMPIRICAL STUDIES, Sonia Altizer,
Charles L. Nunn, Peter H. Thrall, John L. Gittleman, Janis Antonovics,
Andrew A. Cunnningham, Andrew P. Dobson, Vanessa Ezenwa,
Kate E. Jones, Amy B. Pedersen, Mary Poss, and Juliet R.C. Pulliam 517
THE COMMUNITY-LEVEL CONSEQUENCES OF SEED DISPERSAL
PATTERNS, Jonathan M. Levine and David J. Murrell 549
THE ECOLOGY AND EVOLUTION OF SEED DISPERSAL: A THEORETICAL
PERSPECTIVE, Simon A. Levin, Helene C. Muller-Landau, Ran Nathan,
and Jérôme Chave 575
ANALYSIS OF RATES OF MORPHOLOGIC EVOLUTION,
Peter D. Roopnarine 605
DEVELOPMENT AND THE GENETICS OF EVOLUTIONARY CHANGE
WITHIN INSECT SPECIES, Paul M. Brakefield, Vernon French,
and Bas J. Zwaan 633
FLEXIBILITY AND SPECIFICITY IN CORAL-ALGAL SYMBIOSIS:
DIVERSITY, ECOLOGY, AND BIOGEOGRAPHY OF SYMBIODINIUM,
Andrew C. Baker 661
INDEXES
Subject Index 691
Cumulative Index of Contributing Authors, Volumes 30–34 705
Cumulative Index of Chapter Titles, Volumes 30–34 708
ERRATA
An online log of corrections to Annual Review of Ecology,
Evolution, and Systematics chapters may be found at
http://ecolsys.annualreviews.org/errata.shtml

Ecology of Bird Introduction

Uploaded by

Copyright:

Available Formats

Ecology of Bird Introduction

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ecology of Bird Introduction

Uploaded by

Copyright:

Available Formats

30 Sep 2003 14:24 AR AR200-ES34-04.tex AR200-ES34-04.

sgm LaTeX2e(2002/01/18) P1: GJB

Annu. Rev. Ecol. Evol. Syst. 2003. 34:71–98

THE ECOLOGY OF BIRD INTRODUCTIONS

United Kingdom; email: [email protected]

Key Words establishment success, exotic species, introduced birds, biological

72 DUNCAN ¥ BLACKBURN ¥ SOL

Nevertheless, only a small proportion of the species introduced to a new lo-

outside their natural range?

Why Study Bird Introductions?

THE ECOLOGY OF BIRD INTRODUCTIONS 73

Figure 1 Schematic representation of stages in the process of a human-caused inva-

74 DUNCAN ¥ BLACKBURN ¥ SOL

Definitions and Framework

THE ECOLOGY OF BIRD INTRODUCTIONS 75

is termed an introduction event. When we use the term introduction, we refer

environment, regardless of whether it establishes or spreads.

TRANSPORT WITH INTRODUCTION

Which Species Were Selected for Introduction?

TAXONOMIC SELECTIVITY Taxonomic selectivity in global bird introductions has

76 DUNCAN ¥ BLACKBURN ¥ SOL

GEOGRAPHICAL SELECTIVITY As for many other taxa, bird species richness is

CHARACTER SELECTIVITY These taxonomic and geographical patterns of selec-

THE ECOLOGY OF BIRD INTRODUCTIONS 77

To Which Locations Were Birds Introduced?

78 DUNCAN ¥ BLACKBURN ¥ SOL

species-, location-, and event-level factors, respectively (Blackburn & Duncan

Statistical Analysis of Historical Introduction Outcomes

THE ECOLOGY OF BIRD INTRODUCTIONS 79

of factors included in the model (a greater frequency of Type I errors). In plain

if we fail correctly to account for the clustering of observations in our data.

Event-Level Influences on Establishment Success

80 DUNCAN ¥ BLACKBURN ¥ SOL

of establishment [Dawson 1984 (cited in Williamson 1996); Newsome & Noble

THE ECOLOGY OF BIRD INTRODUCTIONS 81

CLIMATE/ENVIRONMENTAL MATCHING One of the most frequently stated hypothe-

82 DUNCAN ¥ BLACKBURN ¥ SOL

Location-Level Influences on Establishment Success

THE ECOLOGY OF BIRD INTRODUCTIONS 83

confounded with variation in other factors that affect establishment, making it

84 DUNCAN ¥ BLACKBURN ¥ SOL

as the number of established species accumulates. Moulton and coworkers argue

overdispersion among the established species.

THE ECOLOGY OF BIRD INTRODUCTIONS 85

Finally, in a global analysis of bird introductions, Blackburn & Duncan (2001a)

Species-Level Influences on Establishment Success

DO SPECIES DIFFER IN THEIR PROBABILITY OF ESTABLISHING? Simberloff &

86 DUNCAN ¥ BLACKBURN ¥ SOL

Preadaptations to establishment One reason for failing to establish is that a

THE ECOLOGY OF BIRD INTRODUCTIONS 87

88 DUNCAN ¥ BLACKBURN ¥ SOL

circumstances, larger-bodied, longer-lived, species with slower rates of population

Constraints on success Some traits may lower establishment probability regard-

TAXONOMIC VARIATION IN ESTABLISHMENT PROBABILITY Given that species dif-

THE ECOLOGY OF BIRD INTRODUCTIONS 89

predict establishment from knowledge of a species’ taxonomic status alone (Daehler

of establishment. Nevertheless, the extent to which at least some variation in es-

90 DUNCAN ¥ BLACKBURN ¥ SOL

Three studies have nevertheless examined the determinants of current geo-

THE ECOLOGY OF BIRD INTRODUCTIONS 91