SEARCH

SEARCH BY CITATION

Keywords:

  • Australian birds;
  • biological invasions;
  • climatic suitability;
  • ecology of invasions;
  • extinction;
  • introduced species;
  • introduction effort;
  • range size

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix
  • 1
    We investigated factors hypothesized to influence introduction success and subsequent geographical range size in 52 species of bird that have been introduced to mainland Australia.
  • 2
    The 19 successful species had been introduced more times, at more sites and in greater overall numbers. Relative to failed species, successfully introduced species also had a greater area of climatically suitable habitat available in Australia, a larger overseas range size and were more likely to have been introduced successfully outside Australia. After controlling for phylogeny these relationships held, except that with overseas range size and, in addition, larger-bodied species had a higher probability of introduction success. There was also a marked taxonomic bias: gamebirds had a much lower probability of success than other species. A model including five of these variables explained perfectly the patterns in introduction success across-species.
  • 3
    Of the successful species, those with larger geographical ranges in Australia had a greater area of climatically suitable habitat, traits associated with a faster population growth rate (small body size, short incubation period and more broods per season) and a larger overseas range size. The relationships between range size in Australia, the extent of climatically suitable habitat and overseas range size held after controlling for phylogeny.
  • 4
    We discuss the probable causes underlying these relationships and why, in retrospect, the outcome of bird introductions to Australia are highly predictable.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

Many species have been deliberately or accidentally introduced to areas outside of where they naturally occur (Long 1981; Lever 1994). Some species have established populations successfully and spread, but others either failed to establish or established but have not spread. Although these introductions have often had disastrous consequences for indigenous biota (e.g. Savidge 1987), they also provide opportunities to improve our understanding of biological invasions and other ecological processes.

The historical record of bird introductions to New Zealand has been used extensively to test hypotheses about the factors influencing introduction success and the subsequent spread of introduced birds (Veltman, Nee & Crawley 1996; Duncan 1997; Green 1997; Sorci, Møller & Clobert 1998; Duncan, Blackburn & Veltman 1999; Legendre et al. 1999). A similarly detailed record of bird introductions to Australia exists: at least 55 species have been introduced to mainland Australia, of which 19 have successfully established (Long 1981; Newsome & Noble 1986). Our aim in this study is to test hypotheses about factors thought to influence introduction success and geographical range size using this historical record. The hypotheses that we test can be divided into three groups:

  • 1
    Availability of climatically suitable habitat. One of the most frequently stated hypotheses in the biological invasion literature is that species should have a greater chance of success if they are introduced to an area with a climate that closely matches that in their original range (Brown 1989; Mack 1996; Williamson 1996). This hypothesis has never been tested using a set of successful and failed introductions. Here we test the hypothesis that introduced species with a larger area of climatically suitable habitat in Australia should have both a higher probability of successful introduction and, if successful, should have a larger geographical range in Australia.
  • 2
    Life history traits. Certain life history traits are hypothesized to increase the probability of successful introduction and to be associated with larger geographical range sizes. In particular, species with a faster population growth rate should have higher introduction success because they will be less vulnerable to local extinction when their population is small and, following successful introduction, should be able to colonize more sites and thereby attain a larger range size (Gaston 1988; Hanski 1991; Hanski, Kouki & Halkka 1993; Holt et al. 1997). We test this hypothesis by searching for significant correlations between introduction success, geographical range size and biological traits that are probable correlates of population growth rate, including body size, clutch size, number of broods per season and incubation time. In addition, species with sexually dichromatic plumage, migratory species, solitary species and species with a narrow diet are hypothesized, or have been shown, to have a lower probability of successful introduction (Mayr 1965; Veltman et al. 1996; Sorci et al. 1998). Birds that are partial migrants in their natural range have also been shown to have a larger geographical range size following introduction to New Zealand (Duncan et al. 1999).
  • 3
    Historical factors. Greater introduction effort (number of introduction attempts, number of locations at which species are introduced or the total number of individuals introduced) is the strongest predictor of success for birds introduced to New Zealand (Veltman et al. 1996), and predicts the success of bird introductions to Australia (Newsome & Noble 1986). For introduced Australian birds, we extend this result by testing if greater introduction effort is associated with a larger geographical range, as is the case for birds introduced successfully to New Zealand (Duncan et al. 1999). We also test if successful introduction elsewhere and the overseas range size of a species predict introduction success in Australia, as has been found elsewhere (Moulton & Pimm 1986; Rejmánek 1996; Reichard & Hamilton 1997).

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

Data sources

We identified 55 bird species introduced to Australia from Newsome & Noble (1986) and Long (1981). We excluded two species that were introduced only to offshore islands (Alopechen aegyptiacus (Linnaeus 1766) and Meleagris gallopavo (Linnaeus 1758)), and one species (Lophura nycthemera (Linnaeus 1758)) that established a population that was subsequently destroyed by humans. Our data set thus comprised a total of 52 species that have been introduced to mainland Australia, of which 19 species have successfully established (Fig. 1, Appendix 1).

image

Figure 1. Hypothesized phylogenetic relationships among the species of birds introduced to mainland Australia. *Indicates species that were successfully introduced. The branch lengths shown are arbitrary and were set as equal in all phylogenetically controlled comparative analyses. Nomenclature follows Sibley & Monroe (1990).

Download figure to PowerPoint

The minimum recorded number of separate introductions, the minimum number of introduction sites and the minimum total number of birds introduced for each species was collated from Long (1981) and Newsome & Noble (1986). We followed Newsome & Noble (1986) in grouping species into three categories: 20 or fewer individuals introduced (including species classed as unknown or cage escapees), 21–100 individuals introduced, or > 100 individuals introduced. Species that had been successfully introduced to locations outside Australia were identified from Long (1981).

For the successful species, two measures of geographical range size in Australia were taken from Blakers, Davies & Reilly (1984): (i) ‘observed range size’, the number of 1° blocks in which the species was confirmed sighted, and (ii) ‘breeding range size’, the number of 1° blocks in which the species was confirmed breeding. The overseas geographical range of each species (i.e. the geographical range excluding Australia) was calculated by digitizing the distribution maps in Long (1981).

A binary variable coding for plumage dichromatism (i.e. a colour difference between males and females of the same species) was taken from Sorci et al. (1998) for those species introduced to both Australia and New Zealand. The remaining species were scored as monochromatic or dichromatic from illustrations in Heinzel, Fitter & Parslow (1984), or from descriptions in Long (1981).

We classed each species as associated with human-dominated habitats or not, from the habitat descriptions in Sibley & Monroe (1990). When a species was listed in Sibley & Monroe (1990) as occupying either towns, gardens, farmland, fields or pastures we classed that species as being associated with human-dominated habitats; species not listed as occupying those habitats were classed otherwise.

Availability of climatically suitable habitat

We used a climate-matching procedure to quantify the climatically suitable habitat available to each introduced species. The climate software (Pheloung 1996) contains data for 16 climate variables (Table 1) for approximately 8000 meteorological stations outside Australia. The geographical range (excluding Australia) of an introduced species (digitized from Long 1981) was plotted on a world map and the climate data from the meteorological stations that fell within the boundaries of that range were used as input data for that species. Using a spatial resolution of 0.5 degrees (latitude × longitude), Australia was divided into 2795 grid points. The 16 climate variables were generated for each grid point based on long-term data from meteorological stations in Australia (Nix 1986). For each of 16 climate variables at each of the input meteorological stations, the difference between the value for each input meteorological station and the value at each Australian grid point was divided by the global standard deviation for the variable to generate a standard score. A Euclidian distance was computed as the square root of the sum of the squares of the standard scores for each of the 16 climate variables, divided by 16. The resultant value was then compared to a normal distribution of reference scores that partition the normal distribution into percentage categories based on the area under the normal distribution curve. Scores within 10% of the mean score are those with the closest possible climate match, and scores of 80% or higher, which fall in the tails of the normal distribution, have the lowest climate match. climate repeats this matching process for all input meteorological stations. The closest matching score is then selected for each Australian grid square because, if a species occurs at an overseas location that closely matches the climate at a given Australian location, whether it also occurs at less well matched overseas locations is irrelevant. Hence, a close climate match is identified for an Australian grid square if at least one of the overseas locations where the species occurs closely matches that grid square. For each species, the number of Australian grid squares allocated to each climate matching class is a measure of Australia’s land area in that climate matching class.

Table 1.  The 16 climate parameters (temperature and rainfall) used to estimate the extent of climatically matched habitat in the climate program (Pheloung 1996). Estimates of these parameters are derived from long-term averages of monthly minimum and maximum temperatures and rainfall for each of the approximately 8000 meteorological stations in the climate database
Temperature parameters (°C)Rainfall parameters (mm)
Mean annualMean annual
Minimum of coolest monthMean of wettest month
Maximum of warmest monthMean of driest month
Average rangeMean monthly CV
Mean of coolest quarterMean of coolest quarter
Mean of warmest quarterMean of warmest quarter
Mean of wettest quarterMean of wettest quarter
Mean of driest quarterMean of warmest quarter

Data analysis

The factors influencing introduction success (a binary variable) were modelled using logistic regression. We first included each predictor variable alone in a logistic regression model and then used a forward selection procedure to identify a subset of predictor variables that could independently predict introduction success. We then used a cross-validation procedure to test the ability of these variables to correctly predict introduction outcomes. We estimated the probability that a particular species would have succeeded by removing that species from the dataset and constructing a logistic regression model using the data for the remaining species. The resulting model was then used to predict the probability that the removed species would have succeeded, a procedure that ensures that the data used to construct the model are independent of the data used to predict the probability of success. This procedure was repeated for all species and the predicted probabilities of introduction success compared with the actual outcomes. For geographical range size we used standard linear regression and analysis of variance to model relationships with predictor variables.

We repeated the analyses using the method of independent contrasts (Felsenstein 1985) to control for phylogenetic relatedness among species. This is important because the birds introduced to Australia are not a random sample of all birds. The families Phasianidae and Anatidae, for example, contain just over 3% of the world’s bird species (following Sibley & Monroe 1990), but species in these families comprise 27% of the birds introduced to Australia. Hence, relationships in our data could be biased by the over-representation of certain taxa as a result of the introduction process (Lockwood 1999). The effect of this taxonomic bias will be reduced if we test for relationships independent of phylogeny.

We constructed a phylogeny for the birds introduced to Australia (Fig. 1) from data in Sibley & Ahlquist (1990), Sibley & Monroe (1990), Martin & Clobert (1996) and Livezey (1997). The method of independent contrasts requires a fully resolved phylogeny, but our phylogeny contained several unresolved nodes. We therefore used a model (Comparative Analysis by Independent Contrasts (CAIC); Purvis & Rambaut 1995) which applies the method of independent contrasts to data sets for which only approximate phylogenies are known. Because we lacked data for all branch lengths, we set these as equal. The independent contrasts for continuous variables were calculated using the crunch option in CAIC (Purvis & Rambaut 1994). Relationships between these sets of contrasts were then analysed using ordinary least squares regression, with the regression line forced through the origin (Garland, Harvey & Ives 1992). Where one variable was dichotomous (e.g. introduction success) and the second variable was continuous, the independent contrasts were calculated using the brunch option (Purvis & Rambaut 1994). If there is no relationship between two variables then, under this option, the independent contrasts should show no significant tendency to be either positive or negative. We tested for such a tendency using a one-sample t-test against an expected value of zero (Purvis & Rambaut 1994). Where one variable was dichotomous and the second variable was categorical we also used the brunch option, treating the categorical variable as continuous by assigning a rank order score to its categories. In all cases the categories could be ranked sensibly.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

Introduction success

Availability of climatically suitable habitat

For the successful and failed bird species introduced to Australia, we compared the mean of the (log-transformed) number of target points allocated to climate matching classes below a progressively higher threshold (≤10%, ≤20%, ≤30% and so on). Successfully introduced species had a significantly greater number of target points in climate matching classes below all but the highest threshold (≤80%; Fig. 2). This difference was greatest in climate matching classes below a threshold of about 50%. Indeed, there was a significant relationship between the probability of successful introduction and the total number of target points in the ≤ 50% climate matching classes, both across-species and having controlled for phylogeny (Table 2). The greater the extent of more strongly climatically matched habitat that was present in Australia, the greater the probability that a species was successfully introduced. Of the nine species with 25 or fewer points in the first two climate matching classes, none were introduced successfully.

image

Figure 2. The mean (± 95% confidence limits) of the log10 (total number of target points) in climate matching classes below a progressively higher threshold (≤10%, ≤20%, ≤30%, and so on), for bird species that were successfully (shaded bars) and unsuccessfully (open bars) introduced to Australia.

Download figure to PowerPoint

Table 2.  Across-species and phylogenetically independent predictors of avian introduction success to Australia. The results of the across-species analyses are the change in deviance resulting from adding each predictor variable in the left-hand column to a logistic regression model with introduction outcome (success or failure) as the response variable. d.f. = the degrees of freedom associated with adding each variable to the model. The results of the analyses of phylogenetically independent contrasts are t-tests of the tendency for the independent contrasts associated with the variable in the left-hand column to be significantly positive or negative. A significant positive t-value indicates that larger values of the variable in the left-hand column are associated with greater introduction success, a significant negative t-value indicates the opposite. n = the number of independent contrasts associated with each t-test
VariableAcross-speciesIndependent contrasts
d.f.Change in deviancent
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

Habitat availability
 log10 (total number of points in the ≤50% climate matching classes)1 8·95**15 2·81*
 human-modified habitat use1 1·29 2− 0·27
Life history traits
 log10 (body mass)1 0·1615  2·78*
 log10 (clutch size)1 2·7515  0·52
 log10 (broods per season)1 5·47*15 1·84
 log10 (incubation period)1 0·0214 1·73
 plumage dichromatism1 2·09 8−1·42
 migration2 4·91 8−1·20
 flocking1 0·71 3      0·26
 diet1 4·53* 6      5·02**
Historical factors
 log10 (overseas range size)1 8·57**15      2·01
 log10 (number of introduction events)122·65***14     7·02***
 log10 (number of introduction sites)122·85***15      6·53***
 number of individuals introduced2 7·00*1010·57***
 introduced successfully elsewhere1 8·49** 9     3·01*
Life history traits and historical factors

The number of broods per season, and diet, were significant but weak predictors of introduction success across-species, and body mass and diet were significant predictors having controlled for phylogeny (Table 2). Species producing more broods per season, species with omnivorous or carnivorous diets and species with larger body mass tended to have a higher probability of introduction success.

All of the historical factors that we considered were significant predictors of introduction success across-species and, with the exception of overseas range size, having controlled for phylogeny. Relative to species that failed, the species introduced successfully to Australia tended to have larger overseas range sizes, to be released at more sites, to be released more often and in greater numbers, and to have been introduced successfully elsewhere.

In addition to the above relationships, there was a marked taxonomic bias in introduction success: of 11 species of gamebird (Order: Galliformes) introduced, only one succeeded (peafowl, Pavo cristatus (Linnaeus 1758); we also excluded a twelfth species of gamebird, Lophura nycthemera, that failed but whose population was destroyed by humans). As a group, gamebirds had a significantly lower success rate than other species (change in deviance = 5·34, P = 0·02).

Modelling introduction success

We coded the number of sites to which species were introduced as a binary variable (introduced to ≤ 2 or > 2 sites) because only two of the 30 species introduced to ≤ 2 sites succeeded, while 17 of the 22 species introduced to > 2 sites succeeded. Five variables were then significant predictors of introduction success in a forward selection model (Table 3). Species that were introduced to > 2 sites, that were not gamebirds, that had a greater area of climatically suitable habitat available in Australia, that had been introduced successfully elsewhere, and that had a larger body mass were more likely to have been introduced successfully. Body mass was the only variable that was not a significant predictor of introduction success when included alone in logistic regression, although it was a significant predictor having controlled for phylogeny (Table 2).

Table 3.  Results of a forward selection multiple logistic regression model with introduction outcome as the response variable (success or failure) and the variables listed in Table 2 as predictors, for 52 bird species introduced to mainland Australia. Ordered = the change in deviance associated with adding each predictor variable to the model in the order of inclusion; Last = the change in deviance associated with adding each predictor variable to the model last, with the remaining predictors already included; coefficient = the estimated regression coefficients. For the continuous variables (climate suitability and body mass), the positive coefficients indicate that species with more area of suitable climate and species with larger body mass had a higher probability of successful introduction. For the categorical variables, a negative coefficient indicates that species in that category had a lower probability of successful introduction relative to species in the category with a 0 coefficient. The total deviance explained by the model, the residual deviance, and the associated degrees of freedom (d.f.) are also shown. The residual deviance is 0 which indicates that the model explains perfectly the patterns in introduction success
Predictor variabled.f.OrderedLastCoefficient
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

Intercept 1  −3696·3
Number of introduction sites 130·0***36·8*** 
 ≤ 2   −1239·5
 > 2       0
Gamebird 1 7·6**32·8*** 
 Yes   −1806·1
 No       0
Log10 (total points in ≤ 50% climate matching class) 1 6·4*20·3***    898·6
Introduced successfully elsewhere 1 5·5*20·0*** 
 No     −988·3
 Yes       0
Log10 (body mass) 118·8***18·8*** 759·4
Total deviance explained 568·3***  
Residual deviance46 0·0  

Remarkably, the forward selection model that included these five variables explained all of the variation in introduction success (Table 3). The results from the cross-validation procedure indicate how well the model performs in prediction: it correctly predicted the outcome for 47 of the 52 species introduced to Australia. Of the five species that were misclassified in cross-validation, three were predicted to succeed but failed (Branta canadensis (Linnaeus 1758), Numida meleagris (Linnaeus 1758) and Plectropterus gambensis (Linnaeus 1766), while two were predicted to fail but succeeded Lonchura punctulata (Linnaeus 1758) and Pavo cristatus).

Geographical range size

For analyses of geographical range size we excluded three predictor variables (human-modified habitat use, flocking behaviour, and whether or not birds were introduced successfully elsewhere) because in each case all but one of the successfully introduced birds used human-modified habitats, flocked or were introduced successfully elsewhere.

Availability of climatically suitable habitat

The observed and breeding range sizes were most strongly correlated with the number of points in the closest matched climate class (≤10%), both across-species and having controlled for phylogeny (Table 4, Table 5).

Table 4.  Results of across-species analyses (Pearson correlation coefficients, r , for the continuous variables or anovaF-values (Type 1 SS) for the categorical variables) comparing the log10-transformed observed and breeding range sizes in Australia with the variables in the left-hand column, for the 19 successfully introduced species
VariableLog10 (observed range size)Log10 (breeding range size)
rFrF
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

Total number of points in climate matching class (log10 transformed)
 ≤10%  0·762***  0·704*** 
 ≤20%  0·643**  0·573* 
 ≤30%  0·533*  0·469* 
 ≤40%  0·497*  0·430 
 ≤50%  0·519*  0·453 
 ≤60% 0·540*  0·483* 
 ≤70% 0·547*  0·472* 
 ≤80% 0·516*  0·442 
Life history traits
 log10 (body mass)−0·594** −0·651** 
 log10 (clutch size)−0·396 −0·388 
 log10 (broods per season) 0·543*  0·567* 
 log10 (incubation period)−0·543* −0·634** 
  plumage dichromatism 0·00 0·16
  migration 2·38 1·95
  diet 0·30 0·62
Historical factors
 log10 (overseas range size)    0·665**  0.636** 
 log10 (number of introduction events)    0·024  0.246 
 log10 (number of introduction sites)−0·128  0.073 
  number of individuals introduced 0·69 2·11
Table 5.  Results of analyses of phylogenetically independent contrasts (Pearson correlation coefficients, r, for the continuous variables or t-tests for the categorical variables) comparing the log10-transformed observed and breeding range sizes in Australia with the variables in the left-hand column, for the 19 successfully introduced species. n = the number of independent contrasts
Variablelog10 (observed range size)log10  (breeding range size)
rtrtn
  • *

    P < 0·05,

  • **

    P < 0·01,

  • ***

    P < 0·001.

Total number of points in climate matching class (log10 transformed)
 ≤10% 0·643**  0·624** 18
 ≤20% 0·526*  0·519* 18
 ≤30% 0·526*  0·535* 18
 ≤40% 0·560*  0·563* 18
 ≤50% 0·554*  0·542* 18
 ≤60% 0·545*  0·521* 18
 ≤70% 0·566*  0·513* 18
 ≤80% 0·582**  0·531* 18
Life history traits
 log10  (body mass)−0·315 −0·255 18
 log10 (clutch size)−0·167 −0·200 18
 log10   (broods per season) 0·380  0·484* 18
 log10  (incubation period)−0·231 −0·265 18
  plumage dichromatism  1·21  1·35 6
  migration  2·84*  1·91 6
  diet −0·16 −0·18 2
Historical factors
 log10  (overseas range size) 0·563*  0·569* 18
 log10 (number of introduction events) 0·215  0·336 18
 log10  (number of introduction sites) 0·088  0·188 18
  number of individuals introduced   
Life history traits and historical factors

Three life history traits were significantly correlated with observed and breeding range size across-species; more widely distributed birds tended to have small body masses, short incubation times and to produce more broods per season (Table 4). The correlations were generally weaker after controlling for phylogeny (Table 5).

Introduced species with a larger geographical range outside Australia tended to have a larger geographical range inside Australia. After controlling for phylogeny, introduced birds that were partial migrants in their natural range also had a significantly larger observed range size in Australia, although this relationship was not significant across-species. None of the remaining historical or life history variables significantly predicted Australian range size.

Modelling geographical range size

A forward selection model included two variables that explained 69% of the variation in geographical range size across-species: the number of target points in the ≤10% climate matching class and body mass (Table 6). Species with a greater area of closely matched climate and species with smaller body mass tended to have larger geographical ranges, although the area of closely matched climate alone could explain most of the variation (58%).

Table 6.  Results of a forward selection regression model with log10   (observed range size) as the response variable, for the 19 bird species successfully introduced to mainland Australia. The predictor variables in the left-hand column are shown in order of their inclusion in the model. Coefficient = the estimated regression coefficients (± one standard error); t and P = the t-value and the associated probability for each regression coefficient, respectively; partial r2 = the increase in r2 associated with successively adding each variable to the model
Predictor variableCoefficient (SE)tPartial r2
  • *

    P < 0·05,

  • ***

    P < 0·001.

log10 (total points in ≤10% climate matching class) 0·91 (0·218)4·2***0·58
log10 (body mass)–0·32 (0·136)2·4*0·11

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

Introduction success

Our results support the claim by Williamson (1996) that ‘propagule pressure, habitat matching and previous success at invasion’ are robust predictors of introduction success. For birds introduced to Australia, these three factors significantly predicted introduction success both across-species and having controlled for phylogeny, and all three factors emerged as independent predictors in our forward selection model (Table 3). Introduction effort (or propagule pressure) was the strongest predictor in our study and is consistently a strong predictor in other retrospective and experimental studies with data to quantify this effect (Crowell 1973; Ebenhard 1989; Griffith et al. 1989; Pimm 1991; Hopper & Roush 1993; Veltman et al. 1996; Williamson 1996; Duncan 1997; Green 1997; Memmott, Fowler & Hill 1998; Forsyth & Duncan 2001). Introduction effort is important because the introduction of more individuals to any one site lowers the chance of stochastic extinction (MacArthur & Wilson 1967), while introductions to multiple sites increase the probability that at least one population will escape stochastic extinction and will encounter habitat suitable for establishment (Crawley 1989). The probability of successful introduction will be enhanced if a species is both introduced to multiple sites and if a large proportion of those sites contain suitable habitat. This is precisely what we found: the number of introduction sites and the extent of climatically suitable habitat independently predict introduction success (Table 3).

Climate matching is reported to have mixed success in predicting introduction outcomes because there are examples of species that have been introduced successfully to areas outside their original climatic range (Mack 1996; Williamson 1996). Despite this, the degree of climate matching significantly predicted the success or failure of birds introduced to Australia. To our knowledge, this is the first time this hypothesis has been tested quantitatively (but see Nix & Wapshere 1986). However, climate matching alone would have failed to predict the success of several birds introduced to Australia. Repeating the cross-validation procedure using the number of target points in the ≤50% climate matching class as the only predictor results in 19 of the 52 introductions being misclassified, including 11 species predicted to fail that actually succeeded.

The low success rate in gamebirds, independent of other variables, suggests that an additional factor limited introduction success in this group. Several species of gamebird introduced to New Zealand have similarly failed to establish despite persistent efforts to introduce them (Heather & Robertson 1996). In Australia, one explanation for the high failure rate is human hunting. We excluded from our analysis one species of gamebird (Lophura nycthemera) whose population was destroyed by humans, and other gamebird populations may have been reduced by hunting. The one species that did succeed (peafowl, Pavo cristatus) is an ornamental species that is not typically hunted.

Overseas range size significantly predicted introduction success across-species but not having controlled for phylogeny or in multiple regression, suggesting that its predictive power derived from correlations with other traits. Wide overseas range could indicate flexible or generalist species that, relative to specialist species, have a high chance of success because they are likely to encounter conditions suitable for establishment (Williamson 1996). As we would predict if this were the case, introduced Australian birds with larger overseas range sizes had a greater area of climatically suitable habitat available in Australia (Spearman r = 0·35, n = 52, P = 0.01). However, widespread species will also be readily available for capture and may therefore be introduced more often and in greater numbers to any one location, increasing their success rate. This holds for birds introduced to Australia: overseas range size is significantly correlated with the number of introduction events, the total number of individuals introduced, and the number of introduction sites (Spearman r = 0·29, 0·44 and 0·30 respectively, n = 52, P all < 0·03). Hence, an indirect relationship between overseas range size and introduction success could arise via two different pathways, the first associated with habitat availability and the second associated with the introduction process.

Of the remaining traits that we considered, only diet showed a consistent relationship with introduction success both across-species and having controlled for phylogeny. Nevertheless, diet failed to explain, independently, significant variation in multiple regression, suggesting that its relationship with introduction success may derive from correlations with other traits. We found no evidence to support the hypotheses that species with a faster population growth rate, species that are non-migrants, species with dichromatic plumage and species that are gregarious were more likely to be introduced successfully.

Geographical range size in australia

A larger geographical range size in Australia is associated with a greater extent of climatically suitable habitat, suggesting that, at a broad scale, climate constrains the distribution of introduced birds. In European birds, changes in the timing (Crick et al. 1997) and success (Visser et al. 1998) of reproduction, and shifts in species northern range-margins (Thomas & Lennon 1999), have been related to changes in climate, mechanisms which could underlie a relationship between climate and geographical range size.

The availability of climatically suitable habitat also appeared to account for the strong but indirect relationship between overseas and Australian range size. Species with larger overseas range sizes had more climatically suitable habitat available in Australia and hence a larger Australian range size. Consequently, the number of target points in the closest climate matching class entirely accounted for the relationship between overseas range size and Australian range size in multiple regression.

Introduced species possessing traits associated with fast population growth rate (small body size, short incubation period and multiple broods per season) were significantly more widely distributed, and body mass remained a significant predictor of range size in multiple regression (Table 6), although these relationships were generally weaker having controlled for phylogeny. Similar relationships between range size and traits associated with population growth rate were found for introduced New Zealand birds and for British breeding birds (Duncan et al. 1999), highlighting a consistent pattern in range size variation.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

This is the first published study confirming that climatic suitability significantly predicts introduction success and the subsequent geographical range size of introduced birds. While many studies stress the unpredictability of biological invasions (Williamson 1999), the outcome of bird introductions to Australia were highly predictable, possibly because we included attributes that are typically unknown, such as introduction effort and climatic suitability, and that were important predictors of introduction outcomes. Further studies in which these factors can be quantified are required to test the generality of these conclusions.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix

We thank the Cowan Vertebrate Museum, Department of Zoology, University of British Columbia, for access to reference material, V. Varsanyi for help with collating the data and P. Pheloung for explaining the workings of the climate programme. R. Duncan was supported by the New Zealand Marsden Fund. Comments by R. Green and S. Albon improved the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix
  • Baicich, P.J. & Harrison, C.J.O. (1997) A Guide to the Nests, Eggs and Nestlings of North American Birds, 2nd edn. Academic Press, San Diego.
  • Bennett, P.M. (1986) Comparative studies of morphology, life history and ecology among birds. PhD Thesis, University of Sussex.
  • Blakers, M., Davies, S.J.J.F., Reilly, P.N. (1984) The Atlas of Australian Birds. Melbourne University Press, Melbourne.
  • Brown, J.H. (1989) Patterns, modes and extents of invasions by vertebrates. Biological Invasions: A Global Perspective (eds J.A.Drake, H.A.Mooney, F.Di Castri, R.H.Groves, F.J.Kruger, M.Rejmánek & M.Williamson), pp. 85109. John Wiley & Sons, Chichester.
  • Cramp, S. & Simmons, K.E.L. (197794) The Birds of the Western Palearctic,Vols 1–9. Oxford University Press, Oxford.
  • Crawley, M.J. (1989) Chance and timing in biological invasions. Biological Invasions: A Global Perspective (eds J.A.Drake, H.A.Mooney, F.Di Castri, R.H.Groves, F.J.Kruger, M.Rejmánek & M.Williamson), pp. 407435. John Wiley & Sons, Chichester.
  • Crick, H.Q.P., Dudley, C., Glue, D.E., Thomson, D.L. (1997) UK birds are laying eggs earlier. Nature, 388, 526.DOI: 10.1038/41453
  • Crowell, K.L. (1973) Experimental zoogeography: introductions of mice to small islands. American Naturalist, 107, 535558.
  • Duncan, R.P. (1997) The role of competition and introduction effort in the success of passeriform birds introduced to New Zealand. American Naturalist, 149, 903915.
  • Duncan, R.P., Blackburn, T.M., Veltman, C.J. (1999) Determinants of geographical range sizes: a test using introduced New Zealand birds. Journal of Animal Ecology, 68, 963975.DOI: 10.1046/j.1365-2656.1999.00344.x
  • Dunning, J.B.J. (1993) CRC Handbook of Avian Body Masses. CRC Press, Boca Raton, Florida.
  • Ebenhard, T. (1989) Bank vole [Clethrionomys glareolus (Schreber, 1780)] propagules of different sizes and island colonization. Journal of Biogeography, 16, 173180.
  • Felsenstein, J. (1985) Phylogenies and the comparative method. American Naturalist, 125, 115.
  • Forsyth, D.M. & Duncan, R.P. (2001) Propagule size and the relative success of exotic ungulate and bird introductions to New Zealand. American Naturalist, 157, 583595.
  • Frith, H.J., McKean, J.L., Braithwaite, L.W. (1976) Sexual cycles and food of the doves Streptopelia chinensis and S. senegalensis in Australia. Emu, 76, 1524.
  • Fry, C.H., Keith, S., Urban, E.K. (1988) The Birds of Africa,Vol. 3. Academic Press, San Diego.
  • Garland, T., Harvey, P.H., Ives, A.R. (1992) Procedures for the analysis of comparative data using phylogenetically independent contrasts. Systematic Biology, 41, 1832.
  • Gaston, K.J. (1988) Patterns in the local and regional dynamics of moth populations. Oikos, 53, 4957.
  • Goodwin, D. (1982) Estrildid Finches of the World. Oxford University Press, London.
  • Goodwin, D. (1983) Pigeons and Doves of the World. Cornell University Press, Ithaca.
  • Goodwin, D. (1986) Crows of the World, 2nd edn. British Museum, London.
  • Green, R.E. (1997) The influence of numbers released on the outcome of attempts to introduce exotic bird species to New Zealand. Journal of Animal Ecology, 66, 2535.
  • Griffith, B., Scott, J.M., Carpenter, J.W., Reed, C. (1989) Translocation as a species conservation tool: status and strategy. Science, 245, 477480.
  • Hanski, I. (1991) Single-species metapopulation dynamics: concepts, models and observations. Biological Journal of the Linnean Society, 42, 1738.
  • Hanski, I., Kouki, J., Halkka, A. (1993) Three explanations of the positive relationship between distribution and abundance of species. Species Diversity in Ecological Communities (eds R.E.Ricklefs, & D.Schluter), pp. 108116. University of Chicago Press, Chicago.
  • Heather, B.D. & Robertson, H.A. (1996) The Field Guide to the Birds of New Zealand. Viking, Auckland.
  • Heinzel, H., Fitter, R., Parslow, J. (1984) The Birds of Britain and Europe with North Africa and the Middle East. Collins, London.
  • Holt, R.D., Lawton, J.H., Gaston, K.J., Blackburn, T.M. (1997) On the relationship between range size and local abundance: back to basics. Oikos, 78, 183190.
  • Hopper, K.R. & Roush, R.T. (1993) Mate finding, dispersal, number released, and the success of biological control introductions. Ecological Entomology, 18, 321331.
  • Del Hoyo, J., Elliott, A., Sargatal, J. (199297) Handbook of the Birds of the World,Vols 1–4. Lynx Editions, Barcelona.
  • Johnsgard, P.A. (1986) The Pheasants of the World. University of Nebraska, Lincoln.
  • Legendre, S., Clobert, J., Møller, A.P., Sorci, G. (1999) Demographic stochasticity and social mating system in the process of extinction of small populations: the case of passerines introduced to New Zealand. American Naturalist, 153, 449463.
  • Lever, C. (1994) Naturalized Animals. Cambridge University Press, Cambridge.
  • Livezey, B.C. (1997) A phylogenetic classification of waterfowl (Aves: anseriformes), including selected fossil species. Annals of the Carnegie Museum, 66, 457496.
  • Lockwood, J.L. (1999) Using taxonomy to predict success among introduced avifauna: relative importance of transport and establishment. Conservation Biology, 13, 560567.DOI: 10.1046/j.1523-1739.1999.98155.x
  • Long, J. (1981) Introduced Birds of the World. David and Charles, London.
  • MacArthur, R.H. & Wilson, E.O. (1967) The Theory of Island Biogeography. Princeton University Press, Princeton.
  • Mack, R.N. (1996) Predicting the identity and fate of plant invaders: emergent and emerging properties. Biological Conservation, 78, 107121.DOI: 10.1016/0006-3207(96)00021-3
  • Maclean, G.L. (1993) Robert’s Birds of Southern Africa, 6th edn. CTP Book Printers, Cape Town.
  • Martin, T.E. & Clobert, J. (1996) Nest predation and avian life-history evolution in Europe versus North America: a possible role of humans? American Naturalist, 147, 10281046.
  • Mayr, E. (1965) The nature of colonizations in birds. The Genetics of Colonizing Species (eds H.G.Baker & G.L.Stebbins), pp. 2943. Academic Press, New York.
  • Memmott, J., Fowler, S.V., Hill, R.L. (1998) The effect of release size on the probability of establishment of biological control agents: gorse thrips (Sericothrips staphylinus) released against gorse (Ulex europaeus) in New Zealand. Biocontrol Science and Technology, 8, 103115.DOI: 10.1080/09583159830478
  • Moulton, M.P. & Pimm, S.L. (1986) Species introductions to Hawaii. Ecology of Biological Invasions of North America and Hawaii (eds H.A.Mooney, & J.A.Drake), pp. 231249. Springer, New York.
  • Newsome, A.E. & Noble, I.R. (1986) Ecological and physiological characters of invading species. Ecology of Biological Invasions (eds R.H.Groves, & J.J.Burdon), pp. 120. Cambridge University Press, Cambridge.
  • Nix, H.A. (1986) A biogeographic analysis of Australian elapid snakes. Atlas of Australian Elapid Snakes (ed. R.Longmore), pp. 415. Bureau of Flora and Fauna, Canberra.
  • Nix, H.A. & Wapshere, A.J. (1986) Biogeographic origins of invading species. Ecology of Biological Invasions (eds R.H.Groves, & J.J.Burdon), p. 155. Cambridge University Press, Cambridge.
  • Pheloung, P.C. (1996) climate: a system to predict the distribution of an organism based on climate preferences. Agriculture Western Australia, Perth.
  • Pimm, S.L. (1991) The Balance of Nature? Chicago University Press, Chicago.
  • Potts, G.R. (1986) The Partridge. Pesticides, Predation and Conservation. Collins, London.
  • Purvis, A. & Rambaut, A. (1994) Comparative Analysis by Independent Contrasts (CAIC), version 2. Oxford University, Oxford.
  • Purvis, A. & Rambaut, A. (1995) Comparative analysis by independent contrasts (CAIC): an Apple Macintosh application for analysing comparative data. Computer Applications in the Biosciences, 11, 247251.
  • Reichard, S.H. & Hamilton, C.W. (1997) Predicting invasions of woody plants introduced into North America. Conservation Biology, 11, 193203.DOI: 10.1046/j.1523-1739.1997.95473.x
  • Rejmánek, M. (1996) A theory of seed plant invasiveness: the first sketch. Biological Conservation, 78, 171181.DOI: 10.1016/0006-3207(96)00026-2
  • Savidge, J.A. (1987) Extinction of an island forest avifauna by an introduced snake. Ecology, 68, 660668.
  • Shurtlett, L.L. & Savage, C. (1996) The Wood Duck and the Mandarin. University of California Press, Berkeley.
  • Sibley, C.G. & Ahlquist, J.E. (1990) Phylogeny and Classification of Birds: A Study in Molecular Evolution. Yale University Press, New Haven.
  • Sibley, C.G. & Monroe, B.L. (1990) Distribution and Taxonomy of Birds of the World. Yale University Press, New Haven.
  • Sorci, G., Møller, A.P., Clobert, J. (1998) Plumage dichromatism of birds predicts introduction success in New Zealand. Journal of Animal Ecology, 67, 263269.
  • Thomas, C.D. & Lennon, J.J. (1999) Birds extend their range northwards. Nature, 399, 213.DOI: 10.1038/20335
  • Urban, E.K., Fry, C.H., Keith, S. (1986) The Birds of Africa,Vol. 2. Academic Press, London.
  • Veltman, C.J., Nee, S., Crawley, M.J. (1996) Correlates of introduction success in exotic New Zealand birds. American Naturalist, 147, 542557.
  • Visser, M.E., Van Noordwijk, A.J., Tinbergen, A.J., Lessells, C.M. (1998) Warmer springs lead to mistimed reproduction in great tits (Parus major). Proceedings of the Royal Society of London B, 265, 18671870.
  • Williamson, M. (1996) Biological Invasions. Chapman & Hall, London.
  • Williamson, M. (1999) Invasions. Ecography, 22, 512.

Received 15 May 2000; revision received 27 February 2001

Appendix

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Appendix
Table Appendix1..  Bird species introduced to mainland Australia. Nomenclature follows Sibley & Monroe (1990). AUS, species was successfully (1) or unsuccessfully (0) introduced to Australia; A, ‘observed’ range size in Australia (1° blocks); B, ‘breeding’ range size in Australia (1° blocks); C, female body mass (g); D, plumage dichromatism, scored as no (0) or yes (1); E, migration, scored as sedentary (1), sedentary and migratory (2), or migratory (3); F, flocking, scored as no (0) or yes (1); G, diet, scored as herbivorous (1), omnivorous (2), or carnivorous (3); H, clutch size; I, number of broods per season; J, days to incubate; K, do not (0) or do (1) use human-dominated habitats; L, overseas range size (×106 km2); M, minimum number of individuals introduced; N, minimum number of introduction events; O, minimum number of introduction sites; P, successfully introduced elsewhere, scored as yes (1) or no (0). n.d. indicates no data available
SpeciesAUSABCDEFGHIJKLMNOP
Struthionidae
 Struthio camelus1  1  199501112 7145·50  8> 500 4 50
Odontophoridae
 Callipepla californica0   .     . 162111114·22n.d.1  3      267 6 51
Numididae
 Numida meleagris0   .     .13880112 81n.d.1 20  170 3 61
Phasianidae
 Alectoris barbara0   .     . 3760112131251  3    2 1 11
 Alectoris gracea0   .     . 501011215·51250 23   48 5 21
 Alectoris rufa0   .     . 439011211·2223·51  1    9 1 11
 Gallus gallus0   .     . 8501112 5·51191  6    4 2 21
 Lophura ignita0   .     .16001102 61240  1    2 1 11
 Lophophorus impejanus0   .     .20001112 4·51270  1    2 1 10
 Pavo cristatus1  2  233751112 6125·51  4   3310 91
 Perdix perdix0   .     . 386111214·6124·51 32    4 2 21
 Phasianus colchicus0   .     . 850110211·81n.d.1 34  75012121
Anatidae
 Aix galericulata0   .     . 6351212 91291  3    2 1 11
 Anas platyrhynchos1 81 131020121212·6127·51 95   86 4 31
 Branta canadensis0   .     .43900211 5·91291 38    4 2 11
 Cygnus olor1  1  196000112 61361 11   12 5 51
 Plectropterus gambensis0   .     .4700011211131·51 16    4 2 20
Psittacidae
 Agapornis roseicollis0   .     .  55·50111 3·5123·31  2    3 2 10
Columbidae
 Columba livia1304 89 2960112 2517·51 80>  500 2 21
 Oena capensis0   .     .  39·51111 22·514·51 20        2 1 11
 Streptopelia chinensis1 89 31 160·50102 25141 11      5710 71
 Streptopelia senegalensis1 34 13  990112 2·55·5131 44        4 3 31
 Streptopelia turtur0   .     . 1320211 22·513·51 34    8 1 10
Pteroclidae
 Pterocles exustus0    .     . 2481112 2·53231 15    2 1 10
Ardeidae
 Bubulcus ibis1172 14 3570213 3·5223·71 54   18 1 11
Corvidae
 Corvus splendens0    .     . 3070112 4·5116·51  5   53 4 21
Muscicapidae
 Erithacus rubecula0    .     .  15·80202 52141 25   47 3 10
 Luscinia megarhynchos0    .     .  19·40302 4·72131 16    4 1 10
 Turdus merula1117 78  82·61212 3·8313·51 31  102 8 41
 Turdus philomelos1   4  4  67·30212 4·7211·51 31  129 8 51
Sturnidae
 Acridotheres tristis1 55 20 111·30112 3·7117·51 11  27711 91
 Sturnus vulgaris1246157  79·81212 4·8214·51 58  292 9 41
Pycnonotidae
 Pycnonotus cafer0     .     .  35·50112 33141  6    4 2 21
 Pycnonotus jocosus1 11  4  32·50112 33131  6    8 3 31
Alaudidae
 Alauda arvensis1 88 37  38·90212 3·7311·51 57  52613 81
Passeridae
 Euplectes albonotatus0     .     .  17·61112 2·51130  6    2 1 10
 Euplectes orix0     .     .  24·51212 3·5212·50 13    4 2 21
 Lonchura malacca0     .     .  12·50111 5·5312·51  7    2 1 11
 Lonchura punctulata1 41 19  13·50111 53161  9    8 4 41
 Padda oryzivora0     .     .  24·80111 5313·51  1  255 2 21
 Passer domesticus1330185  28·81112 3·9312·51110  41420201
 Passer montanus1 20 14  220112 4·73131 70   68 4 31
Fringillidae
 Carduelis cannabina0     .     .  18·51211 4·7213·51 25   32 4 40
 Carduelis carduelis1123 77  15·50211 4·43131 31  223 5 41
 Carduelis chloris1 55 23  291212 4·82131 21  133 7 41
 Carduelis spinus0     .     .  121311 42131 22   80 3 10
 Emberiza citrinella0     .     .  28·21212 3·33131 47   34 3 31
 Emberiza hortulana0     .     .  21·61312 5111·51 30   16 1 10
 Fringilla coelebs0     .     .  23·51211 4·9212·51 27  498 4 21
 Fringilla montifringilla0     .     .  21·41312 62141 31   78 1 10
 Pyrrhula pyrrhula0     .   23·51111 43141 41   14 1 10
 Serinus canarius0     .     .  16·51111 4212·51  0·1   20 2 21