1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Habitat loss, fragmentation and degradation are drivers of major declines in biodiversity and species extinctions. The actual causes of species population declines following habitat change are more difficult to discern and there is typically high covariation among the measures used to infer the causes of decline. The causes of decline may act directly on individual fitness and survival, or through disruption of population processes. We examined the relationships among configuration, extent and status of native vegetation and three commonly used indicators of individual body condition and chronic stress (haemoglobin level, haematocrit, residual body mass condition index) in 13 species of woodland-dependent birds in south-eastern Australia. We also examined two measures of changes to population processes (sex ratio and individual homozygosity) in ten species and alleic richness in five species. We found little support for relationships between site or landscape characteristics and individual or population response variables, notwithstanding that our simulations showed we had sufficient power to detect relatively small effects. We discuss possible causes of the absence of detectable habitat effects in this system and the implications for the usefulness of individual body condition and easily measured haematological indices as indicators of the response of avian populations to habitat change.

Habitat loss and fragmentation are well established as the most prevalent causes of anthropogenically induced bio diversity loss through local and global population decline and extinction (Fahrig 2003, World Resources Inst. 2005). However, the identification of the mechanisms of these negative effects has proven more difficult to establish. Species’ responses to loss and degradation of habitat differ greatly (Mönkkönen and Reunanen 1999, Bennett and Radford 2009) because a wide range of life-history traits can be affected by fragmentation (Banks et al. 2007). There is often covariation of many landscape and site attributes (e.g. habitat clearance, fragmentation of remnants and decrease in mean remnant size and remnant structure and condition) that may influence whether an area can support a population (Saunders et al. 1991, Yates and Hobbs 1997, Ewers and Didham 2005, Lindenmayer and Luck 2005, Radford et al. 2005). Despite this problem, biodiversity protection and restoration require the teasing apart of the processes underlying biodiversity declines that are the consequence habitat loss, fragmentation and degradation (Lindenmayer and Fischer 2007). There is a considerable tension in the literature between those emphasising the importance of extrinsic and stochastic factors in determining population decline following habitat loss (Caughley 1994), and those who argue for an important, though often difficult to detect, role for intrinsic factors (such as individual condition and genetics) in population declines (Arcese 2003). Species declines may result directly from loss of habitat or indirectly through changes in population processes due to habitat fragmentation. Vegetation structure is a common determinant of avian diversity and individual species’ habitat preferences (Rotenberry 1985, Mac Nally 1990). Human-induced changes to vegetation structure decrease habitat suitability for some species while increasing it for others (Lindenmayer et al. 2008). Reduction in the size and increase in edge ratio of patches in fragmented forests and woodlands can lead to decreased vegetation condition (Yates and Hobbs 1997), elevated predation and competition and reduced food availability for woodland-dependent birds (Andren 1992, Zanette et al. 2000, Huhta et al. 2004, Maron et al. 2011).

The effects of clearing on habitat quality stem from the selective clearing of more productive parts of the landscape for agriculture (Vesk and Mac Nally 2006). In the woodlands of south-eastern Australia, the majority of the remnants are in areas of low primary productivity, and have often been heavily grazed, with much of the ground layer and understorey degraded. The lower productivity of the remnants lead to reduced food resouces for insectivorous birds (Watson 2011). The avifaunas of dry woodland systems of southern Australia continue to decline, due primarily to habitat loss compounded by a range of other factors (Robinson and Traill 1996, Ford et al. 2001, Mac Nally et al. 2009, Ford 2011), although the mechanisms generating these declines remain unclear.

The physiological status of individual birds may offer an insight into processes that vary in response to extent, configuration and condition of remnant habitat. Haematological and morphological measures have been used for assessment of individual condition (Norte et al. 2009a). Whole blood haemoglobin levels (Hb) and haematocrit (the ratio of packed blood cells to total blood volume, HCT) have been used to assess condition and physiological response in relation to habitat and to individual behaviour, such as reproductive investment and exercise levels (Campbell 1995). These measures have also been related to the effects of environmental stressors including parasite load, food availability and environmental toxins (Acquarone et al. 2002, Dudaniec et al. 2006, Linkie et al. 2006). Residual body mass (RBM), a measure of mass that accounts for structural size, is frequently used as an index of ‘body condition’ in ecological studies (Schulte-Hostedde et al. 2005, Stevenson and Woods 2006). It reflects variation in stored fuel reserves, particularly lipids, (Seewagen 2008), which have been shown to influence individual inclusive fitness in some birds (Ardia 2005).

HCT, Hb and RBM have been used to assess effects of environmental variation, including habitat fragmentation and habitat quality (and related food availability), on individual condition in wild passerines (Hõrak et al. 1998, Strong and Sherry 2000, Mazerolle and Hobson 2002) and small mammals (Johnstone et al. 2011). These three measures differ between sexes, and with reproductive status, age and season (Norte et al. 2009a), while RBM also varies with moult (Bojarinova et al. 1999). It is necessary to account for these covariates when investigating relationships between physiological condition and habitat. Where a relationship is found, further work is required to determine causality. The relationship may represent a direct effect of habitat on individual physiology, alternatively, individual condition may influence settlement choice (Porlier et al. 2009) or lead to competitive exclusion of individuals in poorer condition from favoured habitat (Latta and Faaborg 2002).

There has been much work assessing the impacts of changes in landscape composition and configuration on individual movement, gene flow and population genetics (Manel et al. 2003, Storfer et al. 2007). Less attention has been given to the influence of landscape characteristics on individual genetic diversity, despite the important role that this quantity plays in evolutionary processes (Porlier et al. 2009). Population processes such as mating systems may be affected by changes in landscape structure and habitat condition (Banks et al. 2007) that could be reflected in individual heterozygosity levels (García-Navas et al. 2009). Heterozygosity may be positively associated with offspring fitness, reproductive success, local survival and recruitment into the adult population (Coulson et al. 1998, 1999, Coltman et al. 1999, Amos et al. 2001, Hansson et al. 2001, Banks et al. 2010) though meta-analysis of such heterozy gosity-fitness correlations suggests that the effects are usually weak (Chapman et al. 2009). Such effects need not be restricted to sessile organisms; where mobile or dispersing individuals assess habitat quality before settling, fitter individuals (with higher individual heterozygosity) may choose and be able to defend higher quality territories (Seddon et al. 2004).

In addition to the individual-based responses to habitat alteration outlined above, disruption of natural patterns of mobility can lead to changes in population parameters, such as sex ratios or genetic diversity, with downstream consequences for individual and population fitness (Banks et al. 2007). For example, disrupted dispersal of the usual dispersing sex, females, in the brown treecreeper Climacteris picumnus in Australian woodlands, has been implicated in low female recruitment, isolated patches containing no females, and local patch extirpation (Cooper and Walters 2002, Cooper et al. 2002a, b). Also, reduction in dispersal and gene flow, recent bottlenecks and/or disruptions of mating systems may lead to decreased levels of population genetic diversity (Palstra and Ruzzante 2008). Thus associations of sex ratios and genetic diversity (measured by allelic richness; AR) with landscape conditions may provide evidence of important responses to landscape alteration.

The dry woodlands of south-eastern Australia have suffered considerable habitat clearance and degradation and there has been a corresponding and ongoing decline of the region's avifauna (Robinson and Traill 1996, Ford et al. 2001, Mac Nally et al. 2009). A pattern of disproportionately large decline in incidence in apparently suitable remnant habitat in many woodland-dependent birds compared to decline in landscape tree-cover has been documented (Radford et al. 2005, Radford and Bennett 2007). Species showing this pattern of disproportionate decline have been termed ‘decliner’, while those that show no relationship of incidence landscape tree-cover have been termed ‘tolerant’ (Bennett and Radford 2009, Amos et al. 2012).

In this study, we examine the relationships of landscape structure and habitat condition with physiological status, individual and population genetic diversity and local sex ratio to explore whether these might be mechanisms underpinning some of the decline of resident woodland birds of south-eastern Australia. Specifically, we explore the possibility that there are impacts of landscape and site attributes on individual physical condition and heterozy gosity, or on population genetic diversity and local sex ratios, that may be contributing to the observed pattern of decline through reduced individual and population condition and reproductive output, disruption of population processes, fitness and function (Hõrak et al. 1998, Kilgas et al. 2006).

We predicted that, if landscape and/or site condition are contributing to the decline of woodland-dependent birds in the study area, evidence of a relationship with RBM, HCT, or Hb should be found in the decliner species, and not in the tolerant species. Site condition is expected to affect sedentary species (those that stay in the same home range year round) more strongly than mobile ones, which may move locally or regionally between areas of varying condition. With regards to the effects of landscape and site on homozygosity-by-locus (HL), AR or sex ratio, evidence of differences related to landscape and site quality would support the hypothesis that population or social processes have been disrupted by change in habitat confi guration or quality (Banks et al. 2007). Relationships between site and landscape variables and the response variables may also be due to condition-dependent settlement patterns. Nevertheless the existence of differences in response variables relating to anthropogenic habitat change would be evidence of disruption of the birds’ interaction with their ecosystem.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Site selection

This study built on the work of Radford et al. (2005), which examined the incidence of 58 species of woodland-dependent birds in 24 landscapes, each 10 × 10 km2, in the woodlands of the box-ironbark region of central Victoria, Australia. Twelve 10 × 10 km landscapes were selected, nine of which were used by Radford et al. (2005). Tree-cover in these nine ranged from 10 to 50% and vegetation configuration in each was ‘dispersed’ or ‘aggregated’ (Radford et al. 2005). Three ‘reference’ landscapes were also selected. These had the highest available extent of tree-cover (72–78%) to represent as near as possible the historical condition of continuous tree-cover (Fig. 1). Within all twelve landscapes, three to six sampling sites were selected to a total of 63 sites (Fig. 1).


Figure 1. The study area in central Victoria, Australia, showing landscapes, sampling sites and remnant tree cover (shaded). Values for landscape treecover are: landscapes with aggregated tree cover; Shelbourne 12%, Glenalbyn 17%, Tunstalls 20%, Crosbie 26%, Havelock 45%. Landscapes with dispersed tree cover; Welha 11%, Stuart Mill 19%, Murchison 27%, Axe Creek 35%. Landscapes with continuous tree cover; Redcastle 75%, Dunolly 79%, Rushworth 79%.

Download figure to PowerPoint

Study species and sampling

We sampled thirteen species of small to medium-sized resident woodland-dependent passerines. These species were the most frequently captured in the study area and represented a range of mobility, and patterns of reduced incidence with respect to landscape scale tree-cover, either ‘decliner’, or ‘tolerant’ species (Table 1).

Table 1.  Classification of studied species according to their response to tree-cover and their expected mobility. Species that were sexed and genotyped are highlighted in bold. Table modified from Amos et al. (2012).
 Response to landscape tree-cover
  1. 1For mobility, the term ‘inconclusive’ is used where there is uncertainty about mobility levels from the literature (Higgins et al. 2001, Higgins and Peter 2002).

Mobile Fuscous honeyeater Lichenostomus fuscus White-plumed honeyeater Lichenostomus penicillatus
 Dusky woodswallow Artamus cyanopterus 
 Brown headed honeyeater Melithreptus brevirostris 
Moderate/inconclusive1 Yellow-tufted honeyeater Lichenostomus melanops Striated pardalote Pardalotus striatus
  Spotted pardalote Pardalotus punctatus  
  Grey shrike-thrush Colluricincla harmonica  
  Weebill Smicoronis brevirostris  
Sedentary/inconclusive1 Eastern yellow robin Eopsaltria australis  
  Superb fairy-wren Malurus cyaneus  
 Buff-rumped thornbill Acanthiza reguloides 
Sedentary Brown treecreeper Climacteris picumnus  

Fifty-seven sites were each visited twice in different seasons, with 4–7 months between visits. Sites were chosen opportunistically where there was a sufficient population of a number of the focal species for sampling to be viable. Six sites, where very few birds or only a single species were caught were not revisited. There were two days of sampling on each visit, between sunrise and 1 h before sunset. Sampling occurred between November 2007 and February 2010, with the sampling of each landscape spread over this period. Birds were captured using 12 and 18 m 31 mm-mesh mist-nets 0.5–3 m above ground. A secondary capture technique was used for ground-feeding birds. Forty spring-loaded net traps (Reilly 1968) baited with mealworms and remotely monitored using UHF digital transmitters (Embedded Communications Systems, Launceston, Tasmania) were used mainly to catch eastern yellow robin Eopsaltria australis and grey shrike-thrush Colluricincla harmonica.

Bird attributes

Birds were aged and sexed from plumage and morphological characteristics following Rogers et al. (1986); the number of distinguishable classes varied from 2 (juvenile or adult plumage) to four (juvenile, hatching, second and third year and older) according to species. For ten species, sex was confirmed genetically (below). Brown-headed honeyeater Melithreptus brevirostris; buff-rumped thornbill Acanthiza reguloides, and dusky woodswallow Artamus cyanopterus were not sexed because they were not subject to a program of genotyping, and they could not be reliably sexed by plumage characters or other external characters (Rogers et al. 1986).

We measured total head-plus-bill length, bill depth at the base of the bill, tarsus length and wing chord to the nearest mm. Multiple blood samples (5–50 μl each) per bird were taken following brachial venipuncture with a 27-gauge hypodermic needle. Samples were collected into a heparinised microcapillary tube for HCT, a non-heparinised tube for genetic sampling and a directly into a cuvette (Hemocue, Ängeholm, Sweden) for Hb. Blood samples for genetic analysis were transferred immediately into 1 ml of ethanol at ambient temperature in the field and 220°C on return to the laboratory.

The HCT sample was spun for 4.5 min at 12 000 r.p.m. in a Zipocrit portable centrifuge (LW Scientific, Atlanta, GA, USA) with a haematocrit rotor. Total blood column length and packed cell length in the microcapillary tube were measured to ± 0.5 mm.

Whole-blood haemoglobin concentration was measured with a Hemocue 2001 B-Haemoglobin photometer (Hemocue, Ängeholm, Sweden) immediately after sampling. The value (g dl−1) obtained by this method is higher than would be obtained by the standard cyanometh haemoglobin methodology for avian blood (Eklom and Lill 2006, Simmons and Lill 2006). Given that our interest was in relative levels of Hb within species, the difference was inconsequential.

Ten species were genotyped for 6–16 polymorphic length-variable nuclear loci per species with a mean of 9.7 alleles/locus (SD = 8.0) (Supplementary material Appendix 1). Assessment of all loci for departures from Hardy–Weinberg and linkage equilibria, sex-linkage and null alleles showed that the genetic markers conformed to Mendelian expectations for codominant, autosomal loci without significant null allele frequencies (Harrisson et al. 2012, in press, Pavlova et al. 2012). Homozygosity-by- locus (HL), which offers an efficient estimate of individual genetic diversity in populations with migration and admixture (Aparicio et al. 2006, Coulon 2010), was calculated across the pooled sample for each species using the R statistical package GenHet (Coulon 2010). Mean Allelic richness (AR) was calculated using R package HierFstat (Goudet 2005) for samples of > 5 individuals for each species mean AR, was rarefied to the smallest included sample for a site.

Genotyped birds were sexed using sex-linked chromosome-helicase-DNA binding protein 1 (CHD1) gene. This gene has different-sized introns on the Z and W chromosomes, allowing homogametic (ZZ) males to be distinguished from heterogametic (ZW) females (Griffiths et al. 1998). PCRs for six species, spotted pardalote, weebill, eastern yellow robin, white-plumed honeyeater, yellow-tufted honeyeater and fuscous honeyeater, were run separately as described in Harrisson et al. (2012, in press). For striated pardalote, grey shrike-thrush, superb fairy wren and brown treecreeper sexing reaction was incorporated into multiplex PCR (Pavlova et al. 2012, Harrisson et al. in press). Individuals of known sex (Australian National Wildlife Collection samples) were used as positive controls on first gels for each species. Each scored gel had clearly detectable Z and W bands. Polymorphism within CHD- Z and/or CHD-W, when detected, did not confound sex determination, as difference between CHD-Z and CHD-W alleles was always much greater than that of two CHD-Z (Pavlova et al. unpubl.).

Sample size

For some individuals, insufficient blood was collected to enable Hb or HCT to be measured, or processing time limited data collected. Hb samples were measured for 2525 and HCT for 2505 birds. For RBM (2239 individuals), only sexed adult birds (for sexed species) were analysed whereas all adults were included for buff-rumped thornbill, brown-headed honeyeater and dusky wood swallow. Sample sizes, and numbers of sites and landscapes where each species was sampled are given in Table 2.

Table 2.  Sample size for each response variable in each species for adults and total by sex and number of sites and landscapes where each species was sampled.
 HaematocritTotal blood haemoglobinResidual body mass  
SpeciesAdultAllAdultAllAdultAllAdultAllAdultAllAdultAllAdultAllAdultAllMale AdultFemale AdultUnsexe dAdultTotal AdultNo. sitesNo. landscapes
Brown treecreeper158207189235  347442173224199243  372467175206 3814812
Eastern yellow robin33394960  829933405263  851034363 1063212
Fuscous honeyeater116158192237  308395115163185230  300393142218 3604112
Grey shrike-thrush20383351  538921383249  53872032 523912
Superb fairy wren52597480  12613957666773  1241396393 1563312
Spotted pardalote16351941  357612331639  28721713 30139
Striated pardalote419066115  107205378272118  1092005378 1313212
Weebill772021222930101018191 29292045 652211
White-plumed honeyeater737720121296125370414778120922110012638642881224 3054011
Yellow-tufted honeyeater128160222251 1350412121157216244 2337403144249 3932811
Brown-headed honeyeater    82978297    83978397  104104209
Buff-rumped thornbill    53535353    54535453  92923312
Dusky woodswallow    54545454    51545154  64642411

Landscape and site-condition attributes

Landscape attributes included in the analysis were percent tree-cover (Dept of Sustainability and Environment 1990–1999) calculated in ARCGIS (Environmental Systems Research Inst. 1999–2008) and configuration of tree-cover: classified as aggregated, dispersed (Radford et al. 2005) or continuous (for the three highest cover landscapes). Site vegetation condition attributes were assessed using the Habitat Hectares methodology (Parkes et al. 2003), with raw data for each habitat component recorded. The Habitat Hectares method uses a measure of ‘deviation’ of the extant vegetation from an idealized structure for the vegetation at that location (Parkes et al. 2003). The area assessed at each site was the minimum convex polygon that included all net and trap locations. All vegetation condition assessments were carried out in the same season (17 August 2009–25 September 2009) and by the same individual (G. Sutter) to avoid seasonal and observer differences in assessments (Gorrod and Keith 2009). Overall site condition scores (which have a possible value of 0–75 (Parkes et al. 2003)) ranged from 12–52 (mean 41, SD 7.3).

A subset of these site condition data was used to calculate three variables that we believed most likely to be related to individual bird condition. These were: CANOPY, projected tree canopy cover (range 10–30%, mean 19, SD 4.1); SHRUBS, the sum of projected cover of small (< 1 m) medium (1–5 m) and large (> 5 m) shrubs (range 0–58%, mean 27, SD 15.2); and LOGS, the length of fallen logs ha21 (0–173 m ha−1, mean 43, SD 35.3). LOGS value was log10-transformed to improve normality of its distribution. Fallen logs form an important foraging resource for several of the target species, particularly the brown treecreeper (Noske 1979, Doerr et al. 2006).

Landscape context was used as a measure of connectedness of the vegetation at the site (Dept of Sustainability and Environment, VIC, NV2005_CON, unpubl. GIS raster dataset, accessed May 2011). Landscape context is a single index of the distance of a site from a large block of remnant vegetation and weighted vegetation cover within radii of 1, 3 and 5 km of the site (Ferwerda 2003, Parkes et al. 2003). Landscape context had a possible range of 0–100; values for the study sites were from 50 to 99 (mean 87, SD 11.4).


Principal component analysis (PCA) of wing chord, tarsus-length, total head length and bill depth was conducted on standardized (zero mean, unit variance) values in R 2.13 (R Development Core Team). A linear model of the first PCA regressed against mass was fitted. The difference of actual mass compared to the model was residual body mass (RBM). This analysis was carried out separately for each species, and for each sex in each species where sex was determined, as the relationship between size and RBM might differ between sexes (Green 2001).

We used generalised linear mixed models (GLMM) to assess evidence for effects of landscape and site variables. GLMM is useful for the analysis of data that are not normally distributed and where there are multiple random effects in large datasets (Bolker et al. 2009). The predictor variables were three landscape components (tree-cover, landscape context, and aggregation) and three components of vegetation condition at sampling sites (CANOPY, SHRUBS, and LOGS, as above). Models initially incorporated spline functions (Lunn et al. 2009) to accommodate non-linear predictor effects, but we found no evidence of non-linear relationships and therefore present results of linear models only. Sine and cosine of ordinal day were included as fixed variables to account for seasonal variation in Hb, HCT and RBM. Both sine and cosine were used to contrast summer/winter and spring/autumn differences. The models fitted for Hb, HCT and RBM also included moult (presence or absence) as a random variable. Sex was included as a random effect for HCT and Hb for species where it was known. All models included landscape and site identity and year of capture as random variables. We used Gaussian errors for RBM, HCT, Hb, HL and AR models and a Binomial model structure for sex ratio (expressed as proportion of the more frequent sex).

We used Bayesian model selection with reversible jump Markov chain Monte Carlo (MCMC) sampling in WinBUGS (Lunn et al. 2000, 2009) to identify landscape and site factors that were associated with each response variable, while accounting for sex, moult and seasonal effects. We modeled the entire sample for each species, subsamples for each sex, and for adults only. Inferences were made on posterior probabilities of inclusion in the best model for each candidate predictor variable. A posterior probability of inclusion > 0.9 (corresponding to a posterior odds ratio of 10:1) is considered strong evidence that the variable is influential, while a probability > 0.75 (odds ratio 3:1) is considered as “substantial” evidence (Jeffreys 1961).

Model sensitivity

We used a dummy response variable to determine the capacity of models to detect influential variables, given our criterion for substantial evidence (probability of inclusion > 0.75). We generated dummy response variables varying the correlation with a predictor variable. All other data were taken from the actual collected samples of four species (representing the range of sample sizes collected n = 60–250). From this we determined the minimum correlation between the predictor and dummy response variable that yielded a probability of inclusion > 0.75. The capacity to detect a substantial effect increased with sample size (Fig. 2). For species with substantial sample sizes (e.g. brown treecreeper), small effects (˜5–6%) were detectible, while in species with small samples sizes (e.g. weebill, grey shrike-thrush) effects would need to be large (˜20%) to be detectable. If effects differed between sexes, a very large effect would be required if it were to be detected in species with small sample sizes.


Figure 2. Minimum R2 required in simulated dataset for an effect to be detected of p > 0.75.

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Haematocrit, whole blood haemoglobin concentration and residual body mass

There was little evidence for relationships between landscape or site condition and Hb, HCT or RBM pooled between sexes. A single relationship for one species (grey shrike-thrush) was supported between cover of shrubs > 1 m in height and HCT (R2= 0.09, probability of effect p = 0.82). All other effects were unsupported (p < 0.71; Supplementary material Appendix 2). There was no support for any of the models when the sexes or adults alone were considered separately (results not shown).

Sex ratio and homozygosity-by-locus and allelic richness

Per-site sample sizes restricted the analysis of site AR to five species. We found little support for a relationship between allelic richness and landscape or site condition. There was support only for a relationship between canopy cover and AR in the fuscous honeyeater (p = 0.78, Supplementary material Appendix 3).

We found no support for inclusion of any of the landscape or site vegetation condition variables as predictors of skewed sex ratio in any of the study species (p < 0.66, Supplementary material Appendix 4). There was no support for inclusion of landscape or site variables on HL in any species (p < 0.52, Supplementary material Appendix 5), and no evidence of differences in HL between the sexes in any species (p < 0.9).


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We found support of an effect of landscape or site condition on individual body condition, levels of AR, HL or sex-ratio skew in only two of 389 combinations of species, responses and predictors. The study therefore provides little evidence that these effects contribute to the observed decline in woodland birds in the box-ironbark region.

Our simulations showed that the models used were capable of detecting moderate effects, R2 > 0.2 with relatively small samples (n = 60) and very small effects with R2 as small as ca 0.05 with large (n > 250) samples. For a few species and response variables, sample sizes were too small to detect any but the largest effect (i.e. Hb and HCT in weebill, and to a lesser extent in buff-rumped thornbill and dusky woodswallow). However for the majority of tests, sample sizes were sufficient to detect an effect if it were present, either for aggregate samples (204 tests where n > 140), or in many cases also for individual sexes (Table 2).

While Hb, HCT and RBM have been identified as useful measurements for estimation of individual condition in relation to environmental factors (Campbell 1995, Norte et al. 2009a, b), they are subject to variation due to age, sex, moult status and between season, year and time of day; breeding status and parasite loads also affect these measures (Hõrak et al. 1998, Ots et al. 1998, Fair et al. 2007, Norte et al. 2009a, 2010). Where used independently as indices of individual condition, these measures may lead to erroneous conclusions. HCT in particular has been challenged as an independent indicator of condition as it is affected by state of hydration (Dawson and Bortolotti 1997, Fair et al. 2007). Relationships of all three indices with individual condition are not monotonic, and similar values may be caused by positive or negative influences (Fair et al. 2007).

Nevertheless, a main aim of this study was to identify any evidence of an effect of our chosen site or landscape condition measures, which are widely used to describe landscape and habitat change, on Hb, HCT, RBM, HL or sex ratio. We incorporated age, sex, moult status, season, and year into our models as random or fixed effects. We also modelled the two factors explaining the greatest variance, sex and age, separately. Despite the capacity of our analyses to detect small effects, we were unable to detect effects in the comparisons of interest (vegetation and individual condition) in this study.

Studies of the relationship between vegetation and individual condition in wild passerines have produced differing results. Effects of habitat fragmentation in passerines have been shown to include elevated stress in chicks, and decreased RBM and HCT in breeding birds (Mazerolle and Hobson 2002, Suorsa et al. 2003b), while Hb level in chicks, and RBM in adults may relate to food availability among sites (Strong and Sherry 2000, Banbura et al. 2007). No relationship was found between landscape forest cover and RBM for some overwintering birds (Tellería et al. 2001, Turcotte and Desrochers 2008).

The studies showing an effect of fragmentation on individual condition mostly examined breeding individuals or their nestlings. Studies restricted to nestlings, or specifically to birds of known breeding status (Suorsa et al. 2003a, Norte et al. 2009b, 2010) allow the removal of the effects of age or reproductive status from analysis. They offer more sensitive probes of response to vegetation variables. Such studies are likely to be limited to single species readily sampled at the nest, or intensive studies of marked populations. Our study attempted a more general, multi-species approach, sampling many sites without repeated sampling of individuals. We could estimate age only from plumage, and could not be sure of the breeding status of birds, unless they had a marked brood patch, and therefore we had limited ability to account for age and breeding status per se.

The region in which the study was undertaken was under extreme climatic stress at the time of the study, having suffered one of the most extreme droughts worldwide from 1997 to 2010 (Leblanc et al. 2009). While earlier studies recorded declines in avifauna related to area of remnant treecover in the landscape (Radford et al. 2005, Radford and Bennett 2007), the more recent reports found that decline in numbers was occurring across the region regardless of amount of remnant tree-cover (Mac Nally et al. 2009). These declines were found across all foraging guilds, including the nectarivores and insectivores of our study. The declines were probably due to reduced food resources (Mac Nally et al. 2009). This may have led to a uniform degree of stress across the entire region, so that effects of landscape configuration or in-site vegetation would be difficult to ascribe. Turcotte and Desrochers (2008) argued that the lack of effect of habitat fragmentation on body condition may be due to differential mortality, predation, or attempted emigration of individuals in poorer condition from fragments. Such a mechanism may explain the lack of effect in our system, although Turcotte and Desrochers’ (2008) birds were subject to regular seasonal stresses rather than the longer-term set of stresses caused by drought in our study system, albeit also causing reduced food abundance, and potentially reduced breeding (Mac Nally et al. 2009). There is a paradox here, if all except the best-conditioned birds are absent from localities (or indeed a whole region due to intrinsic attributes of the individuals), then it may not be possible to detect the effect in the birds themselves, because the poorer conditioned individuals are absent. The result may be fewer birds remaining, that is, the decline in occurrence observed (Radford and Bennett 2007, Mac Nally et al. 2009) with the proximate cause of the decline no longer apparent since the poorer condition birds are absent.

A second factor that may have reduced our ability to detect an effect was the small range of habitat condition in sites that were of sufficient quality to contain woodland-dependent birds. Habitat scores (Parkes et al. 2003) for sample sites had relatively little variation (12–52 out of a possible 75). Most sites had a similar level of degradation; there were no sites in very good condition, and only a few in exceptionally poor condition. Broad modelling of vegetation condition across the state of Victoria showed that the majority of our sites were at the upper end of the available range of condition (Dept of Sustainability and Environment 2008). Our sampling sites, of necessity, were located where sufficient birds were present to make sampling practical; most sites in very poor condition had few woodland-dependent birds present.

Our results indicate that the commonly used measures of avian condition considered here are not useful for discerning the effects of landscape and vegetation change for woodland-dependent birds. Moreover, there is little evidence that stress per se, at least as indicated by these condition measures, is responsible for the decline or otherwise of the woodland birds. There were good grounds for expecting differences, especially between decliner and tolerant species and between sedentary and mobile species, given the rich background of data from prior work. We had substantial capacity to discern effects of stress were these important.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We wish to thank the more than 60 volunteers who gave days or in some cases weeks of their time to assist us with netting and trapping birds. We thank Geoff Sutter of the Arthur Rylah Inst. for Environmental Research, who collected the Vegetation Quality Assessment Data, and Peter Arcese for comments that improved our manuscript. This project was funded by Australian Research Council Linkage Project LP0776322, the Victorian Dept of Sustainability and Environment, Museum of Victoria, Victorian Dept of Primary Industries, Parks Victoria, North Central Catchment Management Authority, and Goulburn Broken Catchment Management Authority. JNA was funded by Monash Univ. Dean of Science Scholarship and a stipend from Birds Australia, and travel funded by Monash Univ. and the Holsworth Wildlife Research Endowment. This is contribution 257 from the Australian Centre for Biodiversity, Monash Univ.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • Acquarone C., Cucco M., Cauli S. L. and Malacarne G. 2002. Effects of food abundance and predictability on body condition and health parameters: experimental tests with the hooded crow. Ibis 144: E155E163.
  • Amos J. N., Bennett A. F., Mac Nally R., Newell G., Radford J. Q., Pavlova A., Thompson J., White M. and Sunnucks P. 2012. Predicting landscape genetic consequences of habitat loss, fragmentation and mobility for multiple species of woodland birds. PLoS One 7: e30888.
  • Amos W., Wilmer J. W., Fullard K., Burg T. M., Croxall J. P., Bloch D. and Coulson T. 2001. The influence of parental relatedness on reproductive success. Proc. R. Soc. B 268: 20212027.
  • Andren H. 1992. Corvid density and nest predation in relation to forest fragmentation: a landscape perspective. Ecology 73: 794804.
  • Aparicio J., Ortego J. and Cordero P. 2006. What should we weigh to estimate heterozygosity, alleles or loci? Mol. Ecol. 15: 46594665.
  • Arcese P. 2003. Individual quality, environment and conservation. – In: Festa-Bianchet M. and Appolino M. (eds), Animal behaviour and wildlife conservation. Island Press, pp. 271296.
  • Ardia D. R. 2005. Super size me: an experimental test of the factors affecting lipid content and the ability of residual body mass to predict lipid stores in nestling European starlings. Funct. Ecol. 19: 414420.
  • Banbura J., Banbura M., Kalinski A., Skwarska J., Slomczynski R., Wawrzyniak J. and Zielinski P. 2007. Habitat and year- to-year variation in haernoglobin concentration in nestling blue tits Cyanistes caeruleus. Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 148: 572577.
  • Banks S. C., Piggott M. P., Stow A. J. and Taylor A. C. 2007. Sex and sociality in a disconnected world: a review of the impacts of habitat fragmentation on animal social inter actions. Can. J. Zool. 85: 10651079.
  • Banks S. C., Dubach J., Viggers K. L. and Lindenmayer D. B. 2010. Adult survival and microsatellite diversity in possums: effects of major histocompatibility complex-linked micro satellite diversity but not multilocus inbreeding estimators. Oecologia 162: 359370.
  • Bennett A. F. and Radford J. Q. 2009. Thresholds, incidence functions and species-specific cues: responses of woodland birds to landscape structure in south-eastern Australia. – In: Villard M. A. and Jonsson B. G. (eds), Setting conservation targets for managed forest landscapes. Cambridge Univ. Press, pp. 161184.
  • Bojarinova J. G., Lehikoinen E. and Eeva T. 1999. Dependence of postjuvenile moult on hatching date, condition and sex in the great tit. J. Avian Biol. 30: 437446.
  • Bolker B. M., Brooks M. E., Clark C. J., Geange S. W., Poulsen J. R., Stevens M. H. H. and White J.-S. S. 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24: 127135.
  • Campbell T. W. 1995. Avain hemotology and cytology. – Iowa State Univ. Press.
  • Caughley G. 1994. Directions in conservation biology. J. Anim. Ecol. 63: 215244.
  • Chapman J. R., Nakagawa S., Coltman D. W., Slate J. and Sheldon B. C. 2009. A quantitative review of heterozygosity–fitness correlations in animal populations. Mol. Ecol. 18: 27462765.
  • Coltman D. W., Pilkington J. G., Smith J. A. and Pemberton J. M. 1999. Parasite-mediated selection against inbred soay sheep in a free-living island populations. Evolution 53: 12591267.
  • Cooper C. B. and Walters J. R. 2002. Experimental evidence of disrupted dispersal causing decline of an Australian passerine in fragmented habitat. Conserv. Biol. 16: 471478.
  • Cooper C. B., Walters J. R. and Ford H. 2002a. Effects of remnant size and connectivity on the response of brown treecreepers to habitat fragmentation. Emu 102: 249256.
  • Cooper C. B., Walters J. R. and Priddy J. 2002b. Landscape patterns and dispersal success: simulated population dynamics in the brown treecreeper. Ecol. Appl. 12: 15761587.
  • Coulon A. 2010. genhet: an easy-to-use R function to estimate individual heterozygosity. Mol. Ecol. Resour. 10: 167169.
  • Coulson T., Albon S., Slate J. and Pemberton J. M. 1999. Microsatellite loci reveal sex-dependent responses to inbreeding and outbreeding in red deer calves. Evolution 53: 19531960.
  • Coulson T. N., Pemberton J. M., Albon S. D., Beaumont M., Marshall T. C., Slate J., Guinness F. E. and Cluttonbrock T. H. 1998. Microsatellites reveal heterosis in red deer. Proc. R. Soc. B 265: 489495.
  • Dawson R. D. and Bortolotti G. R. 1997. Are avian hematocrits indicative of condition? American kestrels as a model. J. Wildl. Manage. 61: 12971306.
  • Dept of Sustainability and Environment 1990–1999. (TREE25) Tree-cover. – Dept of Sustainability and Environment, East Melbourne.
  • Dept of Sustainability and Environment 2008. NV2005_QUAL. – Dept of Sustainability and Environment, East Melbourne.
  • Doerr V. a. J., Doerr E. D. and Jenkins S. H. 2006. Habitat selection in two Australasian treecreepers: what cues should they use? Emu 106: 93103.
  • Dudaniec R. Y., Kleindorfer S. and Fessl B. 2006. Effects of the introduced ectoparasite Philornis downsi on haemoglobin level and nestling survival in Darwin’s small ground finch (Geospiza fuliginosa). Austral Ecol. 31: 8894.
  • Eklom K. and Lill A. 2006. Development pattern of blood oxygen carrying capacity in rainbow bee-eater nestlings. Aust. J. Zool. 54: 17.
  • Environmental Systems Research Inst. 1999–2008. ARCGIS. – Redlands, CA.
  • Ewers R. M. and Didham R. K. 2005. Confounding factors in the detection of species responses to habitat fragmentation. Biol. Rev. 81: 117.
  • Fahrig L. 2003. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34: 487515.
  • Fair J., Whitaker S. and Pearson B. 2007. Sources of variation in haematocrit in birds. Ibis 149: 535552.
  • Ferwerda F. 2003. Assessing the importance of remnant native vegetation for maintaining biodiversity in rural landscapes using geospatial analysis. – RMIT Univ.
  • Ford H. A. 2011. The causes of decline of birds of eucalypt woodlands: advances in our knowledge over the last 10 years. Emu 111: 19.
  • Ford H. A., Barrett G. W., Saunders D. A. and Recher H. F. 2001. Why have birds in the woodlands of southern Australia declined? Biol. Conserv. 97: 7188.
  • García-Navas V., Ortego J. and Sanz J. J. 2009. Heterozygosity-based assortative mating in blue tits (Cyanistes caeruleus): implications for the evolution of mate choice. Proc. R. Soc. B 276: 2931.
  • Gorrod E. J. and Keith D. A. 2009. Observer variation in field assessments of vegetation condition: implications for bio diversity conservation. Ecol. Manage. Restor. 10: 3140.
  • Goudet J. 2005. hierfstat, a package for r to compute and test hierarchical F-statistics. Mol. Ecol. Not. 5: 184186.
  • Green A. J. 2001. Mass/length residuals: measures of body condition or generators of spurious results? Ecology 82: 14731483.
  • Griffiths R., Double M. C., Orr K. and Dawson R. J. G. 1998. A DNA test to sex most birds. Mol. Ecol. 7: 10711075.
  • Hansson B., Bensch S., Hasselquist D. and Åkesson M. 2001. Microsatellite diversity predicts recruitment of sibling great reed warblers. Proc. R. Soc. B 268: 12871291.
  • Harrisson K. A., Pavlova A., Amos J. N., Takeuchi N., Lill A., Radford J. Q. and Sunnucks P. 2012. Fine-scale effects of habitat loss and fragmentation despite large-scale gene flow for some regionally declining woodland bird species. Landscape Ecol. 27: 813827.
  • Harrisson K., Pavlova A., Amos J. N., Takeuchi N., Lill A., Radford J. and Sunnucks P. in press. Disrupted fine-scale population processes in fragmented landscapes despite large-scale genetic connectivity for a widespread and common cooperative breeder: the superb fairy-wren (Malurus cyaneus). J. Anim. Ecol.
  • Higgins P. J., Peter J. M. and Steele W. K. (eds) 2001. Handbook of Australian, New Zealand and Antarctic birds. Volume 5: tyrant-flycatchers to chats. – Oxford Univ. Press.
  • Higgins P. J. and Peter J. M. (eds) 2002. Handbook of Australian, New Zealand and Antarctic birds. Volume 5: pardalotes to shrike-thrushes. – Oxford Univ. Press.
  • Hõrak P., Jenni-Eiermann S., Ots I. and Tegelmann L. 1998. Health and reproduction: the sex-specific clinical profile of great tits (Parus major) in relation to breeding. Can. J. Zool. 76: 22352244.
  • Huhta E., Aho T., Jäntti A., Suorsa P., Kuitunen M., Nikula A. and Hakkarainen H. 2004. Forest fragmentation increases nest predation in the Eurasian treecreeper. Conserv. Biol. 18: 148155.
  • Jeffreys H. 1961. Theory of probability. – Oxford Univ. Press.
  • Johnstone C., Lill A. and Reina R. 2011. Does habitat fragmentation cause stress in the agile antechinus? A haematological approach. J. Comp. Physiol. B 182: 139155.
  • Kilgas P., Tilgar V. and Mänd R. 2006. Hematological health state indices predict local survival in a small passerine bird, the great tit (Parus major). Physiol. Biochem. Zool. 79: 565572.
  • Latta S. C. and Faaborg J. 2002. Demographic and population responses of Cape May warblers wintering in multiple habitats. Ecology 83: 25022515.
  • Leblanc M. J., Tregoning P., Ramillien G., Tweed S. O. and Fakes A. 2009. Basin-scale, integrated observations of the early 21st century multiyear drought in southeast Australia. Water Resour. Res. 45: W04408.
  • Lindenmayer D. and Luck G. 2005. Synthesis: thresholds in conservation and management. Biol. Conserv. 124: 351354.
  • Lindenmayer D. B. and Fischer J. 2007. Tackling the habitat fragmentation panchreston. Trends Ecol. Evol. 22: 127132.
  • Lindenmayer D., Hobbs R. J., Montague-Drake R., Alexandra J., Bennett A., Burgman M., Cale P., Calhoun A., Cramer V., Cullen P., Driscoll D., Fahrig L., Fischer J., Franklin J., Haila Y., Hunter M., Gibbons P., Lake S., Luck G., Macgregor C., Mcintyre S., Nally R. M., Manning A., Miller J., Mooney H., Noss R., Possingham H., Saunders D., Schmiegelow F., Scott M., Simberloff D., Sisk T., Tabor G., Walker B., Wiens J., Woinarski J. and Zavaleta E. 2008. A checklist for ecological management of landscapes for conservation. Ecol. Lett. 11: 7891.
  • Linkie M., Chapron G., Martyr D. J., Holden J. and Leader-Williams N. 2006. Assessing the viability of tiger subpopulations in a fragmented landscape. J. Appl. Ecol. 43: 576586.
  • Lunn D. J., Thomas A., Best N. and Spiegelhalter D. 2000. WinBUGS – a Bayesian modelling framework: concepts, structure, and extensibility. Stat. Comput. 10: 325337.
  • Lunn D. J., Best N. and Whittaker J. C. 2009. Generic reversible jump MCMC using graphical models. Stat. Comput. 19: 395408.
  • Mac Nally R. 1990. The roles of floristics and physionomy in avian community composition. Aust. J. Ecol. 15: 321327.
  • Mac Nally R., Bennett A. F., Thomson J. R., Radford J. Q., Unmack G., Horrocks G. and Vesk P. A. 2009. Collapse of an avifauna: climate change appears to exacerbate habitat loss and degradation. Divers. Distrib. 15: 720730.
  • Manel S., Schwartz M. K., Luikart G. and Taberlet P. 2003. Landscape genetics: combining landscape ecology and population genetics. Trends Ecol. Evol. 18: 189197.
  • Maron M., Main A., Bowen M., Howes A., Kath J., Pillette C. and Mcalpine C. A. 2011. Relative influence of habitat modification and interspecific competition on woodland bird assemblages in eastern Australia. Emu 111: 4051.
  • Mazerolle D. F. and Hobson K. A. 2002. Physiological ramifications of habitat selection in territorial male ovenbirds: consequences of landscape fragmentation. Oecologia 130: 356363.
  • Mönkkönen M. and Reunanen P. 1999. On critical thresholds in landscape connectivity: a management perspective. Oikos 84: 302305.
  • Norte A., Ramos J., Sousa J. and Sheldon B. 2009a. Variation of adult great tit Parus major body condition and blood para meters in relation to sex, age, year and season. J. Ornithol. 150: 651660.
  • Norte C. A., Sheldon B. C., Sousa J. P. and Ramos J. A. 2009b. Environmental and genetic variation in body condition and blood profile of great tit Parus major nestlings. J. Avian Biol. 40: 157165.
  • Norte A. C., Ramos J. A., Sampaio H. L., Sousa J. P. and Sheldon B. C. 2010. Physiological condition and breeding performance of the great tit. Condor 112: 7986.
  • Noske R. 1979. Co-existence of three species of treecreepers in north-eastern New South Wales. Emu 79: 120128.
  • Ots I., Murumagi A. and Horak P. 1998. Haematological health state indices of reproducing great tits: methodology and sources of natural variation. Funct. Ecol. 12: 700707.
  • Palstra F. P. and Ruzzante D. E. 2008. Genetic estimates of contemporary effective population size: what can they tell us about the importance of genetic stochasticity for wild population persistence? Mol. Ecol. 17: 34283447.
  • Parkes D., Newell G. and Cheal D. 2003. Assessing the quality of native vegetation: the ‘habitat hectares’ approach. Ecol. Manage. Restor. 4: 2938.
  • Pavlova A., Amos J. N., Goretskaia M. I., Beme I. R., Buchanan K., Takeuchi N., Radford J. Q. and Sunnucks P. 2012. Genes and song: genetic and social connections in fragmented habitat in a woodland bird with limited dispersal. – Ecology in press.
  • Porlier M., Belisle M. and Garant D. 2009. Non-random distribution of individual genetic diversity along an environmental gradient. Phil. Trans. R. Soc. B 364: 15431554.
  • Radford J. Q. and Bennett A. F. 2007. The relative importance of landscape properties for woodland birds in agricultural environments. J. Appl. Ecol. 44: 737747.
  • Radford J. Q., Bennett A. F. and Cheers G. J. 2005. Landscape-level thresholds of habitat cover for woodland-dependent birds. Biol. Conserv. 124: 317337.
  • Reilly P. 1968. A flame robin trap. – Australian Bird Bander.
  • Robinson D. and Traill B. J. 1996. Conserving woodland birds in the wheat and sheep belts of southern Australia. – RAOU Conservation Statement, Royal Australasian Orinithological Union, Melbounre.
  • Rogers K., Rogers A., Rogers D., Lane B. and Male B. 1986. Bander’s aid a guide to ageing and sexing bush birds. – A. Rogers, St Andrews, VIC.
  • Rotenberry J. T. 1985. The role of habitat in avian community composition: physiognomy or floristics? Oecologia 67: 213217.
  • Saunders D. A., Hobbs R. J. and Margules C. R. 1991. Biological consequences of ecosystem fragmentation: a review. Conserv. Biol. 5: 1832.
  • Schulte-Hostedde A. I., Zinner B., Millar J. S. and Hickling G. J. 2005. Restitution of mass–size residuals: validating body condition indices. Ecology 86: 155163.
  • Seddon N., Amos W., Mulder R. A. and Tobias J. A. 2004. Male heterozygosity predicts territory size, song structure and reproductive success in a cooperatively breeding bird. Proc. R. Soc. B 271: 18231829.
  • Seewagen C. L. 2008. An evaluation of condition indices and predictive models for noninvasive estimates of lipid mass of migrating common yellowthroats, ovenbirds, and Swainson’s thrushes. J. Field Ornithol. 79: 8086.
  • Simmons P. and Lill A. 2006. Development of parameters influencing blood oxygen carrying capacity in the welcome swallow and fairy martin. Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 143: 459468.
  • Stevenson R. D. and Woods W. A. 2006. Condition indices for conservation: new uses for evolving tools. Integr. Comp. Biol. 46: 11691190.
  • Storfer A., Murphy M. A., Evans J. S., Goldberg C. S., Robinson S., Spear S. F., Dezzani R., Delmelle E., Vierling L. and Waits L. P. 2007. Putting the ‘landscape’ in landscape genetics. Heredity 98: 128142.
  • Strong A. M. and Sherry T. W. 2000. Habitat-specific effects of food abundance on the condition of ovenbirds wintering in Jamaica. J. Anim. Ecol. 69: 883895.
  • Suorsa P., Huhta E., Nikula A., Nikinmaa M., Jantti A., Helle H. and Hakkarainen H. 2003a. Forest management is associated with physiological stress in an old-growth forest passerine. Proc. R. Soc. B 270: 963969.
  • Suorsa P., Huhta E., Nikula A., Nikinmaa M., Jäntti A., Helle H. and Hakkarainen H. 2003b. Forest management is associated with physiological stress in an old-growth forest passerine. Proc. R. Soc. B 270: 963969.
  • Tellería J. L., Virgós E., Carbonell R., Pérez-Tris J. and Santos T. 2001. Behavioural responses to changing landscapes: flock structure and anti-predator strategies of tits wintering in fragmented forests. Oikos 95: 253264.
  • Turcotte Y. and Desrochers A. 2008. Forest fragmentation and body condition in wintering black-capped chickadees. Can. J. Zool. 86: 572581.
  • Vesk P. and Mac Nally R. 2006. The clock is ticking – revegetation and habitat for birds and arboreal mammals in rural landscapes of southern Australia. Agric. Ecosyst. Environ. 112: 356366.
  • Watson D. M. 2011. A productivity-based explanation for woodland bird declines: poorer soils yield less food. Emu 111: 1018.
  • World Resources Inst. 2005. Ecosystems and human well-being: biodiversity synthesis. – Millennium Ecosystem Assessment. World Resources Inst., Washington, DC.
  • Yates C. J. and Hobbs R. J. 1997. Temperate eucalypt woodlands: a review of their status, processes threatening their persistence and techniques for restoration. Aust. J. Bot. 45: 949973.
  • Zanette L., Doyle P. and Trémont S. M. 2000. Food shortage in small fragments: evidence from an area-sensitive passerine. Ecology 81: 16541666.

Supplementary material (Appendix JAB5746 at < >). Appendix 1–5.