Diet and Prey Availability of Terrestrial Insectivorous Birds Prone to Extinction in Amazonian Forest Fragments


of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
This study compared niche breath, prey size, and diet variability in two pairs of sympatric species of terrestrial insectivorous birds, each pair containing one species that can persist in small forest fragments and one that does not. The pairs were
   Braz. Arch. Biol. Technol. v.53 n. 6: pp. 1371-1381, Nov/Dec 2010 1371 Vol.53, n. 6: pp.1371-1381, November-December 2010 ISSN 1516-8913 Printed in Brazil BRAZILIAN ARCHIVES OF BIOLOGY AND TECHNOLOGY   AN INTERNATIONAL JOURNAL   Diet and Prey Availability of Terrestrial Insectivorous Birds Prone to Extinction in Amazonian Forest Fragments Luiz Augusto Macedo Mestre 1,2,3 *  , Mario Cohn-Haft 4  and Manoel Martins Dias 2   1 South Dakota State University; Geographic Information Science Center of Excellence; 1021; Medary Ave.; Wecota  Hall, 57007; Brookings - SD - USA. 2 Universidade Federal de São Carlos; Departamento de Ecologia e Biologia  Evolutiva; C.P.: 676; 13.565-905; São Carlos - SP - Brasil. 3 Universidade Federal do Paraná; Rua Pioneiro, 2153; 85950-000; Palotina - PR - Brasil. 4Instituto Nacional de Pesquisas da Amazônia; C.P.: 478; 69083-000; Manaus -  AM - Brasil ABSTRACT This study compared niche breath, prey size, and diet variability in two pairs of sympatric species of terrestrial insectivorous birds, each pair containing one species that can persist in small forest fragments and one that does not. The pairs were Myrmeciza ferruginea  and Sclerurus rufigularis ; and Formicarius colma  and   F. analis  , respectively. The prey availability in forest fragments was also sampled and compared to the availability in continuous forests. Niche breath indices did not differ between pair members, but diet variability differed in the opposite direction from that hypothesized. Although the two bird species most vulnerable to fragmentation fed on larger prey than less vulnerable species, prey availability, including that based on prey size did not differ among  fragmented versus continuous forest sites. Thus, diet per se appeared not to be an important cause of extinction- proneness in these species. The simplest explanation proposed, that vulnerability to fragmentation was directly related to territory size, requires testing. However, it was consistent with observations that the bird species feeding on larger prey also need larger territories. Key words:  Amazonian forests, Bird diet, Forest fragmentation, Insectivores,  Formicarius colma , Formicarius   analis ,  Myrmeciza ferruginea , Sclerurus rufigularis   * Author for correspondence: INTRODUCTION Forest fragmentation affects the composition of forest bird communities, especially in the humid tropics where the rates of forest destruction are high and where birds are generally more specialized in their foraging tactics, live in more specific habitats, and need larger territories than in temperate forests (Stouffer and Bierregaard, 1995; Hagan et al., 1996). Different bird species react differently to deforestation (Canaday, 1996). Species with greater extinction-proneness generally are at their geographic or elevational limits. Higher vulnerability may also be related to life history and ecological characteristics, such as body size, small populations, habitat specialization, or low survival rate (Kattan et al., 1994). Bird species negatively affected by the forest fragmentation are generally restricted to the interior of primary forests and have little ability to use second growth and forest edges (Offerman et al., 1995). In central Amazonia, bird capture rates of some canopy and gap specialists have increased after  Mestre, L. A. M. et al. Braz. Arch. Biol. Technol. v.53 n. 6: pp. 1371-1381, Nov/Dec 2010 1372 forest fragmentation. However, most common understory insectivorous species disappeared from the isolated forest fragments during first years of isolation (Stouffer and Bierregaard, 1995). After this decline, ant followers and some obligate flock species returned to the fragments surrounded by second growth. Despite the fact that some species of terrestrial insectivores were still observed in the forest remnants and second growth sites, this group seems to be one of the most affected by forest fragmentation (Stouffer and Bierregaard, 1995; Stratford, 1997; Stratford and Stouffer, 1999; Borges and Stouffer, 1999; Ferraz et al., 2003; Ferraz et al., 2007). Several hypotheses have been proposed to explain the decline of understory insectivorous birds in forest fragments, including physiological sensitivity to microclimatic changes, habitat specificity, dispersal ability, and food scarcity (Sekercioglu et al., 2002). The food scarcity hypothesis is frequently proposed, because forest fragmentation generally alters the arthropod communities in the region (Brown, 1991; Offerman et al., 1995; Didham et al., 1996; Didham, 1997; Stratford, 1997). Thus, the decline of some groups of litter invertebrates and modifications in biomass of understory arthropods (Didham et al., 1996; Malcolm, 1997) could directly affect the resources and consequently the maintenance of bird populations (Burke and Nol, 1998). Moreover, possible trophic specialization of terrestrial insectivores may allow some bird species to remain in forest fragments, because different habitat preferences and foraging tactics can determine the consumption of different types and number of prey (Sherry, 1984; Cohn-Haft, 1995). If sensitivity to fragmentation is associated with diet, it is reasonable to expect that species with relatively flexible diets will be most likely to persist in forest remnants. We tested this hypothesis by examining if the fragmentation-resistant species have more flexible diets than the vulnerable species. We compared niche breath, prey size and diet variability among four species of terrestrial insectivorous birds (  Myrmeciza  ferruginea, Sclerurus rufigularis, Formicarius colma ,   and  Formicarius analis ) and quantified prey availability in forest fragments. We attempted to control variables other than vulnerability to fragmentation by comparing pairs of species, in which one member can persist in forest fragments and one does not. Our hypothesis was that species more vulnerable to forest fragmentation will differ in diet composition from those resistant to fragmentation, linking the diet characteristics with vulnerability to fragmentation. METHODS Study sites Fieldwork was conducted in terra firme  (upland, non-flooded)   tropical rain forest in the reserves managed by the Biological Dynamics of Forest Fragments Project (BDFFP, a collaborative research project of the Instituto Nacional de Pesquisas da Amazônia and the Smithsonian Institution), located approximately 80 km north of Manaus, Amazonas, Brazil (2 o 20’00”S / 60 o 00’00”W) (Fig. 1). Surrounding the ranches and embedded forest fragments (see below), continuous primary forest extends to the east, west and north, occasionally interrupted by roads, but forest to the south has been progressively disturbed. The continuous forest has a relatively closed canopy, approximately 30 m high, and a relatively open understory dominated by palms. Prey availability data were also collected from seven forest fragments that were created by the clearing of land for cattle ranches. These fragments are separated by distances of 70 to 650 m from the nearest continuous forest (see Lovejoy et al. [1986] for descriptions of the individual fragments). Seasonality at the sites is relative to other lowland moist forests (Gentry 1990). Annual rainfall averages about 2200 mm, with a peak in March and April (>300 mm/month) and a dry season from July to September (<100 mm/month). The mean annual temperature is 26.7 o  C (Salati et al., 1991). More detailed descriptions of these sites are given by Lovejoy et al .  (1983), Gentry (1990), and Stouffer and Bierregaard (1995). The complete avifauna of the sites is presented by Cohn-Haft et al. (1997). Bird species The present study analyzed the diets of two pairs of bird species in the same trophic guild, with similar weight and bill length (size data from Bierregaard, 1988): Formicarius colma  (Rufous-capped Antthrush, Formicariidae, 46 g.) and  Formicarius analis  (Black-faced Antthrush, Formicariidae, 62 g.);  Myrmeciza ferruginea (Ferruginous-backed Antbird, Thamnophilidae,  Diet and Prey Availability of Terrestrial Insectivorous Birds Braz. Arch. Biol. Technol. v.53 n. 6: pp. 1371-1381, Nov/Dec 2010 1373 25 g.) and Sclerurus rufigularis (Short-billed Leaftosser, Furnariidae, 21 g.). The first species of each pair is present in the forest fragments as well as in continuous forests, whereas the second occurs only in pristine, unfragmented forests (Stouffer and Bierregaard, 1995; Stratford and Stouffer 1999). This paired design attempted to control the factors other than diet that could influence the susceptibility to fragmentation (Felsenstein, 1985; Harvey and Pagel, 1991). These four species of terrestrial insectivorous birds forage mainly in the leaf litter. Figure 1 -  Study sites. Brazil and legal Brazilian Amazon (dark); detail of study sites on the right (adapted from BDFFP databases). Furthermore, various other aspects of their biology and ecology have been studied in the same areas, such as comparisons of body dimensions (Bierregaard, 1988), territory sizes (Stouffer, 1997; Stratford, 1997), interspecific interactions (Stouffer, 1997), and population persistence in the studied fragments (Stouffer and Bierregaard, 1995; Stratford, 1997; Stratford and Stouffer, 1999). Diet sampling The study was conducted between February and October 2001, and April and May 2002. The birds were localized by playback with audio equipment. When an individual replied, we set one or two mist-nets and attracted the birds with continued playback. Each capture site was noted to distribute sampling spatially and avoid recapture. The birds were captured between 0700 and 1500 h, so that the captured birds could forage before and after the treatment. Each bird was identified, marked by cutting the tip of a secondary remige, and induced to regurgitate by oral administration of tartar emetic (1.5 % solution of antimony and potassium tartar, 0.8 ml/100 g). The birds were placed in a cardboard box for 30 min and then released. This method allowed diet analysis with low mortality rates (Robinson and Holmes, 1982; Poulin et al. 1994abc, Mallet-Rodrigues et al. 1997), but see Jonson et al. (2001). All the birds induced to regurgitate survived to release, and two of them were recaptured. Those recaptured individuals were not induced to regurgitate. Diet data were collected only in primary continuous forest ( Porto Alegre, Dimona , and  Esteio farms, including the reserves Florestal,   Gavião,  and Km 41 ). No individuals were sampled in the fragments due to the absence of the two extinction-prone species in these sites and low abundance of the other two species in those areas; this avoided inadequate sample sizes and the threat N  Mestre, L. A. M. et al. Braz. Arch. Biol. Technol. v.53 n. 6: pp. 1371-1381, Nov/Dec 2010 1374 of mortality of these few individuals in fragments, where the presence of an undisturbed avifauna is essential to other studies conducted there. Diet analyses Bird stomach contents were preserved in 80% alcohol and processed using a stereo-microscope. Diagnostic prey fragments were identified and sorted into categories as more precise taxonomic level as possible, mainly in order level and life stage (larva or adult). The prey fragments were subsequently matched to determine the minimum number of individuals per prey category. For example, three orthopteran mandibles represented three orthopterans if they all differed in size or shape or if all came from the same side of the arthropod body; they represented two individuals if two of the three could be matched as a pair (Cohn-Haft, 1995). References used to aid in the identification of prey fragments included Borror et al. (1981), CSIRO (1979), and Chapman and Rosenberg (1991). We reexamined earlier samples after becoming familiar with the range of prey fragments. All prey fragments known to represent prey items were successfully identified, and there was no “unidentified” prey category. The prey size was estimated for 96% of all diet components. These estimates were calculated using linear regression from more than 150 arthropods collected in the prey availability samples. Regressions were made using diagnostic prey fragments (such as spider fangs or orthopteran mandibles) and total prey size (with same taxonomic level of diet samples). We estimated the prey size only when this regression had a highly significant slope, different from zero (p<0.001), and Pearson correlation coefficient superior to 0.90 (Rosenberg, 1993). Diet indices Indices were calculated for each species, assuming that the diet composition was a trait that could be characterized for an entire species (or population), that regurgitated samples from different individuals at different times represented a random sample of the dietary variation found within the population, and that diet variation among individuals was equivalent to that within individuals over the study period. Limiting sampling to a single region controlled possible geographic variation in the diet of the same species. Niche breadth was calculated using the inverse of Simpson’s (1949) Index: S = 1/  Σ p i 2 , where p i  is that proportion of the diet comprised by the i th prey category ( Σ  = sum). This index is mathematically related to richness and to the Shannon Diversity Index, but differs only in the importance given to most-represented categories, weighting abundance most heavily (Hill, 1973; May, 1975). Species with a flexible diet should have larger niche breath and we expected that the resistant species fed on a more diverse prey “community”. The variability in stomach contents among individuals within a species was calculated with a population dietary heterogeneity index (PDH; Sherry, 1984; 1990). PDH compares the diet samples of each species as G H  /df, where G H  is the G  statistic for heterogeneity (Sokal and Rohlf, 1981), and df is the degrees of freedom: (no. samples – 1) x (total no. prey categories in all samples for the species – 1). This index measures the “prey category by stomach sample” interaction (Sokal and Rohlf, 1981). In this study, an arbitrary small constant (0.0001) was added to every cell, because G H could not be calculated on tables with zero cells (prey categories not represented in a given sample) (Cohn-Haft, 1995). We expected that extinction-resistant species have more heterogeneous diets. Prey availability sampling The leaf litter arthropod communities were sampled simultaneously with diet sampling from February to September 2001. We collected the samples in four 1ha fragments, three 10ha fragments, and at three sites within the continuous forests, in the same areas where the birds were captured. The sampling sites were divided in 25 equal-sized squares, and we randomly selected five of these sites. We collected an area of 30x30 cm of leaf liter in the center of each selected site. The leaf litter was promptly placed in a plastic bag for sorting at the field camp. Sampled litter was examined in a plastic tray, and leaves were sifted (2mm mesh). Invertebrates (except the mites and collembolans) > 0.5mm in length were collected with tweezers and preserved in 80% alcohol. They were identified later in the lab, being organized in three size classes (0.5-2mm, 2-6mm, and >6mm) and sorted into broad taxonomic groups, order and life stage (larvae or adult).  Diet and Prey Availability of Terrestrial Insectivorous Birds Braz. Arch. Biol. Technol. v.53 n. 6: pp. 1371-1381, Nov/Dec 2010 1375 The means of proportions of total arthropods, and proportion of each taxon among the continuous forests, 1-ha, and 10-ha fragments, were compared with Kruskal-Wallis test complemented by Dunn’s test. The abundance and richness values of arthropods were also compared between the wet and dry seasons with Mann-Whitney test. Abundance and proportions of invertebrates in the three size classes were compared between the fragments and continuous forests with Mann-Whitney U tests. Comparative analyses We tested the hypothesis that dietary indices of the vulnerable species in each pair differed from those of the corresponding fragmentation-resistant species, and predicted that the direction of the difference was the same in both pairs. To compare these indices within the pairs, we estimated by bootstrapping the means and degrees of freedom of the index for each species (Efron, 1982, Lanyon, 1987). The bootstrapping method estimated these values for each species, re-sampling 100 (in this case) pseudoreplicates of each index from the srcinal sample, assuming that these samples were random and represented the diet of the whole population (Cohn-Haft, 1995). Thus, the three dietary indices were calculated 100 times obtaining means and standard deviations and permitting comparisons with Student t  -tests. We used Mann-Whitney U test when the assumptions to run parametric tests were not met. To test the prediction that vulnerable species eat prey of different sizes from those eaten by resistant species, the values of prey sizes were compared within the pairs using Mann-Whitney U tests. Sample adequacy assessment The sample adequacy for each species was assessed by generating prey type saturation curves (by taxonomic order and life stage), in addition to the bootstrap replicates, which provided the confidence estimates for each pairwise comparison. If a species’ diet was adequately characterized in richness, addition of samples should add no new prey types to the known diet, approaching an asymptote with increasing sample size. To avoid the problem of order of inclusion, the saturation curves for each species were drawn using the means and standard deviations obtained by a random subsampling procedure (Cohn-Haft, 1995). RESULTS The present analyses are based on examination and identification of over 1200 prey items from 39 regurgitated samples: 16 from  Myrmeciza  ferruginea , six from Sclerurus rufigularis , eight from Formicarius colma , and nine from F. analis . Twenty-one prey categories were identified, including 12 insect orders, beetle and dipteran larvae, three arachnid orders, egg cases (of spiders and insects), gastropod mollusks (snails), worms, and small vertebrates (lizards or frogs). Individual samples contained from four to 192 prey items. Diet composition  Myrmeciza ferruginea  fed mainly on orthopterans, egg cases, spiders, harvestmen (Opiliones), and beetles (>80% of diet). The main preys of Sclerurus rufigularis  were spiders, egg cases, ants, beetle larvae, and adult beetles. The diets of Formicarius colma and F. analis were comprised mainly of ants, orthopterans, dipteran larvae, and adult beetles (Fig. 2). Contrary to our expectations, pair members were more similar to one another in their overall diets than were extinction-prone or resistant species (Fig. 2). The two Formicarius species did not differ in the proportions of prey types common to both (p>0.05).  M. ferruginea and S. rufigularis  differed in the proportions of harvestmen (Opiliones), orthopterans and beetle larvae (Mann-Whitney, p<0.05). They fed on similar proportions of egg cases, beetles, spiders and ants (p>0.10). Prey saturation curves indicated that only  M.  ferruginea  approached an asymptote. This suggested that population dietary richness of the other species would likely continue to increase with additional sampling (Fig. 2). The numbers of prey types consumed by the four species were similar (  M. ferruginea  = 15, S. rufigularis = 13, F. colma  = 15, F. analis = 16). Results of bootstrap replication indicated that the differences in population dietary richness within the four species were also not statistically different ( t  -tests, p>0.05) (Table 1). The diet diversity was higher in the smaller pair, with S. rufigularis presenting the most diverse diet (6.24) and F. analis  the least (2.77). The comparisons of bootstrapping results showed that population dietary diversity was similar within species pairs ( t- tests, p>0.05), but widely different between the
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks