Biodivers Conserv (2009) 18:2719–2742 DOI 10.1007/s10531-009-9671-x ORIGINAL PAPER

Conservation value for birds of traditionally managed isolated trees in an agricultural landscape of Madagascar Emily A. Martin Æ Lalaı¨na Ratsimisetra Æ Francis Laloe¨ Æ Ste´phanie M. Carrie`re

Received: 28 January 2009 / Accepted: 13 May 2009 / Published online: 30 May 2009 Ó Springer Science+Business Media B.V. 2009

Abstract Isolated trees have distinctive economic, social and cultural value for the Betsileo people living on the edge of the protected forest corridor between Ranomafana and Andringitra national parks, in South-East Madagascar. Many of these trees are Ficus spp., traditionally protected and respected. At the landscape level, they are isolated features in a heterogeneous mosaic, providing fruit, shade and aesthetic services in open cultivated areas. Within the current management system, isolated trees may also contribute significantly to the provision of ecological services by enhancing bird diversity in open areas outside the forest. We identified practices and values linked to isolated tree uses and management through ethnographic data collection. Bird presence and abundance were sampled by 338 point counts in isolated trees and open areas of the agricultural mosaic. Isolated trees were occupied by 18 out of 32 (56%) bird species in the agricultural mosaic, including 8 (25%) endemic forest species. Endemic forest birds were significantly more numerous in isolated trees than in open habitats, both in species richness and abundance (mean P value \ 0.001). Overall bird species richness was significantly higher in open areas containing isolated trees, than in areas without isolated trees. Bird species richness in Ficus spp. was significantly higher than in other isolated tree species, although no differences were detected in abundance or within guilds. Community-based management of isolated trees may thus represent an opportunity for convergence between bird conservation goals outside protected areas and local management values and practices. Keywords Agricultural landscape
Biodiversity conservation
Birds
Community-based resource management
Ficus spp.
Isolated trees
Madagascar
Ranomafana–Andringitra forest corridor
Traditional practices

E. A. Martin
F. Laloe¨
S. M. Carrie`re (&) IRD-Montpellier, 911, Avenue Agropolis, BP 64501, 34394 Montpellier Cedex 5, France e-mail: [email protected] E. A. Martin e-mail: [email protected] L. Ratsimisetra Faculte´ des Sciences, Universite´ Antananarivo, Ankatso, BP 906, 101 Antananarivo, Madagascar

123

2720

Biodivers Conserv (2009) 18:2719–2742

Introduction It is widely acknowledged today that if biodiversity losses due to human modification of ecosystems are to be checked, conservation efforts in agricultural and other production landscapes are necessary (Steffan-Dewenter et al. 2007; Norris 2008; Scherr and McNeely 2008). It is more and more apparent that the sole conservation of remaining forest tracts and ‘pristine’ areas worldwide is insufficient to provide the necessary conditions for most species’ persistence, while still providing satisfactory livelihoods to increasing human populations (Margules and Pressey 2000; Millenium Ecosystem Assessment 2003). This is particularly true on forest margins in tropical areas, where threats to highly endemic biodiversity are often as great as the need for poverty alleviation of neighbouring human populations (Wunder 2001). Combining both conservation and development objectives in these areas is of particular importance, but few efforts effectively seek to balance these seemingly opposite objectives within identical geographical contexts (Garnett et al. 2007). Community-based natural resource management has been advocated as the most efficient way to achieve such a balance (Gadgil et al. 1993; Agrawal and Gibson 1999; Berkes 2004; Papaik et al. 2008). However, this type of management has had mixed results in the tropics (Agrawal and Gibson 1999; Berkes 2004; Garnett et al. 2007). One reason for this is a frequent lack of consistency between ecological conservation criteria, and local stakeholders’ perception of what is important (Sheil et al. 2002; Laumonier et al. 2008). If sustainability rules enforced by managers and conservationists are consistent with local values and resource use regulations, the efficiency of community-based management is likely to increase (Colding and Folke 1997; Berkes 2004; Jones et al. 2008). One role of conservationists is then to strengthen local conservation institutions, not replace them (Berkes 2004). Where possible, it is thus a valid and crucial process for science to identify, evaluate and promote those local practices that contribute to maintaining or increasing ecological value of agricultural landscapes (Colding and Folke 1997; Berkes et al. 2000). Traditional practices identified as benefiting biodiversity in agricultural areas include the protection of sacred groves in several African communities, which act as refuges for remnant biodiversity (Berkes et al. 1995, 2000), the protection or prohibition of use of given species (Colding and Folke 1997), or the management of patchiness at a landscape level towards increased habitat heterogeneity (Gadgil et al. 1993; Scherr and McNeely 2008), for instance by sparing trees during slash-and-burn clearing (Carrie`re 2002; Carrie`re et al. 2002). Such practices are often linked with economic usefulness of protected species and/or their social or religious significance, sometimes exemplified through social or ritual taboos (Berkes et al. 1995; Colding and Folke 1997; Carrie`re et al. 2002). Although practices are not always aimed at protecting biodiversity, protection of biodiversity can be a consequence of their existence (Berkes et al. 1995, 2000), thus bringing special conservation interest to the further use of these practices by local communities. On the eastern edge of the Ranomafana–Andringitra forest corridor in Madagascar, increasing rural populations currently face critical natural resource management issues linked with the need to balance agricultural expansion with strict restrictions of slash-andburn clearing practices in and around the protected forest corridor. These regulations could lead to radical agricultural intensification and landscape homogenization in currently cultivated areas. As part of the national-level Environmental Action Plan for Madagascar (first started in 1990 for 15 years), the law for safe local management of renewable resources (GELOSE) introduced contracts for transfers of power to populations, including for contractual forest management by local communities or CoBa (Communaute´s de base).

123

Biodivers Conserv (2009) 18:2719–2742

2721

In our study region, management is delegated to CoBa since 2003 (Blanc-Pamard and Ramiarantsoa 2007). Formed by voluntary associations of villagers in cooperation with administrative officers, CoBa are the obligate framework within which resource management contracts between populations and authorities are designed and implemented (Blanc-Pamard and Ramiarantsoa 2007). To date, however, these contracts concern only forests, not agricultural areas. The question is raised of how such community-based management can best favour resilience of the surrounding forest ecosystem, as well as benefit local agricultural (economic) sufficiency, within this highly heterogeneous agricultural landscape which includes fallows, forest fragments, small groves, plantations, as well as many isolated trees in the midst of open cultivated areas. We hypothesized (1) that traditional landscape management practices involving the maintenance of habitat heterogeneity through isolated tree management were beneficial to the persistence of species-level biodiversity in this landscape, including biodiversity found primarily in the nearby forest corridor and (2) that because they are anchored in local values and regulations, these practices could provide useful guidelines for communitybased management. This study focused on bird diversity, because of the usefulness of this taxon as a biodiversity indicator in many tropical areas (e.g., Daily et al. 2001; Hughes et al. 2002; Gray et al. 2007). In the context of Madagascar, bird community dynamics are of particular importance in rural areas because of local paucity in species numbers (Goodman and Ganzhorn 1997) coupled with high degrees of endemism (Langrand 1995; Goodman and Benstead 2003). This is especially true of frugivores, which are key seed dispersers for most tropical tree species (Shanahan et al. 2001; Terborgh et al. 2002), but are particularly rare on the island (Fleming et al. 1987; Goodman and Ganzhorn 1997). However, very few studies in Madagascar document bird distribution in rural landscapes outside extensive forest habitats, despite some focus on the effects of habitat fragmentation (Goodman and Raherilalao 2003; but see Watson et al. 2004). Trees and groves are often favourite sites for birds, bats and primates in cultivated areas, as these animals are often reluctant to venture directly into open spaces (Wegner and Merriam 1979; Da Silva et al. 1996). These landscape features appear to play a key role in attracting birds and mammals into cultivated areas, contributing both to overall animal diversity and to diversity and abundance of seed dispersal into the fields beneath tree crowns (Da Silva et al. 1996; Bleher and Bo¨hning-Gaese 2001; Carrie`re et al. 2002). The aim of this paper is firstly to assess the importance of isolated trees in local culture and economy and to identify practices associated with their management in the landscape. We then test the value of these trees for bird tolerance of cultivated areas, by comparing bird diversity in areas containing isolated trees, and in open areas without trees. We predicted (1) that bird species richness would be higher in isolated trees than in areas without trees, within and among bird groups, except in the case of open-habitat specialists; (2) that isolated Ficus trees, a key resource for tropical frugivores (Shanahan et al. 2001), would attract more frugivorous birds than other isolated trees.

Methods Study area and landscape features Research was conducted in an agricultural landscape of the Malagasy highlands, in the village lands of Ambendrana, Igodona and Amindrabe´, on the north-western edge of the

123

2722

Biodivers Conserv (2009) 18:2719–2742

100 km-long forested corridor which connects the national parks of Ranomafana and Andringitra (province of Fianarantsoa, at about 21° South Latitude and 47° East Longitude). Altitude varies between 1,100 and 1,250 m. Mean annual temperature and rainfall are 17.7°C and 1,370 mm, respectively (Randriamalala et al. 2007). Climate is tropical with a dry season between April and September and a rainy season between November and March (Randriamalala et al. 2007). The forested corridor is composed of a dense, midaltitude humid forest dominated by Weinmannia spp. (Cunoniaceae) and Tambourissa spp. (Monimiaceae; Koechlin et al. 1974). As one of the few remaining extensive tracts of forest in Madagascar, it is under strict conservation regulations since a 2007 provisional decree seeking to maintain both the forest’s highly endemic biodiversity, and its assumed function of a biological corridor between two national parks (Carrie`re-Buchsenschutz 2006). These national-level regulations imply a drastic reduction of clearing practices for slash-and-burn and other extractive activities, which particularly impact local-level subsistence agriculture and land use on the forest edge. Consequently, the limit between the forest corridor and the agricultural mosaic is easily perceived in the landscape as the limit between dense, homogeneous forest tracts and a patchy, alternately open and forested human-modified landscape. This agricultural landscape is currently highly heterogeneous. It consists of a finegrained mosaic of cultivated and non-cultivated habitats, including mature forest fragments, secondary forest at different succession ages, eucalyptus and other plantations, slash-and-burn fields on slopes, paddy fields in valley bottoms. On a sampled surface of reference of 3,320 ha in the study area, 960 patches of homogeneous habitat were identified by SPOT 5 satellite imagery from 2004, with 10 m resolution. MapInfo Professional 7.0 was used to analyze and identify habitat classes, with field verification of class identity in 2004–2005. Habitat patches (mean: 3.5 ha per patch) included dry crops 23.7%, herbaceous recruits 10.7%, irrigated rice fields 14.2%, reforestation areas 7.4%, forests 2.1% and shrubby recruits 41.8%. Open areas occupy 48.6% of these habitats. They are defined as areas either without dense vegetation cover, or where the height of dense vegetation is less than 1 m. They include slope and paddy fields, young fallows (\1 year), degraded low herbaceous regrowth after intensive slash-and-burn agriculture, tracks, housing areas and savannah. Trees Outside Forests (TOF), as defined by the Food and Agriculture Organization (Bellefontaine et al. 2001), are frequent in this landscape. They are identified as trees located in non forested land (farmland, grassland, villages and urban areas), within woody patches of less than 0.5 ha and with a height at maturity of at least 5 m. Within the TOF category, isolated trees in the midst of open areas are conspicuous features of the studied landscape. They are defined as single trees or small clumps of trees, with very low species richness, the edge of their crowns located more than 10 m from those of other vertical tree structures. Ethnobotanical data Ethnobotanical data were collected with the aim of sampling the diversity of forest edge situations. Categories and species of TOF were identified, including fruit trees in orchards, plantations of exotic species, isolated trees, live fences, small forest fragments. This paper focuses on isolated trees within open areas such as fields and young fallows outside villages. The area covered by this survey represents a third (1,110 ha) of the 3,320 ha surface of reference; it is located between 0 and 4 km from the forest edge. For each isolated tree with a minimum diameter at breast height (DBH) of 20 cm, the following data

123

Biodivers Conserv (2009) 18:2719–2742

2723

were collected: place name, GPS localization, scientific and vernacular names, ecological habitat, height and DBH. Ethnobotanical information was also recorded through interviews with local farmers and villagers over a period of 24 months (between June 2005 and June 2007). Interviews recorded the reason and local explanation for the tree’s presence, its uses, whether it was planted or spared during previous clearings, its place in Betsileo culture and/or economy. When the species was unknown, a sample was collected and identified at the Tzimbazaza Botanical and Zoological Parc (PBZT) in Antananarivo. Bird sampling Bird counts were carried out between April and May 2004 as part of a more extensive survey in the region. Bird sampling covered the same 1,110 ha landscape portion as the ethnobotanical study. Point counts selecting open areas and isolated trees lasted 5 minutes each and were separated by a distance of at least 200 m. Birds were sampled with an unlimited observation radius. We avoided restricting data collection to a given radius in order to ensure sufficient data for analysis, a preliminary survey having led to anticipate rare observation occurrences, although in practice bird visibility reached a maximum of 50 m around point counts. However, the distance between counts could have been insufficient to avoid double-counting. We thus kept a detailed record of bird localization, corresponding habitat and identifiable movements in order to avoid double-counting at neighbouring points. All counts took place between at least 100 and 4,000 m from the edge of the protected forest corridor and were performed between 6:00 and 10:00 a.m. by the same two observers. A total of 338 point counts were collected, including 109 in open areas without isolated trees and 229 in open areas with isolated trees. Each count recorded the species identified by sight or by sound, its abundance and the corresponding habitat. In sampled habitats, the visibility bias linked to differences in structure and vegetation density was relatively weak. In open areas without trees, counts were performed with the nearest tree always at least 100 m away. Isolated trees were sampled as such if the area around them contained no other trees on a radius of at least 100 m. In isolated trees, the area 50 m around the tree was visible for bird detection. However, in order to have clear information on the functional value of the tree itself, it appeared essential for subsequent interpretation to evaluate whether it was directly occupied by birds, in addition to the general effect of tree presence on diversity in the area patch. In order to obtain a clear measure of bird occupation of the isolated trees, we sampled an area within the extent of tree crowns, and counted only the birds directly occupying the trees. Moreover, birds detected within the visibility radius (approx. 50 m) around isolated trees were observed to systematically fly into the trees, either directly or after settling shortly on the ground nearby. This was consistent both with the observations of a preliminary study, and with behavioural descriptions by local experts and researchers (S. M. Carrie`re, pers. obs.). Consequently, measures of bird diversity directly occupying trees were considered reliable estimates of general bird diversity in open areas with trees, thus allowing lumped analyses of the data both for tree occupation, and for general effect of tree presence. Although this introduces possible underestimation of diversity values in areas with trees, this bias was observed to be negligible during sampling; estimates in areas with trees should nevertheless be considered conservative in our results. Tree species were identified at each point count and 167 counts in Ficus spp. were distinguished from 62 counts in non-Ficus spp. At the time of sampling, none of the isolated trees we sampled had fruit except one, a Ficus, which was at the end of its fruiting time.

123

2724

Biodivers Conserv (2009) 18:2719–2742

Bird nomenclature followed mainly Langrand (1995); Goodman and Benstead (2003) was also used for recent taxonomic updates. Data analysis Bird species were classified into main dietary and habitat groups based on a comprehensive bibliographic description of the species referenced in the study region, compiled from a number of authoritative sources (Appendix). Species that are specialists of humid habitats (marshes, rivers, paddy fields), mainly insectivores and piscivores, are very well documented locally as highly exclusive to the large and ever-growing network of irrigated paddy fields; this is confirmed by a parallel study (in prep.). Such fields are permanent monocultures with a priori very low potential for contributing to bird diversity in the agricultural mosaic, with the exception of these humid area specialists (seven species). As components of bird diversity that remain clearly unthreatened by current agricultural intensification, these birds were of little interest for the present study and were excluded from the analysis in order to avoid a bias in favour of paddy fields, hence of open habitats, which would not necessarily reflect the tolerance of other, less well known species for these habitats. Habitat classification was based on main, but not exclusive, habitat preferences. Species defined as forest species are often tolerant of some or many human-modified forest areas, except where indicated (see Table 3). Other habitat categories include generalist species and those with preference for open areas. Diet classification was based firstly on main diet occurrence as cited by bibliographic sources, leading to the identification of species that are mainly frugivores, granivores, insectivores and carnivores. In order to take into account functional aspects linked to seed dispersal by fruit consumption, a supplementary category (fruit-eaters) was defined as including species that consume fruit, irrespective of its proportion in the diet. In tropical areas, detection of bird species by point counts is frequently limited by habitat complexity (Terborgh et al. 1990). Biases can occur both because of differences in habitat structure between sampling points and because of heterogeneity in individual species’ detectability (Boulinier 1998). Although counts were performed in open areas with good visibility and homogeneous vegetation structures, it is necessary to account for species that were not detected during the sampling. Completeness of the samples was assessed by estimating overall true species richness in each habitat. We used the nonparametric first-order jackknife method originally designed for capture-recapture data (Burnham and Overton 1979). When applied to point count data, the jackknife estimator allows detection probabilities to vary among species and among habitats (Boulinier 1998), thus accounting for spatial heterogeneity in the study area. This estimator was well adapted to our data because of its good performance when the number of rare species is low (species detected in only one or two samples) (Chao 1987). Overall species richness estimators and their standard errors were computed using EstimateS software (Colwell 2006) with 50 randomizations. Taxon sample-based rarefaction curves were used for comparison between habitats of species density and of species richness after rescaling to the number of individuals, as recommended by Gotelli and Colwell (2001). Sample-based curves were chosen instead of individual-based curves because they account for natural patchiness in species’ distribution and are relevant for sample data (Gotelli and Colwell 2001). Population density was also compared by plotting the number of individuals against the number of samples. Point count data were then analyzed among bird groups between the two habitat categories (open areas and isolated trees) and between the two categories of isolated tree

123

Biodivers Conserv (2009) 18:2719–2742

2725

species (Ficus spp. or non-Ficus spp.). This method allowed testing differences in species richness and abundance among functional bird groups in addition to pooled species samples. Because of variable non normality, we used the non-parametric Wilcoxon rank sum test to compare means of observed species richness per point count. Means of the number of individuals per point count were compared using a generalized linear model for count data (Poisson distribution and log link) with F significance tests (McCullagh and Nelder 1989), corrected for overdispersion. The Bonferroni correction for type I errors was used to calculate significance of differences between habitats. Estimates are given ± standard errors. In addition to the response of functional bird groups, we considered the response of individual species to habitat differences (MacNally et al. 2000). Species by species response was analyzed by comparing the mean number of individuals of each species per point count between habitat types and isolated tree species. We used a Poisson linear regression corrected for overdispersion to test this response. The Bonferroni correction for type I errors was again used to calculate significance of species-by-species response. Estimates are given ± standard errors. Estimates and rarefaction curves were computed using EstimateS v.8.0.0 software (Colwell 2006). Statistical tests were performed with R Statistical Software v.2.8.0 (R Development Core Team 2007).

Results Place and functions of isolated trees In this section, we describe the nature and functions of different categories of Trees Outside Forests (TOF) in the study area, which included various types of forest fragments, regenerated fallows, hedges, live fences and fruit orchards, in addition to isolated trees. Trees Outside Forests were located within several landscape elements: on cultivated land (tanimboly), in fallows, in eucalyptus, acacia or pine plantations, and on uncultivated land as in the case of forest fragments (songon’ala, ‘‘lock of hair’’) recorded as never having been cleared. Forest fragments include ‘‘primary’’ forest sites (valamaty), old houses, sites of ancient villages (tananahaolo) where isolated Ficus trees are numerous, and where utensils used to wash and makeup the dead (fanarinandra) must be deposited; they also include tombs (fasana) surrounded by ancient sacred groves. Some structures, such as the abundant hedges and live fences (fahitra or fefy), can be either planted or natural growth, and serve as protection for crops against pests and zebu cattle grazing, or as protection around stockyards. These hedges and other TOF also serve as reserves for firewood and construction wood. In this study, only single trees isolated within open areas were recorded. Overall, 48 trees were observed in dry crop fields, in villages and in herbaceous fallows (named ‘‘pseudosavannahs’’ because of their low rates of return to a woody state; Randriamalala et al. 2007), representing a maximum of 20 different species (Table 1). Trees were dispersed throughout the hilly landscape, which follows an organized pattern of land allocation between humid valley bottoms, slopes and dryer hilltops. Isolated trees were thus located either near villages and stockyards, in the first ring of dry crops (tambina) around irrigated rice fields of valley bottoms, in fields of dry crops on the hill slopes (tanety), or in herbaceous pastures on the hilltops (tamboho). They occupied formerly cultivated fields, which had reverted to an herbaceous state, as well as newly cleared fields where they were spared during clearing.

123

2726

Biodivers Conserv (2009) 18:2719–2742

Isolated tree distribution in the landscape is not random. In addition to socioeconomic, cultural and symbolic factors influencing the trees’ presence, tree distribution depends on original presence in the forest before clearing, on geophysical factors and on land-use attributes. For instance, Ficus tiliifolia are mostly found on hill slopes (in tanety), both because they are well-adapted to these drier areas and because they are preferentially maintained there by farmers; whereas the hygrophilous Pandanus spp. mostly grow in humid valley bottoms near irrigated rice fields (Table 1). One or several isolated trees are systematically found in each tanety or valley of the study area, with a high proportion of large, conspicuous Ficus trees. Isolated trees can be planted from seeds, grown from cuttings or spared during the slashing of forest while creating a slash-and-burn field (tavy). In Betsileo societies, planting a tree is not a trivial act: it is highly symbolic and only useful or symbolically significant trees are planted. Several factors also influence the decision to protect a tree (Table 1). They include historical and political obligations as well as traditional beliefs or taboos (e.g., trees possessing supernatural forces in connection with the ancestors; official prohibitions or fady; the ‘‘magic’’ force transmitted to the trees by traditional healers or ombiasy; some trees’ ability to ward off bad luck or evil spirits). Other reasons for protecting isolated trees are their usefulness for symbolic protection of fields, crops and stockyards, the lack of time for clearing, the provision of shade in fields and pastures, their role as live fences, the need for land boundary markers and economic purposes such as the production of edible fruit or of medicine (Table 1). Examples of useful trees are Anthocleista spp., which acts as a lightning rod in fields; Pandanus spp., which is widely used in basketry and mat construction; and Ficus spp., the most abundant, accounting for more than 33% of recorded trees (Table 1). The symbolic connotation of Ficus trees is very strong as they symbolize the highly respected Betsileo great ancestors. The respect people show to these trees is a measure of their respect for their ancestors. For many generations, Ficus trees were often kept and protected in the name of ancestrality throughout successive clearing phases. They are thus markers of formerly cultivated territories, old stockyards (valamaty) or ancient villages (tananahaolo) inhabited in the era of the Betsileo kingdoms (hova). There is no official prohibition (fady) to cut them, but those who choose to do so may incur many reprisals by evil spirits. For example, cutting a Ficus tiliifolia may stop lactation of a woman of childbearing age or provoke blindness for a family member. Other Ficus trees, such as Ficus reflexa, are often found near stockyards where they act as live fences if pruned, or as symbolic protectors of the cattle, in this region where cattle robbery is a constant threat. Bird point counts In our point counts, 749 individuals were sighted representing 21 species. Of these, 8 were found in isolated trees only, 3 were in open areas only and 10 were in both habitats. These 21 species represented 66% of the 32 species found in the agricultural mosaic, and 41% of the 51 species found in the study region including the corridor (humid-area specialists excluded, cf. Methods; unpublished manuscript). We recorded 8 forest species, all endemic, representing 47% of forest species in the mosaic and 30% of forest species in the study region. Mean overall estimated species richness was lower in open areas without isolated trees than in isolated trees (15.97 ± 1.70 and 20.99 ± 1.72, respectively; with standard errors of overall estimates). With the standard error of the difference of means estimated by H (1.72 ? 1.722) = 2.41, the difference of means (D = 5.02 ± 2.41) presents a strictly

123

kininina fanerana hazombato

sampin

vakoandrano

dendemilahy

dendemivavy

Local name

Moraceae

Ficus tiliifolia Baker 18 (cauliflorous)

Strelitziaceae

Moraceae

Moraceae

17 Ficus tiliifolia Baker

19 Ficus trichopoda Baker Ravenala madagascariensis 20 Sonn.

nonoka zamborizany

Moraceae Myrtaceae

15 Ficus reflexa Thunb. 16 Eugenia jambos Linn.

ravinala

aviavy

voarandambo

voara be

volomborona fahavalonkazo

1

16

1 1

1 1

1

field 1 field, fallow, disturbed forest 3 48

fallow

field pasture fallow near old village and stockyard field fields, shrubby and herbaceous fallows, pastures, orchards

field

mananitra, hazotokana

1 1

6.25 100.00

2.08

2.08

33.33

2.08 2.08

2.08 2.08

2.08

2.08 2.08

2.08

field field

fallow, close to a tomb 1

2.08 2.08 2.08

6.25

4.17

4.17

14.58

% per species

4.17

1 1 1

3

2

2

7

Nb. of ind.

2

fallows

field fallow orchard

fallows and fields

burnt forest riparian forests, edges of rice fields, humid

Habitat shrubby and herbaceous fallows, fields

hazondrano fompohona

Brachylaena ramiflora DC. 12 (Humbert) Asteraceae Albizia gummifera (J.F. Gmelin) 13 C. A. Smith Mimosaceae 14 Indeterminata

Aquifoliaceae Moraceae

Harungana madagascariensis Lam. ex Poir. Clusiaceae harongana Abrahamia oblungifolia (Engl.) Randrianasolo & Lowry Anacardiaceae taratambavy

10 Ilex mitis L. Radlk. 11 Ficus sp.

9

8

Eucalyptus robusta Sm. Myrtaceae Psorospermum fanerana Baker Clusiaceae Deinbollia pervillei Radlk. Sapindaceae

5 6 7

Pinaceae

Pinus spp.

4

Loganiaceae

Loganiaceae

Pandanus acanthostylus Martelli Pandanaceae

Anthocleista amplexicaulis Baker Anthocleista madagascariensis Baker

Family

3

2

1

N° Scientific name

medicine (malaria) firewood, construction

firewood

firewood

firewood

mats, baskets planks, fire lighter planks, charcoal

commemorative funeral stone and tomb marker, symbolizes rapid growth (etym.)

boundary marker boundary marker, commemorative funeral stone marker sacred tree

boundary marker

id.

lightning rod, weakens one's enemies, opposes evil spirits

Social and symbolic value

boundary marker, strong ancestral symbolic value, prejudice if cut used as hazomanga: ritual posts in villages

boundary marker, strong ancestral symbolic value, prejudice if cut, symbolizes respect for ancestors

shade, ornamental, fruit, herb tea, pharmacopoeia, basketry, attracts birds and bats shade, ornamental, fruit, herb tea, pharmacopoeia, basketry, attracts birds and bats shade, ornamental, fruit, herb tea, basketry, attracts birds and bats ornamental, ecological humidity level indicator

attracts birds

symbolizes a bird's crest (etym.) sacred plant used by soothsayers boundary marker, stockyard marker, symbolic stockyard protection

leaves used to ripen bananas pharmacopoeia

chases evil spirits from houses fruit, construction (musical instruments, knives) boundary marker attracts birds for hunting etym. (hazotokana ): "isolated tree in the fields". Sacred tree, commemorative funeral stone marker construction wood, pharmacopoeia

shade, construction wood, honey, fruit, pharmacopoeia

firewood, construction wood charcoal, construction wood, stretchers, firewood, pharmacopoeia construction wood, fruit, pharmacopoeia support for grape vines

mats, baskets, tents

id.

Economic uses Cultural uses shade, ornamental, prevent thunder, ward off evil spirits, pharmacopoeia, pestles for shelling rice, charcoal

Table 1 Place and function of isolated trees in Betsileo society. Etymology (etym.) is referred to if the local name has symbolic meaning

X

X

X

X

X

X

X

X

X

X X

X

X X

X

X

X X X

X

X

Planted Spared

Biodivers Conserv (2009) 18:2719–2742 2727

123

2728

Biodivers Conserv (2009) 18:2719–2742

positive 95% confidence interval, thus showing a significant difference between these values. Detection probabilities were 81.4% in open areas and 85.8% in isolated trees. Overall, open areas also had lower observed species density of 1.14 ± 0.08 species per sample than isolated trees, with 1.32 ± 0.04 species per sample (Wilcoxon rank sum test, W = 14,182.5, P = 0.011; Table 2). Rarefaction curves confirm these results independently of differences in sampling effort, suggesting that species density and species richness were lower in open areas than in isolated trees (Fig. 1a, b), with both curves nearing an asymptote. Population density was higher in open areas (Fig. 1c) with a non-significant difference between values of 2.53 ± 0.17 and 3.49 ± 0.86 individuals per sample in isolated trees and open areas, respectively. Bird groups Carnivores, frugivores, fruit-eaters and forest fruit-eaters had more species and individuals in isolated trees than in open areas (Table 2). Furthermore, among these groups, only 3 species out of 7 were found at any point in open areas without trees (Centropus toulou, 4 individuals; Hypsipetes madagascariensis, 1 ind.; Acridotheres tristis, 13 ind.; Table 3). Granivores had significantly more species in isolated trees than in open areas, but abundance was higher in open areas than in isolated trees. Insectivores were more numerous and had significantly more species in open areas. The 4 species of fruit-eaters recorded in these habitats represent 44% of the 9 fruiteaters found in the agricultural mosaic, and 29% of the 14 fruit-eaters recorded in the study region (unpublished manuscript). Three of these are recorded as forest fruit-eaters (H. madagascariensis, Zosterops maderaspatana, Leptopterus chabert; Table 3). They were present almost exclusively in isolated trees, with only one individual in open habitats (H. madagascariensis). Differences in abundance between habitats were significant for H. madagascariensis and Z. maderaspatana. Forest species were significantly more numerous in areas with isolated trees (in species richness and abundance; Table 2). Habitat generalists presented no significant difference between the two habitats in species richness; however, abundance was higher in open areas. Open habitat species were found in higher species richness and abundance in open areas. Species by species analysis Out of the 21 bird species, 8 (38%) were more abundant in isolated trees than in open areas without trees (Table 3), with at least 75% of individuals per species in isolated trees (not considering species with less than 10 individuals); 4 of these (Falco newtonii, H. madagascariensis, Nectarinia notata, and Z. maderaspatana) presented a significant difference between values. These 21 species include 8 endemic forest species representing 299 detected individuals, of which only 5 were found at any point in open areas, all others having been detected only in isolated trees (1 H. madagascariensis, 3 Terpsiphone mutata, 1 Nectarinia souimanga; Table 3). Overall, 8 species out of 21 were found exclusively in isolated trees, including 4 with sufficient sampling numbers (F. newtonii, N. notata, Z. maderaspatana, L. chabert; Table 3), and 4 other species had at least 75% of detected individuals in isolated trees (C. toulou, Nesillas typica, N. souimanga and Lonchura nana). One habitat generalist, F. newtonii, was present only in isolated trees (n = 11; Fig. 2) and one open habitat species, L. nana, was also more numerous in isolated trees. All other

123

0.29 ± 0.03

0.48 ± 0.03

0.42 ± 0.04

0.11 ± 0.02

21

2

2

13

All species

Frugivores

Granivores

Insectivores

0.48 ± 0.04

0.32 ± 0.03

3

8

8

5

Forest fruit-eaters

Forest species

Habitat generalists

Open habitat species

0.72 ± 0.06

0.38 ± 0.06

0.05 ± 0.02

0.02 ± 0.01

0.07 ± 0.03

0.03 ± 0.02

0.69 ± 0.07

0.35 ± 0.05

0.01 ± 0.01

1.14 ± 0.08

Open areas (n = 109)

10,678

12,329

16,853

15,964.5

15,625.5

13,499.5

9,960

14,130

15,693.5

14,182.5

W

\0.001** 0.014**

478

183

213 299

\0.001** 0.823

30 237

0.011** \0.001**

373 390

0.022** \0.001**

960 143

0.011**

N

\0.001**

P

0.83 ± 0.07

0.42 ± 0.05

1.28 ± 0.16

0.93 ± 0.14

0.97 ± 0.14

0.11 ± 0.02

1.00 ± 0.14

0.75 ± 0.07

0.62 ± 0.09

2.53 ± 0.17

Isolated trees (n = 229)

2.64 ± 0.85

0.80 ± 0.15

0.05 ± 0.02

0.02 ± 0.01

0.13 ± 0.05

0.04 ± 0.02

1.47 ± 0.22

1.85 ± 0.85

0.01 ± 0.01

3.49 ± 0.86

Open areas (n = 109)

Individuals per point count

0.0013**

0.0285**

\0.001**

\0.001**

\0.001**

0.2496

0.7236

0.2329

\0.001**

1.0000

P values (glm)

** Denotes statistical significance of tests (P \ 0.05)

Total number of species (S) and individuals (N) in each group are also given. Wilcoxon test statistic (W) and P values (P) are given for species richness comparison. P values of the Poisson generalized linear model are given for abundance comparison, accounting for overdispersion and Bonferroni correction

0.52 ± 0.03

0.33 ± 0.04

0.35 ± 0.04

3

4

Carnivores

Fruit-eaters

1.32 ± 0.04

Isolated trees (n = 229)

Species per point count

S

Bird group

Table 2 Mean species richness and abundance per point count in isolated trees and open habitats among bird groups, expressed as mean ± SE

Biodivers Conserv (2009) 18:2719–2742 2729

123

2730

(a) Number of species

Fig. 1 Sample-based rarefaction curves showing estimated a bird species richness, b bird species density and c bird population density in open areas and in open areas with isolated trees

Biodivers Conserv (2009) 18:2719–2742

20

isolated tree

15

open area 10 5 0 0

200

400

600

800

Number of individuals

Number of species

(b)

20 isolated

15 open area

tree

10 5 0 0

50

100

150

200

250

Number of samples

Number of individuals

(c) 700

isolated

600

tree

500

open area

400 300 200 100 0 0

50

100

150

200

250

Number of samples

habitat generalists showed non-significant difference between the two habitats, except for Cisticola cherina which was detected only in open areas (n = 42; P \ 0.001). All other open habitat species were significantly more abundant in open areas than in isolated trees, including Mirafra hova which was present only in open areas. Independently of main diet classification, out of five seed-eating species, two were significantly more numerous in open areas (M. hova, insectivore, and Foudia madagascariensis, granivore), two were present only in these areas (Numida meleagris and M. hova, insectivores) and L. nana, granivore, was mostly found in isolated trees. Isolated tree species Isolated Ficus trees had slightly higher overall bird species density than non-Ficus trees, with, respectively, 1.40 ± 0.05 and 1.15 ± 0.04 species per sample (Wilcoxon rank sum

123

Biodivers Conserv (2009) 18:2719–2742

2731

test, W = 6,064.5, P = 0.008). However, no separate bird group showed significant difference in species density or abundance between tree species. Five individual species showed significant difference of the test between Ficus and nonFicus trees (Table 3), although the number of observations was often too low for interpretation. Of these, one carnivore (Accipiter henstii, n = 3), 2 strict insectivores (Merops superciliosus, n = 8; Tylas eduardi, n = 1) and 1 omnivore (Acridotheres tristis, n = 11) were more abundant or detected only in non-Ficus trees; 1 strict insectivore (Nesillas typica, n = 53) was more abundant in Ficus trees.

Discussion Ethnographic interviews and observations show that isolated trees hold special economic, cultural and symbolic value for Betsileo farmers. They are frequently spared or planted within open areas with generally high symbolic significance, various levels of protection and considerable impact on bird diversity in the landscape. The 48 recorded individuals were sampled in view of accounting for the diversity of situations, species and uses linked with isolated tree management. They thus represent only a fraction of total isolated trees in this landscape, given in particular the fact that sampling did not account for formerly isolated trees, now integrated within regenerated fallows. Why and how are isolated trees protected? Few isolated trees are under clear taboos for protection: reasons for protection vary depending upon the species considered. Eight out of 20 recorded species (40%) were protected partly for economic reasons, because their products may add valuable income to rural households through the sale of firewood, construction wood, basketry, charcoal and even medicine (Brachilaena ramiflora; Table 1: sp. 12). However, all these economically useful trees also have importance in Betsileo culture, daily use or symbolism. Most recorded species combine cultural and symbolic importance, which ranges from ornamental value, shade provision, medicinal or hunting uses, to land marking functions or the protection of humans from spirits (Table 1). These values of isolated trees are frequent in several tropical landscapes, where they are retained for timber, firewood, fodder, the provision of fruit and shade, and other daily, economic and cultural uses. Among isolated tree species of our study area, only F. tiliifolia (Table 1: sp. 17 and 18) is protected by socially enforced reprisals in case of cutting, although these do not represent a formal prohibition or fady. This species is also among the 5 species recorded as planted instead of spared (this represents 25% of recorded species, among which Pinus spp. and Eucalyptus robusta are plantation trees). In the case of Ficus spp., planting is an act with no economic value, but strong practical, cultural and symbolic implications. Their high proportion in the landscape (41.7% of recorded species) is linked to their preferential planting (especially of F. tiliifolia) and is a measure of their local importance. In addition to many practical and cultural uses, they are highly respected ‘‘ancestor trees’’ and are frequently used for rituals, for instance as sacred posts (hazomanga) during village ceremonies (F. trichopoda, Table 1—species 19). Such symbolic importance of Ficus trees is found throughout the Old World (Gadgil and Berkes 1991; Berkes et al. 1995), where these trees are afforded widespread protection for their ability to provide fruit, materials, and various ecological services such as the provision of shelter for birds and other animals.

123

123 Insectivore (I, N)

For

Nectariniidae

R

Insectivore (I, N)

For

Nectarinia souimanga Gmelin

Insectivore (I)

Insectivore (I)

Insectivore (I)

Insectivore (I)

For

Ge

O

Frugivore (F, I)

Insectivore (I)

Insectivore (I, G)

Insectivore (I)

Insectivore (I)

Muscicapidae Terpsiphone mutata Linne´ M Nectariniidae Nectarinia notata Mu¨ller M

Cisticola cherina Smith

Sylviidae

M

n

10 0.00 ± 0.00

0.01 ± 0.01

0.00 – . 0.00

49 0.21 – 0.06 24 0.10 ± 0.03

0.03 ± 0.02

0.39 ± 0.09

0.17 ± 0.05

9 0.03 ± 0.02

42 0.00 ± 0.00

71 0.23 ± 0.03

0.25 ± 0.08

0.01 – 0.01

75 0.32 – 0.05

41 0.06 ± 0.02

0.37 ± 0.08

0.17 ± 0.11

0.00 ± 0.00

0.00 ± 0.00

45 0.02 ± 0.01

19 0.00 ± 0.00

1 0.00 ± 0.00

8 0.03 ± 0.03

0.04 ± 0.02

0.09 ± 0.09

0.00 ± 0.00 0.00 – 0.00

Open areas (n = 109)

0.01 ± 0.01

Isolated trees (n = 229)

11 0.05 – 0.01

3

N

Individuals per point count

Carnivore (I, V; 16 0.05 ± 0.01 birds)

Ge

Nesillas typica Hartlaub

Sylviidae

For

O

O

Ge

Ge

Ge

Insectivore (I,G)

R

Saxicola torquata Linne´

M

Pycnonotidae Hypsipetes madagascariensis Mu¨ller

Turdidae

M

Motacilla flaviventris Hartlaub

M

Motacillidae

M

Upupa epops Linne´

Mirafra hova Hartlaub

n

Merops superciliosus Linne´

Meropidae

Alaudidae

M

Centropus toulou Mu¨ller

Cuculidae

Upupidae

i

Numida meleagris Linne´

Numididae

Ge

M

Falco newtonii Gurney

Falconidae

Carnivore (I, V)

For* Carnivore (V)

Ge

Accipiter henstii Schlegel M

Accipitridae

End. Hab. Diet

Species

Family Ficus (n = 167)

Non Ficus (n = 62)

0.06 – 0.03

0.236

0.002**

1.000

0.05 – 0.05 23 0.08 ± 0.03 0.15 ± 0.07

49 0.28 – 0.08

6 0.01 ± 0.01 0.08 ± 0.07

0 0.00 ± 0.00 0.00 ± 0.00

\0.001**

14 0.07 ± 0.03 0.03 ± 0.03 53 0.29 ± 0.04 0.06 ± 0.05

1.000

0.063

0.23 – 0.06

5 0.02 ± 0.02 0.02 ± 0.02

\0.001**

74 0.36 – 0.07

0 0.00 ± 0.00 0.00 ± 0.00

<0.001**

1 0.01 ± 0.01 0.00 ± 0.00

1.000

8 0.00 ± 0.00 0.13 ± 0.13

12 0.06 ± 0.02 0.03 ± 0.02

0 0.00 ± 0.00 0.00 ± 0.00

11 0.04 – 0.02

3 0.00 ± 0.00 0.05 ± 0.04

N

\0.001**

1.000

1.000

0.175

0.007**

1.000

P values (glm)

Table 3 Species-by-species response to isolated tree presence in open areas, and to isolated tree species (Ficus spp./non Ficus spp.)

1.000

0.899

0.534

1.000

0.035**

1.000

1.000

1.000

1.000

1.000

0.043**

1.000

1.000

1.000

\0.001**

P values (glm)

2732 Biodivers Conserv (2009) 18:2719–2742

Foudia madagascariensis Linne´

Lonchura nana Pucheran

Ploceidae

Estrildidae

O

M

O

Ge

i

M

Insectivore (F, I)

Granivore (G)

Granivore (G, I, N)

Omnivore (F, G, I, V)

For* Insectivore (I)

For

Frugivore (F, I, N, flo)

Isolated trees (n = 229)

198 0.69 ± 0.06

0.37 ± 0.11

1.49 ± 0.85

0.12 ± 0.04

175 0.06 ± 0.03

0.00 ± 0.00

1 0.00 ± 0.00

0.00 ± 0.00

0.00 – 0.00

Open areas (n = 109)

24 0.05 ± 0.02

70 0.31 ± 0.11

68 0.30 – 0.07

N

Individuals per point count

0.171

0.003**

1.000

1.000

0.074

<0.001**

P values (glm)

Ficus (n = 167)

Non Ficus (n = 62)

158 0.80 ± 0.08 0.40 ± 0.08

13 0.05 ± 0.03 0.08 ± 0.06

11 0.01 ± 0.01 0.16 ± 0.07

1 0.00 ± 0.00 0.02 ± 0.02

70 0.17 ± 0.09 0.66 ± 0.33

68 0.34 – 0.08 0.19 – 0.12

N

0.070

1.000

0.002**

0.043**

1.000

1.000

P values (glm)

** Statistical significance of the test for differences of bird abundance between categories, accounting for overdispersion and Bonferroni correction (P \ 0.05)

Diet classification is based on estimated relative importance of diet elements. Diet elements are given in brackets (F fruit, G seeds, I insects, V vertebrates (non bird), N nectar, flo flowers)

Habitat (Hab.) For*: strictly dense forest tracts, For: mainly forest with tolerance for disturbed or human-dominated forests, Ge generalist (dense or degraded forest, open and human-dominated areas), O open areas (clearings, savannah, crop fields, rice paddies, marshes)

Endemism (End.) M Madagascar, R region (Comores, Mascareignes, Seychelles), i introduced, n non endemic

Bird abundance is expressed as the total number of individuals recorded in isolated trees and open areas (N) or isolated trees only (n). For each measured category, abundance is expressed as mean ± SE of the number of individuals recorded per point count. Species in bold were significantly more abundant in isolated trees than in open areas without trees

Acridotheres tristis Linne´

Tylas eduardi Hartlaub M

Vangidae

Sturnidae

M

Leptopterus chabert Mu¨ller

Vangidae

For

End. Hab. Diet

R

Species

Zosteropidae Zosterops maderaspatana Linne´

Family

Table 3 continued

Biodivers Conserv (2009) 18:2719–2742 2733

123

2734

Biodivers Conserv (2009) 18:2719–2742

Isolated trees enhance bird diversity in the agricultural landscape This study shows that up to 18 species of birds out of 21 (86% of sample) may be attracted to isolated trees in open, cultivated areas, including 12 (57%) which were otherwise not found or less frequent in open areas. This includes four species (three endemic forest species) which were significantly less frequent in open areas without trees, while the eight endemic forest species found occupying isolated trees showed very low presence in open areas. Overall, species density and richness were higher in isolated trees than in open areas without trees, thus amounting to a positive effect of tree presence on bird diversity in open areas (cf. Methods). This was true for frugivores, fruit-eaters and carnivores (Table 2), consistent with their dietary and behavioural habits: trees are the main source of fruit in these areas and serve as look-outs for prey observation by carnivores (Thorstrom et al. 2003). Species referenced as forest birds were often tolerant of open areas only in the presence of trees. However, consistent with these species’ dependency upon tree habitats (Raherilalao 2001; Goodman and Raherilalao 2003), their presence in these areas implies higher tolerance than could be expected of forest birds for what is frequently considered extremely degraded habitat (Randriamalala et al. 2007). This result thus confirms qualitative observations on forest species such as H. madagascariensis and Z. maderaspatana, which are known to have well adapted to human modification of habitats (Langrand 1995; Goodman and Benstead 2003). Despite this, however, the presence of trees, even isolated, does remain a necessary condition for forest birds’ occupation of the agricultural landscape. Inversely, insectivores and granivores were more abundant in open areas without trees, than in isolated trees (Table 2). This is consistent with dietary preferences of granivores (Foudia madagascariensis, L. nana) and seed-eating insectivores (Numida meleagris, M. hova; Table 3), seeds being most frequent in grass-rich open areas. Although species density was slightly higher for granivores in isolated trees (P = 0.022; Table 2), this may be linked with flocking habits of L. nana, which were often found perching in groups in isolated trees (Fig. 2), whereas this species was more sparsely distributed in the wider open areas. Higher frequencies of non seed-eating insectivores in open areas (Motacilla flaviventris, Saxicola torquata, Cisticola cherina) could be linked to preferred insect species’ availability. Open habitat birds were among the most abundant species and could thus account for higher abundance values in open habitats. However, results showing higher abundance in open areas should be taken with caution because of the sampling bias previously described. These results are consistent with most available studies worldwide. In temperate areas, birds and other taxa were found to benefit from isolated trees (e.g., Fischer and Lindenmayer 2002; LeCœur et al. 2002; Lumsden and Bennett 2005). Fewer studies in the tropics document tree utility for several taxa including bird, bat and insect presence in agricultural mosaics (e.g., Dunn 2000; Luck and Daily 2003), where the contribution of even the smallest habitat patches such as isolated trees to overall landscape heterogeneity appears to be an important factor for biodiversity conservation (Bodin et al. 2006). Birds are widely used in these studies and are considered major indicators of landscape conservation value, both because of their relative ease of sampling, and because of their central position in ecological functioning, particularly for frugivores (Luck and Daily 2003; Bleher and Bo¨hning-Gaese 2001; Shanahan et al. 2001). However, to our knowledge and despite widely advertised conservation efforts and urgency in Madagascar (Peters 1998), available results on birds and general biodiversity in Malagasy agricultural

123

Biodivers Conserv (2009) 18:2719–2742

2735

Fig. 2 Response of bird species to isolated tree presence in open areas, expressed as the number of individuals detected per habitat. Total species abundance is given in brackets. Habitat preference classification is based on bibliographic reference (see Appendix)

landscapes are surprisingly rare (but see Watson et al. 2004, for edge and matrix effects on birds; Bodin et al. 2006, for bee and lemur diversity in an agricultural landscape), and no previous studies have focused on the role of isolated trees for biodiversity, despite their conspicuousness and sociocultural importance in several Malagasy landscapes (Styger et al. 1999). Ecological function of isolated trees By attracting birds and other frugivorous animals into cultivated areas (Luck and Daily 2003), isolated trees may contribute considerably to the maintenance of ecosystem processes in these areas. According to several studies, increased animal movements and seed deposition by frugivores contribute to a local increase of forest tree regeneration (e.g., Carrie`re et al. 2002; Guevara et al. 2004). Isolated trees are then considered foci of regeneration for persistent species that may coalesce with others according to the nucleation model (Yarranton and Morrison 1974; Laborde et al. 2008). By enhancing landscape connectivity and providing stepping stones for animals between heterogeneous habitat patches (Manning et al. 2006), ‘‘attractive’’ isolated trees may also contribute to the viability of remnant forest patches in the agricultural landscape, by allowing gene flows between isolated patches and the nearby forest (Bacles et al. 2006; Jordano 2007). The importance of Ficus spp. for these processes is likely to be particularly strong (Guevara et al. 2004), because Ficus trees are shown to be keystone resources worldwide for frugivores due to their year-round and abundant fruiting phenology (Fleming et al. 1987; Goodman and Ganzhorn 1997; Shanahan et al. 2001). In Madagascar, where

123

2736

Biodivers Conserv (2009) 18:2719–2742

endemic frugivore bird species are particularly depauperate, this is of special conservation and biogeographical interest (Fleming et al. 1987; Leigh et al. 2007). Moreover, the diet of most fruit-eating birds is highly mixed in this region and frequently includes a large variety of other food types (e.g., seeds, insects and other invertebrates; cf. Langrand 1995). However, such mixed diets are often a symptom of high spatiotemporal patchiness of food resources (Fleming et al. 1987), to which these birds may thus have been exposed. In view of the keystone role of most Ficus trees for frugivores worldwide, Madagascar’s paucity in specialized frugivores has been linked by some authors to the island’s low availability in endemic Ficus species, as compared to other regions of the Old World (Goodman and Ganzhorn 1997). Human-induced, accentuated presence of Ficus trees in some areas could thus contribute to lowering the spatiotemporal patchiness of food resources for frugivores, potentially impacting diet speciation processes if these practices are maintained on an evolutionary timescale. In these circumstances, preference for and symbolic importance of isolated Ficus trees (especially F. tiliifolia) for Betsileo farmers and others worldwide (Gadgil and Berkes 1991) is particularly intriguing and relevant to conservation, both of fruit-eating animals and of ecosystem processes. Indeed, what may have brought about such practices and values concerning Ficus trees? Are the ecological implications conscious or unconscious elements of traditional regulations? In our study, comparison of bird diversity between Ficus spp. and other isolated tree species showed higher species density in Ficus trees, which may imply a leading role of Ficus for bird conservation in isolated trees. However, the study did not reveal a clear difference among guilds; in particular, fruit-eaters showed no clear preference for Ficus trees over other isolated tree species (Table 3). One reason for this could be the relative absence of fruiting Ficus, indeed of any fruiting trees in our sample except one. The fact that birds were still detected using isolated trees preferentially, particularly fruit-eating forest birds, is consistent with other studies where seed rain is shown to be independent of fruiting times (Carrie`re et al. 2002; Guevara and Laborde 1993). This is of interest because it indicates that fruit is not the only driver for birds to occupy these trees. Although other dietary aspects may play a role, such as seasonal caterpillar presence for birds with mixed diets (Carrie`re et al. 2002), such a case was not observed here. In addition to possible foraging and sheltering functions, these results thus indicate that isolated trees may have permanent function as stepping stones between habitats (Manning et al. 2006) and could even represent fixed, habitual routes for bird movement throughout the landscape (Van Dyck and Baguette 2005). Isolated trees in traditional ecosystem management For Berkes et al. (1995), social restraints such as traditional beliefs of habitat sacredness and respect for ancestral linkages can be vectors of (conscious or unconscious) preservation of biodiversity and environmental conditions. Even if environmental utility of traditional management practices is socially unconscious, they may still contribute to essential ecosystem services, thus reinforcing the need to document these practices before their disappearance following socioeconomic change (Berkes et al. 1995; Colding and Folke 1997; Berkes 2004). In keeping with the findings of this study, all recorded Ficus species are considered by Betsileo farmers to attract birds and other animals for hunting (Table 1: Ficus sp 11; F. reflexa 15; F. tiliifolia 17 and 18; F. trichopoda 19). Although the primary aim of this function is subsistence, not the enhancement of agroecosystem biodiversity, the latter is a

123

Biodivers Conserv (2009) 18:2719–2742

2737

conscious implication of Ficus trees’ protection. Such conscious management of biodiversity at the landscape level is found in many regions in the form of practices that enhance landscape heterogeneity and patchiness (Gadgil and Berkes 1991; Gadgil et al. 1993), for instance in Africa (Warren and Pinkston 1998), Asia (Sankhala 1993), and Oceania (Gadgil et al. 1993). Although this was not recorded in the course of this study, protection of isolated trees in open areas may also be a part of the farming system, with trees serving to regulate soil humidity and enhance fertility, in particular by favouring regeneration in fallows. Utility for crop management is recorded as a reason for maintaining isolated trees in fields in many areas (Altieri 2002), for instance for farmers in Cameroon (Carrie`re 2002), in India (Sankhala 1993) or in Costa Rica (Harvey and Haber 1999); and the ecological value of isolated trees for management of fertility levels, fallow rotations and plant regeneration is shown to be important in several cases (Harvey and Haber 1999; Bleher and Bo¨hningGaese 2001; Carrie`re et al. 2002; Guevara et al. 2004). In current Betsileo society, isolated trees’ utility for crop management may not be a conscious preservation factor, but may involve underlying, empirically developed and secular ecological knowledge, traditionally enforced through these social regulations (Colding and Folke 1997; Berkes et al. 2000). This is particularly consistent with the importance of ancestrality in the Betsileo belief complex (sensu Berkes 1999), and in light of how ancestrality is symbolized, even embodied by the centuries-old, protected Ficus, standing majestically isolated on cultivated hillsides.

Conclusion These results stress the importance of filling the gap in understanding biodiversity dynamics in Malagasy agricultural landscapes, particularly in the context of current management issues for conservation and development outside protected areas like the Ranomafana–Andringitra forest corridor. There is a need to clarify under what conditions these areas may act as buffer zones near protected areas, while still allowing local subsistence and economic development. This study shows that maintaining and promoting traditional practices of isolated tree management could contribute to the efficiency of bird conservation in Malagasy landscapes, particularly endemic forest birds. As a guideline for management very much in line with traditionally implemented practices, this result is particularly relevant for community-based natural resource management, represented in our study area by the CoBa. However, rapid social, cultural and economic change as currently observed in our study area often implies the loss of a great deal of traditional knowledge and practices. This is why we also stress the importance of documenting these practices and their relevance for conservation, in view of their further use within community-based or other management systems. This study is also a first account of Malagasy bird distribution in open, cultivated areas outside the forest. It represents a step towards understanding the resilience of bird populations in this human-modified landscape, thus furthering our ability to make pragmatic conservation choices in this area under high competition between land use and nature conservation. Although isolated trees are shown to impact favourably on bird presence in open areas of the agricultural mosaic, this does not imply, however, that they are sufficient to maintain stable populations outside the forest corridor (Daily et al. 2001). More work is needed to understand their dynamics, in relation to other patches of the mosaic and to

123

2738

Biodivers Conserv (2009) 18:2719–2742

neighbouring forest areas, and the importance of these dynamics for the sustainability of ecological processes. Acknowledgments This study was funded by the French Institut de la Recherche pour le De´veloppement (IRD). We are grateful to the communities of Ambendrana, Amindrabe´ and Igodona for allowing us to conduct botanical and bird surveys, as well as for their patience in answering our questions. We especially thank Marion Viano and our expert guide M. Rakotomaro for bird point counts, Ms. Modeste and Ms. Noly for local logistical support. We are grateful to Drs. Olivier Gimenez and Thierry Boulinier for providing scientific support and statistical advice in the course of the work leading to this article, as well as to Drs. Emmanuel Torquebiau, Jochen Krauss and two anonymous reviewers for their helpful comments on the manuscript.

Appendix List of bibliographic references used for diet and habitat classification of bird groups Billing TM, Mulder RA (2003) Centropus toulou, Madagascar coucal, Kotrohake, Monjo, Toloho. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1108–1110 Bo¨hning-Gaese K, Gaese BH, Rabemanantsoa SB (1995) Seed dispersal by frugivorous tree visitors in the Malagasy tree species Commiphora guillaumini. Ecotropica 1:41–50 Bo¨hning-Gaese K, Gaese BH, Rabemanantsoa SB (1999) Importance of primary and secondary seed dispersal in the Malagasy tree Commiphora guillaumini. Ecology 80:821– 832 Bollen A, Van Elsacker L (2004) The feeding ecology of the Lesser Vasa Parrot, Coracopsis nigra, in south-eastern Madagascar. Ostrich 75:141–146 Bollen A, Van Elsacker L, Ganzhorn GU (2004) Relations between fruits and disperser assemblages in a Malagasy littoral forest: a community-level approach. J Tropical Ecol 20:1–14 Carrie`re S, Roche P, Viano M, Iftice`ne E, Picot Manuel M, Tatoni T (2007) He´te´roge´ne´ite´ des paysages, dispersion des graines et biodiversite´ : le cas du terroir d’Ambendrana. In: Serpantie´ G, Rasolofoharinoro, Carrie`re S (eds) Transitions agraires, dynamiques e´cologiques et conservation: le corridor Ranomafana–Andringitra (Madagascar). CITE/IRD Ekstrom JM (2003) Psittaciformes: Coracopsis spp., parrots. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1098–1102 Goodman SM, Ganzhorn JU, Wilme´ L (1997) Observations at a Ficus tree in Malagasy humid forest. Biotropica 29:480–488 Goodman SM, Putnam MS (1996) Chapter 18. The birds of the eastern slopes of the Re´serve Naturelle Inte´grale d’Andringitra, Madagascar. In: Goodman SM (ed) A floral and faunal inventory of the eastern slopes of the Re´serve Naturelle Inte´grale d’Andringitra, Madagascar: with reference to elevational variation. Fields Museum of Natural History (Fieldiana Zoology new series no. 85), USA, pp 171–190 Goodman SM, Raherilalao MJ (2003) Effects of forest fragmentation on bird communities. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1064–1066 Goodman SM, Wilme´ L (2003) Cuculiformes: Coua spp., Couas. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1102–1108

123

Biodivers Conserv (2009) 18:2719–2742

2739

Hawkins AFA, Goodman SM (2003) Chap. 12. Birds. Introduction to the Birds. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1019–1044 Hawkins AFA, Seddon N (2003) Mesitornithidae, Mesites. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1095–1098 Langrand O (1995) Guide des oiseaux de Madagascar. Delachaux et Niestle´, p 415, Paris Prum RO, Razafindratsita VR (1997) Lek behavior and natural history of the Velvet Asity (Philepitta castanea: Eurylaimidae). Wilson Bull 109:371–560 Prum RO, Razafindratsita VR (2003) Passeriformes : Philepittinae, Asities and sunbirdasities. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1123–1130 Raherilalao MJ, Goodman SM (2005) L’avifaune de sous-bois de la foreˆt humide de montagne du couloir forestier entre les parcs nationaux d’Andringitra et de Ranomafana (Madagascar). Rev E´col (Terre Vie) 60:167–178 Raherilalao MJ, Razafindratsita V, Goodman SM, Rakotoniaina JC (2001) Chapitre 8. L’avifaune du Parc National Ranomafana et du couloir forestier entre Andringitra et Ranomafana. In: Goodman SM, Razafindratsitra VR (eds) Inventaire biologique du Parc National de Ranomafana et du couloir forestier qui le relie au Parc National d’Andringitra, vol. 17. Recherches pour le De´veloppement—Se´rie Sciences biologiques, Antananarivo, Madagascar, pp 165–195 Rakotomanana H, Hino T (1998) Fruit preference in the Velvet Asity Philepitta castanea in a rain forest of Madagascar. Jpn J Ornithol 47:11–19 Rakotomanana H, Hino T, Kanzaki M, Morioka H (2003) The role of the Velvet Asity Philepitta castanea in regeneration of understory shrubs in Madagascan rainforest. Ornithol Sci 2:49–58 Rakotomanana H, Rene de Roland L-A (2007) Breeding ecology of the endemic, Madagascan, Velvet Asity Philepitta castanea. Ornithol Sci 6:79–85 Ravokatra M, Wilme´ L, Goodman SM (2003) Bird weights. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1059–1063 Rene de Roland L-A, Thorstrom R (2003) Accipiter spp., Goshawk and sparrowhawks. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1091–1094 Schulenberg TS (2003) Vangidae, Vangas. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1138–1143 Shanahan M, So S, Compton SG, Corlett R (2001) Fig-eating by vertebrate frugivores: a global review. Biol Rev 76:529–572 Thorstrom R, Rene de Roland L-A, Watson RT (2003) Falconiformes and strigiformes : ecology and status of raptors. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1080–1085 Wilme´ L (1996) Composition and characteristics of bird communities in Madagascar. In: Lourenc¸o WR (ed) Bioge´ographie de Madagascar. ORSTOM Editions, Collection Colloques et Se´minaires, Paris, pp 349–362 Young HG (2003) Freshwater birds. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1071– 1077

123

2740

Biodivers Conserv (2009) 18:2719–2742

Young HG, Lewis RE, Razafindrajao F (2003) Anseriformes : Anatidae, wildfowl. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1077–1080

References Agrawal A, Gibson CC (1999) Enchantment and disenchantment: the role of community in natural resource conservation. World Dev 27:629–649. doi:10.1016/S0305-750X(98)00161-2 Altieri MA (2002) Agroecology: the science of natural resource management for poor farmers in marginal environments. Agric Ecosyst Environ 93:1–24. doi:10.1016/S0167-8809(02)00085-3 Bacles CFE, Lowe AJ, Ennos RA (2006) Effective seed dispersal across a fragmented landscape. Science 311:628. doi:10.1126/science.1121543 Bellefontaine R, Petit S, Pain-Orcet M, Deleporte P, Bertault J-G (2001) Les arbres hors foreˆt. Vers une meilleure prise en compte. CIRAD/FAO, Cahier FAO no 35, p 214, Rome Berkes F (1999) Sacred ecology: traditional ecological knowledge and resource management. Taylor & Francis, Philadelphia Berkes F (2004) Rethinking community-based conservation. Conserv Biol 18:621–630. doi:10.1111/j.15231739.2004.00077.x Berkes F, Folke C, Gadgil M (1995) Traditional ecological knowledge, biodiversity, resilience and sustainability. In: Perrings CA (ed) Biodiversity conservation. Kluwer, Dordrecht, pp 281–299 Berkes F, Colding J, Folke C (2000) Rediscovery of traditional ecological knowledge as adaptive management. Ecol Appl 10:1251–1262. doi:10.1890/1051-0761(2000)010[1251:ROTEKA]2.0.CO;2 Blanc-Pamard C, Ramiarantsoa HR (2007) Normes environnementales, transferts de gestion et recompositions territoriales en pays betsileo (Madagascar). Nat Sci Soc 15:253–268. doi:10.1051/nss:2007055 Bleher B, Bo¨hning-Gaese K (2001) Consequences of frugivore diversity for seed dispersal, seedling establishment and the spatial pattern of seedlings and trees. Oecologia 129:385–394 Bodin O, Tengoe M, Norman A, Lundberg J, Elmqvist T (2006) The value of small size: loss of forest patches and ecological thresholds in southern Madagascar. Ecol Appl 16:440–451. doi:10.1890/10510761(2006)016[0440:TVOSSL]2.0.CO;2 Boulinier T (1998) Estimating species richness: the importance of heterogeneity in species detectability. Ecology 79:1018–1028 Burnham KP, Overton WS (1979) Robust estimation of population size when capture probabilities vary among animals. Ecology 60:927–936. doi:10.2307/1936861 Carrie`re SM (2002) ‘Orphan trees of the forest’: why do Ntumu farmers of southern Cameroon protect trees in their swidden fields? J Ethnobiol 22:133–162 Carrie`re SM, Letourmy P, McKey D (2002) Effects of remnants trees in fallows on diversity and structure of forest regrowth in a slash-and-burn agricultural system in southern Cameroon. J Trop Ecol 18:375–396 Carrie`re-Buchsenschutz SM (2006) L’urgence d’une confirmation par la science du roˆle e´cologique du corridor forestier de Fianarantsoa. Etud Rurales 178:181–196 Chao A (1987) Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43:783–791. doi:10.2307/2531532 Colding J, Folke C (1997) The relations among threatened species, their protection, and taboos. Ecol Soc 1:6 Colwell RK (2006) EstimateS: statistical estimation of species richness and shared species from samples. Version 8. Persistent URL \purl.oclc.org/estimates[ Da Silva JM, Uhl C, Murray G (1996) Plant succession, landscape management, and the ecology of frugivorous birds in abandoned Amazonian pastures. Conserv Biol 10:491–503. doi:10.1046/j.15231739.1996.10020491.x Daily GC, Ehrlich PR, Sanchez-Afofeifa GA (2001) Countryside biogeography: use of human-dominated habitats by the avifauna of southern Costa Rica. Ecol Appl 11:1–13. doi:10.1890/1051-0761(2001) 011[0001:CBUOHD]2.0.CO;2 Dunn RR (2000) Isolated trees as foci of diversity in active and fallow fields. Biol Conserv 95:317–321. doi: 10.1016/S0006-3207(00)00025-2 Fischer J, Lindenmayer DB (2002) The conservation value of paddock trees for birds in a variegated landscape in southern New South Wales. 1. Species composition and site occupancy patterns. Biodivers Conserv 11:807–832. doi: 10.1023/A:1015371511169 Fleming TH, Breitwisch R, Whitesides GH (1987) Patterns of tropical vertebrate frugivore diversity. Annu Rev Ecol Syst 18:91–109. doi:10.1146/annurev.es.18.110187.000515

123

Biodivers Conserv (2009) 18:2719–2742

2741

Gadgil M, Berkes F (1991) Traditional resource management systems. Resour Manage Opt 8:127–141 Gadgil M, Berkes F, Folke C (1993) Indigenous knowledge for biodiversity conservation. Ambio 22:151– 156 Garnett ST, Sayer J, du Toit JT (2007) Improving the effectiveness of interventions to balance conservation and development: a conceptual framework. Ecol Soc 12:2 Goodman SM, Benstead JP (eds) (2003) The natural history of Madagascar. The University of Chicago Press, Chicago and London, p 1709 Goodman SM, Ganzhorn JU (1997) Rarity of figs (Ficus) on Madagascar and its relationship to a depauperate frugivore community. Rev Ecol (Terre Vie) 52:321–329 Goodman SM, Raherilalao MJ (2003) Effects of forest fragmentation on bird communities. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1064–1066 Gotelli N, Colwell RK (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecol Lett 4:379–391. doi:10.1046/j.1461-0248.2001.00230.x Gray MA, Baldauf SL, Mayhew PJ, Hill JK (2007) The response of avian feeding guilds to tropical forest disturbance. Conserv Biol 21:133–141. doi:10.1111/j.1523-1739.2006.00557.x Guevara S, Laborde J (1993) Monitoring seed dispersal at isolated standing trees in tropical pasture: consequences for local species availability. Vegetatio 107–108:319–338 Guevara S, Laborde J, Sa´nchez-Rios G (2004) Rain forest regeneration beneath the canopy of fig trees isolated in pastures of Los Tuxtlas, Mexico. Biotropica 36:99–108 Harvey CA, Haber WA (1999) Remnant trees and the conservation of biodiversity in Costa Rican pastures. Agrofor Syst 44:37–68. doi:10.1023/A:1006122211692 Hughes JB, Daily GC, Ehrlich PR (2002) Conservation of tropical forest birds in countryside habitats. Ecol Lett 5:121–129. doi:10.1046/j.1461-0248.2002.00294.x Jones JPG, Andriamarovololona MM, Hockley N (2008) The importance of taboos and social norms to conservation in Madagascar. Conserv Biol 22:976–986. doi:10.1111/j.1523-1739.2008.00970.x Jordano P (2007) Differential contribution of frugivores to complex seed dispersal patterns. P Natl Acad Sci 104:3278–3282 Koechlin J, Guillaumet JL, Morat P (1974) Flore et ve´ge´tation de Madagascar. Editions J. Cramer, p 667, Vaduz Laborde J, Guevara S, Sa´nchez-Rı´os G (2008) Tree and shrub seed dispersal in pastures: the importance of rainforest trees outside forest fragments. Ecoscience 15:6–16. doi:10.2980/1195-6860(2008)15 [6:TASSDI]2.0.CO;2 Langrand O (1995) Guide des oiseaux de Madagascar. Delachaux et Niestle´, Paris Laumonier Y, Bourgeois R, Pfund J-L (2008) Accounting for the ecological dimension in participatory research and development: lessons learned from Indonesia and Madagascar. Ecol Soc 13:15 LeCœur D, Baudry J, Burel F, Thenail C (2002) Why and how we should study field boundary biodiversity in an agrarian landscape context. Agric Ecosyst Environ 89:23–40. doi:10.1016/S0167-8809(01) 00316-4 Leigh EJ, Hladik A, Hladik CM, Jolly A (2007) The biogeography of large islands, or how does the size of the ecological theater affect the evolutionary play? Rev Ecol (Terre Vie) 62:105–168 Luck G, Daily G (2003) Tropical countryside bird assemblages: richness, composition, and foraging differ by landscape context. Ecol Appl 13:235–247. doi:10.1890/1051-0761(2003)013[0235:TCBARC] 2.0.CO;2 Lumsden LF, Bennett AF (2005) Scattered trees in rural landscapes: foraging habitat for insectivorous bats in south-eastern Australia. Biol Conserv 122:205–222. doi:10.1016/j.biocon.2004.07.006 MacNally R, Bennett AF, Horrocksa G (2000) Forecasting the impacts of habitat fragmentation. Evaluation of species-specific predictions of the impact of habitat fragmentation on birds in the box–ironbark forests of central Victoria, Australia. Biol Conserv 95:7–29. doi: 10.1016/S0006-3207(00)00017-3 Manning AD, Fischer J, Lindenmayer DB (2006) Scattered trees are keystone structures—implications for conservation. Biol Conserv 132:311–321. doi:10.1016/j.biocon.2006.04.023 Margules CR, Pressey RL (2000) Systematic conservation planning. Nature 405:243–253. doi:10.1038/ 35012251 McCullagh P, Nelder JA (1989) Generalized linear models, 2nd edn. Chapman & Hall, London Millenium Ecosystem Assessment (2003) Ecosystems and human well-being: a framework for assessment. Island Press, Washington Covelo London Norris K (2008) Agriculture and biodiversity conservation: opportunity knocks. Conserv Lett 1:2–11. doi: 10.1111/j.1755-263X.2008.00007.x Papaik MJ, Sturtevant B, Messier C (2008) Crossing scales and disciplines to achieve forest sustainability. Ecol Soc 13:30

123

2742

Biodivers Conserv (2009) 18:2719–2742

Peters J (1998) Transforming the integrated conservation and development project (ICDP) approach: observations from the Ranomafana National Park project, Madagascar. J Agric Environ Ethics 11:17– 47. doi:10.1023/A:1007796628731 R Development Core Team (2007) R: A language and environment for statistical computing v2.8.0. R Foundation for Statistical Computing, Vienna, Austria Raherilalao MJ (2001) Effets de la fragmentation de la foreˆt sur les oiseaux autour du Parc National de Ranomafana (Madagascar). Rev Ecol (Terre Vie) 56:389–406 Randriamalala JR, Serpantie´ G, Carrie`re SM (2007) Influence des pratiques culturales et du milieu sur la diversite´ des jache`res d’origine forestie`re (Hautes-Terres, Madagascar). Rev Ecol (Terre Vie) 62:169– 189 Sankhala K (1993) Prospering from the desert. In: Kemf E (ed) Indigenous peoples and protected areas. Earthscan, London, pp 18–22 Scherr SJ, McNeely JA (2008) Biodiversity conservation and agricultural sustainability: towards a new paradigm of ‘ecoagriculture’ landscapes. Philos T R Soc B 363:477–494. doi:10.1098/rstb.2007.2165 Shanahan M, So S, Compton SG, Corlett R (2001) Fig-eating by vertebrate frugivores: a global review. Biol Rev Camb Philos Soc 76:529–572 Sheil D, Puri RK, Basuki I Heist Mv, Wan M, Liswanti N, Rukmiyati, Sardjono MA, Samsoedin I, Sidiyasa K, Chrisandini, Permana E, Angi EM, Gatzweiler F, Johnson B, Wijaya A (2002) Exploring biological diversity, environment and local people’s perspectives in forest landscapes: methods for a multidisciplinary landscape assessment. Center for International Forestry Research (CIFOR), Jakarta, Indonesia Steffan-Dewenter I, Kessler M, Barkmann J, Bos MM, Buchori D, Erasmi S, Faust H, Gerold G, Glenk K, Gradstein SR, Guhardja E, Harteveld M, Herteld Dietrich, Ho¨hn P, Kappas M, Ko¨hler S, Leuschner C, Maertens M, Marggraf R, Migge-Kleian S, Mogea J, Pitopan R, Schaefer M, Schwarze S, Sporn SG, Steingrebe A, Tjitrosoedirdjo SS, Tjitrosoemito S, Twele A, Weber R, Woltmann L, Zeller M, Tscharntke T (2007) Tradeoffs between income, biodiversity, and ecosystem functioning during tropical rainforest conversion and agroforestry intensification. P Natl Acad Sci 104:4973–4978. doi: 10.1073/pnas.0608409104 Styger E, Rakotoarimanana JEM, Rabevohitra R, Fernandes ECM (1999) Indigenous fruit trees of Madagascar: potential components of agroforestry systems to improve human nutrition and restore biological diversity. Agrofor Syst 46:289–310. doi:10.1023/A:1006295530509 Terborgh J, Robinson S, Parker TI, Munn C, Pierpont N (1990) Structure and organization of an Amazonian forest bird community. Ecol Monogr 60:213–238. doi:10.2307/1943045 Terborgh J, Pitman N, Silman M, Schichter H, Nu´n˜ez P (2002) Maintenance of tree diversity in tropical forests. In: Levey D, Silva W, Galetti M (eds) Seed dispersal and frugivory. CABI, Wallingford, Oxon, pp 1–17 Thorstrom R, Rene de Roland L-A, Watson RT (2003) Falconiformes and strigiformes: ecology and status of raptors. In: Goodman SM, Benstead JP (eds) The natural history of Madagascar. The University of Chicago Press, Chicago and London, pp 1080–1085 Van Dyck H, Baguette M (2005) Dispersal behaviour in fragmented landscapes: routine or special movements? Basic Appl Ecol 6:535–545. doi:10.1016/j.baae.2005.03.005 Warren DM, Pinkston J (1998) Indigenous African resource management of a tropical rainforest ecosystem: a case study of the Yoruba of Ara, Nigeria. In: Berkes F, Folke C (eds) Linking social and ecological systems: management practices and social mechanisms for building resilience. Cambridge University Press, Cambridge, pp 158–189 Watson JEM, Whittaker RJ, Dawson TP (2004) Habitat structure and proximity to forest edge affect the abundance and distribution of forest-dependent birds in tropical coastal forests of southeastern Madagascar. Biol Conserv 120:311–327. doi:10.1016/j.biocon.2004.03.004 Wegner JF, Merriam G (1979) Movements of birds and small mammals between a wood and adjoining farmland habitats. J Appl Ecol 16:349–357 Wunder S (2001) Poverty alleviation and tropical forest: what scope for synergies? World Dev 29:1817– 1833. doi:10.1016/S0305-750X(01)00070-5 Yarranton GA, Morrison RG (1974) Spatial dynamics of a primary succession: nucleation. J Ecol 62:417– 428. doi:10.2307/2258988

123