The Wallacean biodiversity ‘hotspot’ (Myers et al. 2000) supports diverse bird communities with a high proportion of endemic species, although these remain poorly studied (Cannon et al. 2007, Eaton et al. 2021). Though some early studies provided significant insight into the avifauna of particular islands, such as the work of Verheijen (1964) on Flores and Noske (2003) on Timor, significant gaps in knowledge persisted for many years, as highlighted in the foundational compilations of regional ornithological knowledge (White & Bruce 1986, Coates & Bishop 1997). In recent years knowledge regarding the taxonomy, distribution and conservation status of the region’s species has improved via significant new field and museum-based research (e.g. Indrawan et al. 2006, Martin et al. 2012, Reeve et al. 2015, Bashari & Arndt 2016, Verbelen et al. 2017, Monkhouse et al. 2018, Ng et al. 2018, O’Connell 2019, Irham et al. 2020, Rheindt et al. 2020) and efforts to organise citizen science data such as eBird ( https://ebird.org/home). These advances have been compiled in a new regional field guide (Eaton et al. 2016, 2021) and informed the growing Birds of the world ( https://birdsoftheworld.org/bow/home) online database of the world’s avifauna.

However, basic information concerning the natural history of most Wallacean species is still lacking, hampering an understanding of how species within the region might respond to environmental challenges (Lees et al. 2020). Outside the detailed work of Verheijen (1964), who rigorously collected >1,000 records of breeding birds on Flores, the relatively sparse data on avian biology from Wallacea has largely come from opportunistic observations by field workers. Accumulating and quantifying data this way is a slow process (Fierro-Calderón et al. 2021), which has been hampered by the fact that Wallacea is still relatively understudied, with numerous endemic bird species going unrecognised until recently (O'Connell et al. 2019c, Rheindt et al. 2020), highlighting the relatively low baseline scientific knowledge of the region's avifauna. This shortfall in natural history data was quantified by Noske (2017), who demonstrated that 57% of all Wallacean endemics possess no published data on their breeding biology and 72% no data regarding nest characteristics. As breeding by birds in Wallacea is strongly seasonal, being driven by changes in precipitation between the wet and dry seasons (Verheijen 1964, Noske 2003), filling knowledge gaps concerning breeding patterns in Wallacean birds provides vital information to understand how likely future changes in regional climate may affect species' ecologies. As movements by some Wallacean species are complex, with seasonal short-range breeding migrations noted for species such as Island Monarch Monarcha cinerascens and Elegant Pitta Pitta elegans, collecting data on breeding behaviour informs our understanding of key breeding sites for species with complex life histories (Eaton et al. 2021).

In this study, we respond to Noske's (2017) call for more published breeding data on Wallacean birds by summarising all information related to breeding biology obtained throughout a series of long-running bird surveys of south-east Sulawesi and its offshore islands, Indonesia. These surveys formed part of near-annual field seasons between 1999 and 2018, in a collaborative partnership between Operation Wallacea ( www.opwall.com), Trinity College Dublin (TCD), Universitas Halu Oleo (UHO), Bogor Agricultural University (IPB) and University of Wollongong (UOW).

Study area and Methods

Data on the breeding biology of the birds of south-east Sulawesi were collected during two parallel ornithological research projects focusing on different aspects of the regional avifauna. Systematic research was conducted on Buton most years during the study period (2004–14 and 2018) by Operation Wallacea, IPB, UHO and UOW ornithologists, surveying set transect routes in forest and forest edge habitats in central and north Buton to monitor bird populations, along with limited mist-netting (see Martin et al. 2012 for detailed methods, survey periods and habitat descriptions). In addition, researchers from TCD and UHO undertook a long-term study of avian speciation and biogeography in south-east Sulawesi, collecting data from most islands in the region (including Buton) over a number of field seasons during 1999–2017 (Table 1); for detailed methods, survey periods and habitat descriptions see O'Connell et al. (2017, 2019d, 2020) and Monkhouse et al. (2018). Briefly, these surveys involved mist-netting and transects in edge habitats (scrub, mangrove, farmland, forest edge), moving regularly and covering a large variety of sites for short periods.

The goals of these surveys were not explicitly to obtain information on the breeding biology of Wallacean birds. Rather they were to address questions relating to ecological (Martin & Blackburn 2014), biogeographical (O'Connell et al. 2019a, Ó Marcaigh et al. 2021, Ó Marcaigh et al. 2022) and evolutionary (Kelly et al. 2014, O'Connell et al. 2019b,c) processes in the region. However, during the course of these surveys, numerous opportunistic observations were made of breeding biology, in the form of nests and juvenile birds, handling individuals in breeding condition during mist-netting work, and behavioural observations. Together, these efforts encompassed work on mainland south-east Sulawesi, the five large islands off the coast of south-east Sulawesi (Buton, Muna, Wawonii, Kabaena and Menui), the four major Wakatobi Islands (Wangi-wangi, Kaledupa, Tomia and Binongko, uplifted Quaternary coral limestone islands always separated from Sulawesi; Milsom et al. 1999, Nugraha & Hall 2018), a number of small islets off the coasts of these islands, and isolated Runduma Island (Fig. 1, Table 1). Breeding bird records were collected on 16 of these islands (Table 1). With the exception of a visit to the Wakatobi Islands in April/May 2013, all field work was conducted during July–September (Wakatobi Islands) and June–August (all other islands), with a particular concentration of effort in July and August. Comparable quantification of field work effort is impossible for the collection of breeding records, as many important observations were made outside structured surveys, and survey personnel, timing and focus varied significantly between years and islands. However, broadly, Buton has received the most consistent and detailed survey effort, followed by the islands of Kabaena, mainland Sulawesi, Hoga, Kaledupa, Wangi-wangi and Tomia (Table 1) (Martin et al. 2012, 2015, 2017a, O'Connell et al. 2017, 2019d, 2020, Monkhouse et al. 2018). Visits to other islands were generally brief or limited in scope.

Figure 1.

South-east Sulawesi and its offshore islands. The inset shows the location of the region within Indonesia and the wider area.

In the Sulawesi region, precipitation follows a tropical monsoon climate, with a June–September dry season and November–April wet season (Whitten et al. 2002). Mean annual rainfall ranges between 1,500 and 2,000 mm, peaking in April–June. Mean annual temperatures range from 25°C to 27°C (Whitten et al. 2002). Birds have been noted to breed year-round in Wallacea (Noske et al. 2013), but generally display seasonal peaks related to rainfall. How precipitation patterns affect breeding in Sulawesi's bird species remains an open question. In the most detailed year-round breeding survey in Wallacea, Verheijen (1964) found that peak activity on Flores was at the start of the dry season, and Linsley et al. (1998) reported a similar result based on more fragmentary data from Sumba. However, Noske (2003) discovered a distinctly different pattern on Timor, with breeding largely ceasing in the dry season (June–September) and peaking twice in the wet season (March–April and November–December). Our survey (concentrated on July–August) is at the start of the dry season in south-east Sulawesi, so may also capture peak breeding activity if the pattern is similar to Flores (Verheijen 1964). On most islands surveyed, edge habitats were those most accessible for survey work; therefore much of our effort focused on scrub, forest edge, farmland and mangroves across all islands (Fig. 2B) (O'Connell et al. 2017, 2019d, 2020, Monkhouse et al. 2018). The exception to this was Buton, where large tracts of lowland rainforest with a canopy height of >25 m (Whitten et al. 2002) remain, and significant survey effort was focused on Lambusango Forest Reserve, a 650 km² expanse of tropical monsoon forest in central Buton, and Buton Utara Nature Reserve (920 km²) in northern Buton (Fig. 2A) (Martin et al. 2012). In addition, short surveys were undertaken of paddyfield areas on Buton, Wawonii and Sulawesi (Fig. 2C) (Martin et al. 2012, O'Connell et al. 2019d), the UHO urban forest in Kendari city (mainland Sulawesi), the small monsoonal dry forest on Menui (23.5 km²) (Fig. 2D) (Monkhouse et al. 2018) and the remaining lowland rainforest on Kabaena (fragmentary areas with a closed canopy at c.12 m) (Robinson-Dean et al. 2002).

TABLE 1

Summary of islands visited in south-east Sulawesi, and the years in which breeding records were collected. Islands denoted ‘(W)’ form part of the Wakatobi chain. The total number of records of breeding evidence, with number of species in parentheses, across the full time period on each island is presented in the final column. Field work was heavily weighted towards July and August. ‘NA’ indicates periods where no field work was undertaken on that island.

Evidence of breeding was provided by observations of: (1) active nests, (2) juveniles (with reference to Coates & Bishop 1997 and Eaton et al. 2016 for juvenile plumage features), (3) birds in breeding condition, i.e. females with clear brood patches or carrying eggs (gravid), and males with obvious cloacal protuberances, identified during mist-net surveys, with reference to Rogers et al. (1986) and Svensson (1992) for identifying these, and (4) characteristic breeding behaviour such as nestbuilding, display flights or copulation. Active nests were considered the primary evidence of breeding as this constitutes irrefutable proof linked to a known location, followed by juveniles (confirmed breeding, location and timing not certain as the distance juveniles travel, and duration juvenile plumage persists, vary significantly between species), birds in breeding condition (indicative of breeding but not confirmed) and finally behaviour (more weakly indicative of breeding). Breeding records were aggregated by species and island, and arranged in taxonomic order following Gill et al. (2021). In addition, for Wallacean endemics (following Myers et al. 2000) and threatened species (IUCN 2021), we provide further details of our observations and assess how these add to the existing data on these species' breeding biology. Where available, photos of notable records were uploaded to the Macaulay Library ( https://www.macaulaylibrary.org/) and photo codes are provided in the format ML #.

Figure 2.

A range of habitats covered by the survey effort in south-east Sulawesi showing: (A) lowland rainforest on Buton (Thomas E. Martin), (B) beach and coastal scrub on Hoga in the Wakatobi Islands (Thomas E. Martin), (C) rice paddies on Buton and secondary forest transitioning into (D) mature dry forest on Menui (Joseph Monkhouse)

Results

Overall, we collected 1,064 observations of potential breeding evidence for 66 species, of which 27 were Wallacean endemics (including the Critically Endangered Maleo and Near Threatened Red-backed Thrush Geokichla erythronota), and 39 species with a wider range (including the Endangered Milky Stork and Near Threatened Rufous-bellied Eagle Lophotriorchis kienerii) (Table 2,  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx)). Active nests were considered the strongest evidence of breeding, followed by juveniles, adults in breeding condition and finally adults exhibiting breeding behaviour. For the 66 species recorded as proven or potentially breeding, evidence of breeding came from active nests for 17 species (19 records in total, providing strongest evidence of breeding for 17 species), from juveniles for 48 species (313 records in total, providing strongest evidence of breeding for 38 species), from observations of birds in breeding condition for 35 species (705 records in total, providing strongest evidence of breeding for nine species) and observations of birds displaying breeding behaviour for nine species (27 records in total, providing primary evidence of breeding for two species). Therefore, breeding was confirmed for 55 species (active nests and juveniles) with evidence of potential breeding (breeding condition and behaviour) for an additional 11 species. The 66 species with breeding records represent 32 families, but 73% of all records are from three families of small passerines common in the region: white-eyes (33%), sunbirds (28%) and flowerpeckers (11.5%) (Table 2). The majority of breeding records, and the greatest diversity of breeding birds, were recorded on the larger islands that were visited more regularly (mainland Sulawesi, Buton and Kabaena), likely reflecting greater habitat diversity, although a significant number of records were also collected on the Wakatobi Islands of Kaledupa and Tomia where field work was undertaken in multiple years (Table 1). Buton had the greatest species diversity of breeding records (Table 1), particularly of active nests (Table 2), doubtless reflecting the greater effort (Table 1) at intensively surveyed forest sites.

TABLE 2

Summary of new breeding records of 66 bird species on the islands of south-east Sulawesi. Taxonomy follows Gill et al. (2021). Species in bold and denoted * are endemic to the Wallacean biodiversity hotspot as defined by Myers et al. (2000). After the common name, an (I) indicates species introduced to Wallacea; CR = Critically Endangered, EN = Endangered, NT = Near Threatened (IUCN 2021). Initials in the ‘Island’ column correspond to the following island names: Sul = south-east Sulawesi mainland, But = Buton, Siu = Siumpu, Mun = Muna, Waw = Wawonii, Kab = Kabaena, Sag = Sagori, Men = Menui, Wan = Wangi-wangi, Oro = Oroho, Hog = Hoga, Kal = Kaledupa, Tom = Tomia, Lin = Lintea Selatan, Bin = Binongko, Run = Runduma. Initials in the ‘Category’ column correspond to the following categories: BB = breeding behaviour, BC = breeding condition, J = Juvenile, N = active nest. Months are denoted: Apr = April, August = Aug and September = Sep. All individual records providing date, location details and additional notes are presented in  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx). Field work effort was heavily focused on July and August.

Continued

Continued

Continued

Continued

Continued

Continued

Individual accounts for Wallacean endemics (denoted *), Critically Endangered (CR) and Endangered species (EN) are presented below, covering 28 species in detail, with the remaining 38 species covered in Table 2 and  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx).

SULAWESI HAWK-EAGLE Nisaetus lanceolatus *

Few breeding data available (Clark & Kirwan 2020b). One nest recorded in August on Sulawesi (White & Bruce 1986), and season thought to be May–August, based on fledged young also seen in the latter month (Ferguson-Lees & Christie 2001). Two juveniles observed during our study; one in Labundobundo village (central Buton) on 22 July 2014 by TEM, and one in Buton Utara Reserve (north Buton) on 12 July 2018 by PGA (Table 2,  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx), ML 204689311).

Figure 3.

Juveniles of: (A) Speckled Boobook Ninox punctulata, Buton, August 2013 (Thomas E. Martin), (B) Spot-tailed Sparrowhawk Accipiter trinotatus, Buton, July 2009 (Thomas E. Martin), (C) Grey-sided Flowerpecker Dicaeum celebicum, Kabaena, August 2003 (David J. Kelly and Nicola M. Marples) and (D) Moluccan Starling Aplonis mysolensis, Menui, August 2017 (Joseph Monkhouse)

PYGMY HANGING PARROT Loriculus exilis *

Few breeding data available. Known to breed in February and August and to nest in holes in dead palms (White & Bruce 1986, Collar et al. 2020), with reports of breeding from Lore Lindu NP in October–November (D. D. Putra unpubl. data). Possible nesting behaviour in this species (and a possible active nest) was observed in Kakenauwe village, Buton, on 10 July 2018 by PGA. A pair was seen on this date repeatedly entering a cavity atop a derelict wooden-cored concrete utility pillar in the village (Fig. 4C). The cavity entrance was c.25 cm in circumference and broadly round in shape. This behaviour was subsequently noted several further times over the next three weeks, although by late July only the male was seen entering and exiting the cavity, with another bird (presumably the female) calling from inside it. During these observations, an all-male flock of five Pygmy Hanging Parrots was also seen regularly in a coconut tree near the pillar, which the (presumably nesting) pair appeared to tolerate (Table 2,  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx)). These observations demonstrate the species can probably breed in anthropogenically modified habitats, and artificial wood-core pillars may offer alternative nesting sites to the palms previously noted as being used by the species.

Figure 4.

Photos of observed/assumed nesting attempts by: (A) Ashy Woodpecker Mulleripicus fulvus, Buton, July 2018 (Panji G. Akbar), (B) Sulawesi Pitta Erythropitta celebensis, Buton, June 2016 (Arini Wijayanti), (C) Pygmy Hanging Parrot Loriculus exilis, Buton, July 2018 (Panji G. Akbar) and (D) Ivory-backed Woodswallow Artamus monachus, Buton, June 2018 (Panji G. Akbar)

Discussion

Our results provide an extensive summary of breeding data for bird species in south-east Sulawesi based on nearly 20 years of field work. The data we present improve understanding of the basic natural history of Wallacean species. This paper provides extensive observations of breeding by 66 bird species. These include rare observations of breeding behaviour and nest structure for several poorly studied endemics, including Pygmy Hanging Parrot, Sulawesi Pitta and Ivory-backed Woodswallow. In addition, we present important evidence of breeding activity by one Critically Endangered and one Endangered species. The Milky Stork nesting attempt recorded on Kabaena sheds light on potential breeding areas for the poorly understood Sulawesi population of this widespread but rare species. The record of a juvenile Maleo confirms that this species is both extant and breeding on Buton (as of 2014). Our observations of breeding condition in Wangi-wangi White-eye provide the first data regarding life history parameters in this single-island endemic, which could be classified as Endangered due to its restricted range (O'Connell et al. 2019c, 2021).

Our recording period was focused on the start of the dry season on Sulawesi, which is the peak breeding period on nearby Flores (Verheijen 1964). Breeding of birds in Wallacea is strongly seasonal and influenced by precipitation, but how this affects the timing of breeding varies on different islands (Verheijen 1964, Linsley et al. 1998, Noske 2003). As the impacts of climate change progress, the region is likely to experience more extreme seasonality, with wetter wet seasons and drier dry seasons (Chou et al. 2013). Therefore, filling gaps in our knowledge of avian breeding biology on Sulawesi is vital to provide baseline data on timing of breeding, informing how these species might respond to anthropogenic habitat and climate change in this part of Wallacea (Lees et al. 2020).

Our assessment of avian breeding biology in south-east Sulawesi has several clear limitations. First, our survey methods were not designed to maximise the recording of breeding behaviour, with the focus instead being on biodiversity (Buton) and speciation (all other islands covered by TCD & UHO). Therefore, nest searches and behavioural observations were not prioritised, and our data are very much opportunistic in nature. The majority of our records of breeding behaviour involve common small passerines that are easily trapped using mist-nets (Table 2,  Table S1 (BBRC_142_3_01OConnell_Supplemenatry_information_Table_S1_all_breeding_records.xlsx)), but this contribution should be considered a descriptive account of breeding behaviour, rather than a strict quantification of such activity at each field site. Second, field work was largely concentrated in late June–August and consequently provides only a snapshot of potential seasonality. These limitations make it difficult to infer phenological trends for the species in our dataset. Additionally, tropical birds tend to have longer breeding seasons than those in temperate regions, and timing of breeding can vary significantly between trophic guilds (Stutchbury & Morton 2001). Our study region is also geologically and ecologically complex (Whitten et al. 2002, Nugraha & Hall 2018), and this may cause the breeding biology of bird species to differ even on neighbouring islands. For example, the islands of Kabaena and Wawonii have distinctly different underlying geology (ultramafic) and associated plant communities (Galey et al. 2017, Trethowan et al. 2020, Ó Marcaigh et al. 2021). The entire study area is also strongly influenced by El Niño–Southern Oscillation cycles, which exert influence on breeding probability and phenology (Duursma et al. 2018, Smart et al. 2021). Thus, breeding patterns on different islands, and in extreme El Niño/La Niña years, may produce distinctive patterns that our data are unable to reveal.

It is also interesting to note which species are missing from our dataset, particularly in the mature lowland rainforest of Buton Island. Many species considered relatively common by Martin et al. (2012), e.g. Green Imperial Pigeon Ducula aenea, Knobbed Hornbill Rhyticeros cassidix and Sulawesi Hornbill Rhabdotorrhinus exarhatus, lack a single observation of breeding behaviour despite nearly 20 years of field work. It is acknowledged that surveying the demographics of tropical birds is often difficult (Robinson et al. 2018), but the absence of many common species from our dataset highlights that obtaining breeding data for tropical birds is also problematic unless targeted methods (see Fierro-Calderón et al. 2021) are employed. Even these may be inhibited in ‘canopy-heavy’ ecosystems such as on Buton, where many species rarely come close to ground level (Martin et al. 2017b). Indeed, the many conspicuous gaps in our dataset indicate that difficulties in data acquisition may be an important contributing factor to the dearth of information for Wallacean birds (Noske 2017). However, the lack of experienced observers in the region certainly is also a key limitation, although initiatives such as Atlas Burung Indonesia are beginning to remedy this shortfall (Taufiqurrahman et al. 2016).

Overall, however, this paper fills a number of knowledge gaps, particularly with respect to the breeding biology of 27 Wallacean endemics, many of which have few or no previous nesting data. This study also highlights the importance of collating opportunistic natural history data in poorly studied areas, even when assembling such data is not the primary focus. Like Noske (2017), we encourage other researchers in Wallacea (and other under-studied regions) to publish summaries of their breeding data, to further knowledge of avian breeding biology in this unique biodiversity hotspot.