Checklist of the bees (Hymenoptera, Apoidea) of New Caledonia

Abstract Background In a world where insects and notably bees are declining, assessing their distribution over time and space is crucial to evaluate species status and highlight conservation priorities. However, this can be a daunting task, especially in areas such as tropical oceanic islands where exhaustive samplings over time have been lacking. This is the case in New Caledonia, an archipelago located in the southwest Pacific. Historical records of bee species are piecemeal and, although contemporary samplings have significantly advanced our knowledge of the bee fauna of New Caledonia, the status of several species remains to be elucidated. New information Here, we provide an updated checklist of the 51 bee species recorded for New Caledonia using previous publications and personal samplings. We documented their distribution, origin (i.e. endemic, native or alien) and the year and location of their occurrences. Based on the year of their first capture and the year of their last capture, we determined an occurrence status for each species. Thus, 10 years after the last checklist of the New Caledonian bee fauna, the literature review and recent samplings allowed us to add six new species to the list. Half of them are recently introduced species including one firstly mentioned in this paper (i.e. Hylaeusalbonitens). We consider here that 30 species are effectively present on the territory and the presence of 21 species could not be determined due to a lack of data, which highlights the need to increase sampling efforts across New Caledonia. Given the difficulty of exhaustively sampling the entire archipelago, we would recommend taking, as a starting point, altitude environments and areas where data-deficient species were captured. In a broader perspective, biomolecular analyses are crucial to confirm species identifications. This is also needed to make comparisons between archipelagoes and thus clarify the distribution and status of species at the scale of the southwest Pacific.


Introduction
Insects and notably bees are declining worldwide (Wagner 2020, Zattara and Aizen 2021). For 50 years, many studies have reported a decrease in their abundance and species richness, even in protected areas (Hallmann et al. 2017, Seibold et al. 2019. The magnitude and rate of this biodiversity loss have urged taxonomists and naturalists around the globe to exhaustively assess species diversity and their distribution over time and space (Orr et al. 2021). This baseline information is crucial to evaluate species status and provide informative tools for conservation, such as IUCN Red Lists (Rodrigues et al. 2006, Potts et al. 2010, Cardoso et al. 2011, Fox et al. 2019. While in developed countries, where bee species have been studied for a long time, the status of bee populations is already difficult to assess -as a single example, the European Red List of Bees did not evaluate the status of more than a half of the 2000 species present on the continent due to a lack of data (Nieto et al. 2014) -this issue is all the more relevant in other parts of the world where bee species have been seldomly sampled, notably in oceanic tropical and subtropical islands, such as New Caledonia (Orr et al. 2021).
New Caledonia is located in the southwest Pacific. This French overseas archipelago is part of the Australasian biogeographic realm. Historically, the archipelago went through a total submersion event between 75 and 60 Mya and likely at least a partial submersion event between 34 and 25 Mya (Maurizot and Campbell 2020) and has been isolated from any continent since, allowing an important species radiation (Grandcolas et al. 2008, Nattier et al. 2017). The archipelago covers 18 519 km² with a mosaic of edaphic and topographic conditions supporting a high diversity of habitats. Notably, a third of the main island is covered by ultramafic soils, originating from the overlap of peridotites on the continental crust during an obduction event in the late Eocene (Pelletier 2006, Pillon et al. 2020. These soils are characterised by a low nutrient availability and high concentrations of metals and metalloids, strongly constraining plant growth. As such, plants growing on ultramafic substrates present a very high rate of endemism. Thus, New Caledonia is considered as one of the main biodiversity hotspots at global scale for vascular plants with an endemic rate of 74%, up to 96.7% in ultramafic habitats (Morat et al. 2012, Isnard et al. 2016, Munzinger et al. 2022. However, regarding animals, while some groups also display a remarkable diversity and a high rate of endemism (e.g. Squamata; Bernstein et al. (2021) , Slavenko et al. (2023)), others, such as bees, appear remarkably poor, illustrating a convincing example of island disharmony (Vargas 2012).
From the end of the 19 century to the middle of the 20 century, very few mentions and descriptions of bees were published for New Caledonia (Vachal 1897, Friese 1905, Vachal 1907, Vachal 1908, Cockerell 1911, von Schulthess 1915, Cockerell 1939, Cheesman 1953, Donovan 1983. These records are fragmentary and mostly based on the examination of old collections. The first review of the bee species of New Caledonia is attributed to Michener in 1965, who listed nine species (previously mentioned in the literature) and one new species retrieved from an unexamined collection (Michener 1965). Following this, Donovan (1983) made additional captures between 1979 and 1981 and updated the number of New Caledonian bee species to 28, unfortunately naming only two of them. This number did not change until 2003, when Pauly & Muzinger updated this list to 22 species, based on specimens stored in collections and additional captures realised between 1998 and 1999 (Pauly and Munzinger 2003). An important step in the knowledge of the New Caledonian bee fauna was achieved by  , who studied plant-bee interactions using previous collections and additional captures realised between 1979 and 2008. This substantial work resulted in a two-fold increase in the number of bee species recorded for New Caledonia. However, amongst the 43 bee species listed, 24 were undetermined species at this stage. Subsequently, the works of Pauly et al. (Pauly et al. 2013a, Pauly et al. 2013b, Pauly et al. 2015 on the description and revision of New Caledonian Austronomia, Lasioglossum and Homalictus genera allowed a species name to be given to 15 of them. Since then, no review has provided an updated assessment of the bees of New Caledonia. For the last ten years, we have realised three bee sampling campaigns and opportunistic captures in several localities through New Caledonia (e.g. Zakardjian et al. (2020), Zakardjian et al. (2023a) and we felt the lack of an updated checklist clarifying the advances of the last 50 years. For example, many bee species recorded for New Caledonia during the 20 century have not been seen for several decades, if not since their first detection. Conversely, new alien bee species did enter the territory (see below). This is why, in this paper, we provide an updated checklist of the bee fauna of New Caledonia. We provide their status, distribution and main synonymies. We also provide every occurrence (date and location) found in the literature and in our samplings, together with a comprehensive list of the plant species they visit.

Literature search
The checklist provided here results from: i) a review of existing literature and ii) additional data from recent field samplings.
Reviewed publications were retrieved from three methods. First, we retrieved 15 publications from GoogleScholar using the key words "New Caledonia" and "bees". Then, eight were retrieved from the personal library of H. Jourdan. Finally, we retrieved 14 additional publications of interest (i.e. mentioning New Caledonian bee species) cited in the former publications. In total, we scanned 37 publications. Amongst them, 23 were of interest and used in this work (Suppl. material 1). In those publications, we looked for mentions of bee species in New Caledonia. For each bee species, we gathered information concerning their taxonomy (i.e. genus, subgenus, species, subspecies, synonymies and past combinations), occurrences in New Caledonia (i.e. localities and dates of capture), plant species visited, distribution outside New Caledonia if not endemic and original status (i.e. endemic, native, alien). In order to increase the reliability of this information, we verified the valid name and synonymies of each species according to the INPN's (French National Inventory of the Natural Patrimony, https://inpn.mnhn.fr) taxonomic repository TaxRef (Gargominy et al. 2022). For example, one species was named Lasioglossum instabilis (Cockerell 1914) in the publications reviewed ( Pauly et al. 2013b). However, according to TaxRef, this name is a synonym and its valid name is Lasioglossum instabile (Cockerell, 1914). Therefore, this species appears as L. instabile in this checklist. We also systematically verified species distribution outside New Caledonia using GBIF (Global Biodiversity Information Facility, www.gbif.org) and completed the information based on our own expertise (Hervé Jourdan, pers. com.). We noted every country in which each species has been observed.

Additional samplings
In addition to the occurrences retrieved from the existing literature, we implemented new occurrences and plant species visited from our personal published (Zakardjian et al. 2020) and unpublished datasets and opportunistic samplings -all realised between 2017 and 2023.
The first dataset comes from a sampling conducted in 2017, from March to May, in Ouvéa, Tiga, Koumac and Belep Islands areas. In each area, bees were captured along 24 transects of 50 m placed to represent their diversity of environments (e.g. forest, coastal vegetation, fallow land, shrubland and subsistence agriculture). Each transect was walked for 25 min during which bees interacting with flowers were captured using nets. In addition, in each site, 16 pan traps (4 blue, 4 red, 4 white and 4 yellow) were distributed during 48 h.
The second dataset comes from a sampling conducted in 2019, from February to April, in Nouméa and the Tontouta Valley. Nouméa is the main city of the archipelago, with 99,926 inhabitants for an area of 45.7 km². Within the city, three sites were sampled, separated by at least 3 km. These three sites were: (i) the park of the French National Research Institute for Sustainable Development (Institut de Recherche pour le Développement); (ii) the Parc Zoologique et Forestier Michel Corbasson and (iii) the Tjibaou Cultural Centre. The Tontouta Valley is located 30 km north of Nouméa. Within the Tontouta Valley, three sites were sampled, separated by at least 3 km. In each site, the plant species presenting the most abundant floral resources were sampled. Depending on the number of available flowering species, up to four plant species were observed during each sampling session. For each plant species, bees contacting flowers in a 1 m² quadrat were captured using nets during 10 minutes. In total, each Tontouta Valley site was sampled 10 times and each Nouméa site 11 times.
The third dataset comes from a sampling realised in 2022, from March to June, in 14 sites along an urbanisation gradient running from Nouméa to the Col de Mouirange (Suppl. material 2). In each site, the plant species presenting the most abundant floral resources were sampled. Depending on the number of available flowering species, up to five plant species were observed during each sampling session. For each plant species, bees contacting flowers in a 1 m² quadrat were captured using nets during 10 minutes. Each site was sampled four times.
Records retrieved those additional samplings have bee published through GBIF (Zakardjian et al. 2023b) and are provided within this checklist together with historical data.

Occurrence status
For each bee species, we determined an occurrence status in New Caledonia, based on the year of first capture and the year of last capture. Species were determined as "present" if individuals were captured in more than one year including at least one during the last 50 years (i.e. later than 1973) or "data deficient" if individuals were captured in only one year.

Origin status
Information on the distribution of species was used to determine or correct species origin status. Indeed, for most of the species listed here, their status does not appear in the reviewed publications or may not be valid anymore. For example, Megachile albomarginata was mentioned as endemic to New Caledonia in Pauly and Munzinger (2003). However, a subsequent work showed that this species is also present in Fiji (Davies et al. 2013). Thus, M. albomarginata appears as native to New Caledonia in the present checklist. Species present elsewhere than in New Caledonia without any mention of an alien origin in the archipelago were categorised as native. Then, native species only present in New Caledonia were categorised as endemic here. Finally, for alien species, we used the definition given by Russell and Blackburn (2017), according to which a species is considered as alien if its presence is due to a geographical displacement through anthropogenic activities (e.g. voluntary importation, accidental transport via commercial trade). Thus, species that naturally colonised New Caledonia were not considered as alien in this checklist. First, species mentioned as alien or likely alien in literature were systematically categorised in the same way in the checklist. Then, species that were recorded for the first time during our additional samplings were considered as alien.
in Fiji and French Polynesia where it is an alien species (Groom et al. 2017, Groutsch et al. 2018 Notes: First detected in Nouméa in 2016, the species is now established in the city and has started to spread north along the west coast (Hervé Jourdan, pers. com.).

Apis mellifera subsp. mellifera L., 1758
Feeds on: A large range of endemic, native and alien plants (including invasive ones).

Native status: Alien
Distribution: Cosmopolitan subspecies. Ubiquitous in New Caledonia, but absent from Belep Islands and Tiga.
Notes: Before the arrival of Europeans, there were no social bees or honey bees in New Caledonia. Apis mellifera mellifera was first introduced in 1848 from France by priests on Lifu Island for the production of wax candles. It subsequently spread to the mainland and the Isle of Pines in 1853, as well as to Maré during the same period (Lamaignere 2001, Hervé Jourdan, pers. com.). To our knowledge, Apis mellifera is still absent from Belep Islands and Tiga and abundant elsewhere in the archipelago.

Apis mellifera subsp. ligustica Spinola, 1806
Feeds on: A large range of endemic, native and alien plants (including invasive ones).

Native status: Alien
Distribution: Cosmopolitan subspecies. Ubiquitous in New Caleodnia, but absent from Belep islands and Tiga.
Notes: Apis mellifera ligustica was introduced periodically by the end of the 19 and the beginning of 20 centuries on the mainland as mentioned by von Schulthess (1915), but most of colonies were brought from Australia and New Zealand between 1985 and 1988 on the mainland, following American foulbrood outbreaks (Thevenon et al. 1989, Lamaignere 2001. Notes: Likely a new alien species, potentially introduced via the Vavouto industrial harbour facility. We nevertheless cannot exclude that one of the three below-mentioned Hylaeus morphospecies ) may also fit with H. albonitens.
Occurrence status: data deficient.
Notes: This species is named Hylaeus sp. indet. One in . Notes: This species is named Hylaeus sp. indet. Three in .
Occurrence status: data deficient.

Native status: Endemic
Distribution: Historical data in New Caledonia: Nepoui Valley, Jul 1940, one male Munzinger 2003, Donovan et al. 2013). Dec 1979-Sep 2008, one male and one female, no location .
Notes: This species has been recorded in forested or maquis areas including in altitudes (up to 900 m). The recorded host plants and known localities are restricted to ultramafic substrates.
Occurrence status: data deficient.
Native status: Likely alien. Donovan (1983) hypothesised that this species could be alien because individuals were captured in a few localities in the vincinity of Noumea, suggesting a recent establishment.

Distribution: Vanuatu
Notes: This bee appears as the most common native bee in New Caledonia. It has been observed in a wide range of habitats (including urban, non-ultramafic, ultramafic and altitude environments) in the whole archipelago.
Potential confusion with Homalictus urbanus from Australia; Pauly and Villemant (2009) highlighted a slight difference between the specimens of Vanuatu and New Caledonia compared to specimens from Australia. Based on preliminary biomolecular analysis (unpublished data), we confirm that H. aponi is a different species from H. urbanus and we confirm its status as a native species from Vanuatu and New Caledonia.

Notes: This species was observed in altitudinal (> 500 m) forested areas on ultramafic substrates.
Occurrence status: data deficient.
Notes: This species has been observed in ultramafic substrates in both maquis and forested areas.
Notes: This species has been observed in forested areas on both ultramafic and volcano sedimentary substrates, in high altitudes (450 to 900 m).  Nomenclature:

Native status: Endemic
Notes: This species is widespread in the archipelago and has a broad range of host plants, including endemic, native and alien species. It can be found from low to high altitudes on both ultramafic and volcano sedimentary substrates.

Notes:
The status of H. risbesci crotalariae as a subspecies remains to be confirmed through molecular analysis.
Notes: This species is named Homalictus sp. indet. Seven in .
Occurrence status: data deficient.

Lasioglossum (Austrevylaeus) sp.*
Notes: This species is mentioned in Donovan (1983) based on a personal communication. Pauly and Munzinger (2003) stated that they did not examine the species which, therefore, needs confirmation.
Occurrence status: data deficient.
Notes: This species has been observed in high altitude (up to 1350 m) maquis and forested areas on ultramafic substrates and high altitude (up to 850 m) forested areas on volcano sedimentary substrate.

Distribution: Australia
Notes: This species was first described from New Caledonia before being discovered in Australia. The status of the three subspecies should be confirmed through molecular analysis.
Native status: Likely alien. Pauly & Munzinger (2003) hypothesised that this species could be alien because it was captured in Noumea, while visiting an alien plant species.
Distribution: Native range: India to Malaysia, Nepal, Vietnam.
Alien rage: Singapore, Maldives, South Africa (remains to be confirmed), widely distributed in the Pacific Region (Vanuatu, Fiji, French Polynesia).
Historical data in New Caledonia: Nouméa, Jul 1957, one female. Dec 1979-Sep 2008, one male and one female . Nouméa, 5 Feb 1999 one male and one female (Pauly and Munzinger 2003 Notes: This species was previously considered as endemic to Vanuatu (Pauly and Villemant 2009). At this stage, it is difficult to know if it is native to New Caledonia and it has remained undetected since 2017 or if its presence is due to a recent humaninduced displacement or a natural range expansion.
Occurrence status: data deficient.

Discussion
We found 23 publications, dated from 1897 to 2020, mentioning bee species in New Caledonia (Suppl. material 1). From the first published mention of New Caledonian bees to the present day, 51 bee species have been recorded, including 10 undetermined ones. Halictidae is the most represented family (26 species), followed by Megachilidae (11 species), Colletidae (10 species) and Apidae (4 species).

Occurrence status
Based on their first and last year of capture, we applied occurrence status to each species. Thus, we can confirm the presence of 30 species in the archipelago and we lack data to do so for the remaining 21 species (i.e. data-deficient species).
Amongst the 21 data deficient species, 17 have been captured only once, three have been captured twice during the same year (i.e. Lasioglossum webbi in 1992) and one is recorded for New Caledonia, based on personal communication with no specimen reported (i.e. Lasioglossum (Austrevylaeus) sp.). As bee samplings in New Caledonia prior to the 21 century have been piecemeal and far from exhaustive, it is difficult to adjudicate on their current presence. Several hypotheses may explain why those species went unrecorded during recent samplings. First, several species may have been misidentified. For example, Megachile australasiae may be, in fact, a misidentification of Megachile albomarginata . Thus, this species should no longer be considered as present in New Caledonia. Then, some species may be cryptic and contemporary samplings may have not covered their period of activity and/or their New Caledonian range. Finally, we cannot exclude the possibility that some species may have become extinct. However, the lack of exhaustive samplings of the New Caledonian bee fauna does not allow us to discriminate between the two last hypotheses. A significant part of the bee species of New Caledonia needs clarification (41%). An exhaustive sampling of the entire territory at different times of the year would provide a more precise picture of the New Caledonian bee fauna and potentially allow new species to be detected. A starting point to such a project could be to realise further samplings at periods and locations where data-deficient species were captured. This first step could provide additional information to support the presence or absence of these species. Moreover, it seems crucial to apply biomolecular analysis to the New Caledonian bee fauna. Bee species recorded from the archipelago have been described and identified exclusively, based on morphological criteria. This may have induced an over-or underestimation of the actual number of species (e.g. Praz et al. (2022)).

Origin status
Amongst the 41 identified bee species recorded in New Caledonia, 20 were categorised as endemic, 12 as native and nine as alien. However, the status of endemic and native species may evolve over time. First, it is possible that some species considered as endemic to date actually have a wider distribution across the Pacific. For example, Megachile albomarginata was considered endemic in Pauly and Munzinger (2003), but a later work showed that this species is also present in Fiji (Davies et al. 2013). Part of the southwest Pacific bee fauna remains unrecognised, as shown by the checklist of the bees st of Vanuatu, in which five of 21 species are undetermined (Pauly and Villemant 2009). Thus, it would not be surprising if future studies reveal a wider distribution of certain species.
Concerning native bee species, biomolecular analysis could clarify the origin of certain species, notably megachilid bees. Biomolecular analysis suggested that Fijian megachilid bees are mostly, if not entirely, the result of anthropogenic displacements (Davies et al. 2013). Thus, the number of alien bee species in New Caledonia could be underestimated.
For now, alien bee species account for almost 20% of the New Caledonian bee fauna. Within the last seven years, three new alien bee species have entered the territory and established themselves (i.e. Amegilla pulchra in 2016, Braunsapis puangensis in 2019 and Hylaeus albonitens in 2022). New arrivals of alien bee species will likely continue in the years to come (Seebens et al. 2017) and their establishment raises serious concerns for native species. Alien bees may negatively impact native pollinators in a multitude of ways (i.e. competition for food and nesting resources, transmission of diseases and parasites, hybridation; Zakardjian et al. (2022)). Moreover, they may disrupt plant-pollinator interactions and impact native and alien plant reproductive success (Dohzono and Yokoyama 2010). While most native bees in New Caledonia are short-tongued species, almost all alien ones (except Hylaeus albonitens) are long-tongued species. This morphological trait may allow alien bees to effectively pollinate alien plants that are poorly pollinated by native bees (i.e. awakening of sleeper weeds; Groom et al. (2014), Groom et al. (2017)). Additionally, if alien bees show morphological mismatches with native plant flowers, they may visit them without effectively transporting pollen, potentially disrupting their reproductive success.

Conclusion
As in other Pacific islands (see Villemant (2009) for Vanuatu andNaaz et al. (2021) for Fiji), the New Caledonian bee fauna is relatively poor, despite a remarkable and highly endemic plant diversity. Amongst the 51 species recorded through previous publications and recent samplings, only 30 are certainly or likely present. In order to have a more accurate picture of the New Caledonian bee fauna, there is a need to confirm species identifications and their original status through biomolecular analysis. Additionally, further samplings covering wider periods and areas are crucial to detect cryptic species and to potentially highlight local extinctions.