A global checklist of the 932 fruit fly species in the tribe Dacini (Diptera, Tephritidae)

Abstract The correct application of the scientific names of species is neither easy nor trivial. Mistakes can lead to the wrong interpretation of research results or, when pest species are involved, inappropriate regulations and limits on trade, and possibly quarantine failures that permit the invasion of new pest species. Names are particularly challenging to manage when groups of organisms encompass a large number of species, when different workers employ different philosophical views, or when species are in a state of taxonomic flux. The fruit fly tribe Dacini is a species-rich taxon within Tephritidae and contains around a fifth of all known species in the family. About 10% of the 932 currently recognized species are pests of commercial fruits and vegetables, precipitating quarantines and trade embargos. Authoritative species lists consist largely of scattered regional treatments and outdated online resources. The checklist presented here is the first global overview of valid species names for the Dacini in almost two decades, and includes new lure records. By publishing this list both in paper and digitally, we aim to provide a resource for those studying fruit flies as well as researchers studying components of their impact on agriculture. The list is largely a consolidation of previous works, but following the results from recent phylogenetic work, we transfer one subgenus and eight species to different genera: members of the Bactrocera subgenus Javadacus Hardy, considered to belong to the Zeugodacus group of subgenera, are transferred to genus Zeugodacus; Bactrocera pseudocucurbitae White, 1999, stat. rev., is transferred back to Bactrocera from Zeugodacus; Zeugodacus arisanicus Shiraki, 1933, stat. rev., is transferred back to Zeugodacus from Bactrocera; and Z. brevipunctatus (David & Hancock, 2017), comb. n.; Z. javanensis (Perkins, 1938), comb. n.; Z. montanus (Hardy, 1983), comb. n.; Z. papuaensis (Malloch, 1939), comb. n.; Z. scutellarius (Bezzi, 1916), comb. n.; Z. semisurstyli (Drew & Romig, 2013), comb. n.; and Z. trilineatus (Hardy, 1955), comb. n. are transferred from Bactrocera to Zeugodacus.


Introduction
Despite the current 'phylogenomic' age and the generation of large amounts of data on relatively few, selected, organisms, discovering and classifying new species is an ongoing endeavor of basic science that is far from complete (Zhang 2011). Major challenges to advance taxonomic work lie, among others, in the correct application of scientific species names, which in turn depends on the availability of accurate reference databases. Global initiatives to provide reference lists of species names (e.g., Roskov et al. 2017) all include major gaps that can only be filled by taxonomic specialists. Some groups of organisms are particularly challenging to manage because of the number of species they encompass, conceptual differences between workers, or the existence of unresolved problems with species identities or concepts themselves. Simultaneously, those same groups will likely benefit the most from an authoritative overview.
The fruit fly tribe Dacini is a species-rich radiation within Tephritidae and contains around a fifth of all known species in the family , Pape et al. 2011, Schutze et al. 2017. All Dacini members are frugivorous or florivorous and about 10% of the 932 currently recognized species are pests of commercial fruits and vegetables (Fletcher 1987, White and Elson-Harris 1992, Vargas et al. 2015, Freidberg et al. 2017). Among these are some of the world's economically most important pests, such as the widely introduced oriental fruit fly, Bactrocera dorsalis (Hendel, 1912), carambola fruit fly Bactrocera carambolae , and the melon fly, Zeugodacus cucurbitae (Coquillett, 1899) (De , Ekesi et al. 2016. The tribe as a whole has received considerable taxonomic attention and new species are continuously being discovered ( Fig. 1; Leblanc et al. 2015a, David et al. 2016. Dacini flies are phenotypically very similar and therefore also one of the most difficult groups of Tephritidae to identify to species-level. Whereas many Tephritidae can be identified from their intricate wing patterns, which are commonly thought to have evolved to deter predators (such as Salticidae jumping spiders [Whitman et al. 1988]), for mating rituals, or territorial behavior, the wings of most Dacini are clear with only a costal band and, usually, an anal streak. The adult chaetotaxy is a set of characters that is usually of value in dipteran species identification, but in Dacini the number of setae is reduced and similar configurations may often be homoplaseous (Hardy 1955, Hancock andDrew 2015). Their body colors, various combinations of black and yellow to red, are commonly thought to have resulted from wasp mimicry and may be under se-Figure 1. Accumulation of described species in Dacini through time, with publications that featured large numbers of newly described species indicated at their respective moment. The first species was described in 1790, but during the past fifty years the number of recognized [or known] species has more than doubled to reach the current 932. lective pressure . Diagnostic body color patterns used to separate species are further confounded by considerable intraspecific variation (Leblanc et al. 2015b). The combination of these factors has resulted in a long history of unstable classification and even though molecular phylogenetic studies are now reaching a general consensus, this has not fully trickled down to the nomenclatural level.
Dacini is a tropical and subtropical evolutionary radiation of flies with centers of diversity in Southeast Asia and Sub-Saharan Africa. Early molecular phylogenetic studies focused on pest species, often of a particular region, leading to biased results on the relationships between species that may not accurately reflect monophyletic origins or sister-group assignments (Smith et al. 2003, Nakahara and Muraji 2008. With phylogenetic studies expanding their scope beyond the pest species and increased use of molecular data, it became clear that the initial morphology-based classifications had to be revised and, in particular, the large genus Bactrocera had to be split into Bactrocera and Zeugodacus because the latter is more closely related to Dacus (Krosch et al. 2012in press, Dupuis et al. 2017. Following the most recent results, there are currently four genera in Dacini: Monacrostichus Bezzi, Dacus Fabricius, Bactrocera Macquart and Zeugodacus Hendel , Freidberg et al. 2017, although some authors (e.g.,  continue to include Zeugodacus within Bactrocera. Ichneumonopsis Hardy is now placed in Gastrozonini , Freidberg et al. 2017. Aside from shifts in generic assignments, taxa have been variably assigned to species complexes, species groups, subgenera and species-complex groups to provide some additional systematic structure, primarily for the purpose of identification keys (Clarke et al. 2005. These intermediate taxonomic ranks are mostly groups of convenience defined by unique combinations of characters rather than by synapomorphic characters. The largest and most intesively studied is the Bactrocera dorsalis complex with 88 species; the group that, incidentally, also holds the largest number of pest species. This complex, like most others, is not monophyletic (Leblanc et al. 2015b in press) and there has been synonymy of several significant pest species, such as B. papayae , and B. invadens Drew, Tsuruta & White, 2005with B. dorsalis (San Jose et al. 2013, Schutze et al. 2015a. To facilitate communication and progress of our understanding of the group, a reliable taxonomic starting point is badly needed and will enable further studies into the taxonomy and systematics of the tribe. The most recently published catalogs that covered Dacini globally are now almost two decades old , Norrbom 2004) and scattered regional treatments and keys currently comprise the largest body of references for Dacini. For Southeast Asia, there is a relatively recent two-part work including a revision  and the accompanying keys that followed (Drew and Romig 2016). These books have incorporated the previous keys for the B. dorsalis complex of Drew and Hancock (1994), but they did not adopt the latest results from a series of molecular phylogenetic works, including the split of Bactrocera into Bactrocera and Zeugodacus. For other regions, all treatments are older with increased confusion due to differing morphological terminology, species designations, and assignments. For Africa, the most recent works are two treatments from 2006 Drew 2006, White 2006), and for Australasia there is a treatment from 1989 , including keys, a proposed subgeneric classification, and revisions for the species in the region. As of 2017, the Drew and Romig books on the Asian fauna are in print and available for sale, and the other works are available digitally online and provide important catalogue references. Online resources that aim to provide up-todate species checklists such as Systema Dipterorum (Pape and Thompson 2013), the Catalogue of Life (Roskov et al. 2017), or the pest-oriented database of the Centre for Agriculture and Biosciences (CABI 2017) are outdated and have not been able to combine the regional treatments appropriately. Valid and invalid names can be verified using the Tephritidae Databases on the COFFHI website (https://coffhi. cphst.org/), but it was primarily designed for host plant information and the tephritid name searches are currently undergoing revision. Other websites, such as the "True Fruit Flies of the Afrotropical Region" (De Meyer and White 2016) or the "PACIFLY" website, covering the Pacific region (Pest Management in the Pacific Project 2003), contain valuable information, but are limited in scope and are irregularly maintained due to sporadic funding. The checklist presented here is a global overview of valid species names of Dacini. By publishing this list in paper and digital format we hope to provide a resource for those studying fruit fly taxonomy as well as researchers concerned with their impacts on agriculture. The list is largely a consolidation of previous works, but following the results from recent phylogenetic work  in press), we transfer one subgenus and eight species to different genera: Bactrocera subgenus Javadacus Hardy, considered to belong to the Zeugodacus group of subgenera by Hancock and Drew (2017), is transferred to genus Zeugodacus; Bactrocera pseudocucurbitae White, 1999, stat. rev., is transferred back to Bactrocera from Zeugodacus; Zeugodacus arisanicus Shiraki, 1933, stat. rev., is transferred back to Zeugodacus from Bactrocera; and Z. brevipunctatus (David & Hancock, 2017a), comb. n., Z. javanensis (Perkins, 1938), comb. n., Z. montanus (Hardy, 1983), comb. n., Z. papuaensis (Malloch, 1939), comb. n., Z. scutellarius (Bezzi, 1916), comb. n., Z. semisurstyli , comb. n., and Z. trilineatus (Hardy, 1955), comb. n. are transferred from Bactrocera to Zeugodacus.

Checklist
The source data is, for a large part, comprised of regional treatments , with additions and revisions from more recent studies , Yu et al. 2012, Hancock and Drew 2015, Hendrichs et al. 2015, Schutze et al. 2015b, David et al. 2016, Freidberg et al. 2017, Han et al. 2017. Species included in the list are ordered alphabetically by genus. We do not indicate subgeneric or species complex ranks because their biological significance is, at present, unclear (Leblanc et al. 2015b). We do, however, provide the checklist also in spreadsheet form in supplementary material (S1) where these ranks are included and users can sort the species to their preference. For distribution, we use a coarse geographical indication: African or Asia-Pacific. The native region is indicated in the line with the species name, invasive regions are mentioned in the notes. We also include known male lure records for each species: cue-lure, methyl eugenol, isoeugenol and zingerone. Zingerone, first reported as a male lure by Tan and Nishida (2000), has received increased attention in recent years, with new lure records, including species not attracted to the two other lures, for a number of species in Australia and Papua New Guinea (Fay 2012, Royer et al. 2015. These records are included in the list, along with previously unpublished new records, indicated as such, from our team's recent surveys carried out in Taiwan, Vietnam, Sri Lanka, Bangladesh and Nepal. For morphological terminology we follow White et al. (2000), which follows that in standard usage for other Diptera and differs somewhat from the older treatments.

Conflicting views
For some species that have recently been synonymized or where there are conflicting views by different authors, we have indicated this under the 'notes' for the respective species, so that this may help users to place different views in perspective. It should also be noted that some authors do not follow the elevation of Zeugodacus to genuslevel, because this is currently only supported by molecular data and morphological studies are inconclusive ). This affects the placement of nearly 200 species and although we agree that the reassignment of species may have initially been premature, recent studies corroborate the need to recognize Zeugodacus as a genus to maintain Bactrocera as monophyletic. A 168-species seven-gene phylogeny, including multiple Bactrocera subgenera, shows that Zeugodacus, Bactrocera and Dacus each are monophyletic, and provides moderate statistical support for a sister relationship between Zeugodacus and Dacus (San . A phylogeny with less representatives from Dacini, but 878 molecular loci, provides full statistical support for the sister relationship of Zeugodacus with Dacus, and Bactrocera as sister to both (Dupuis et al. 2017). As such, Bactrocera in the old sense is paraphyletic.

Gender agreement
Because Dacini includes both masculine and feminine genera and because species have been moved between different genera over time, there is some confusion in the literature regarding the correct application of gender agreement. We have paid particular attention to this in the checklist. Most notably, several species names ending in -fer have originally been described without the author indicating if the name should be regarded as a noun or as an adjective. Following section 31.2.2 of the Zoological Code of Zoological Nomenclature (ICZN 1999), such names should be treated as a noun in apposition and the ending should not change when the species is moved to a different genus. This applies to Bactrocera terminifer (Walker, 1860), B. speculifer (Walker, 1865) and B. curvifer (Walker, 1860).

Results
An overview of the current numbers of species split per genus, worldwide and according to the region where they are native, is shown in Table 1 (Vargas et al. 2015).   suggest. Hardy (1954) considered them synonyms, but they are treated as different species in . Their populations are likely allopatric, but there appears to be some morphological overlap. Bactrocera allwoodi (Drew, 1979). Asia-Pacific. Non-pest. Cue-lure. Bactrocera bancroftii (Tryon, 1927). Asia-Pacific. Fruit pest (oligophagous). Methyl eugenol. Bactrocera banneri White, 1999. Asia-Pacific. Non-pest. Notes: B. banneri and B. coracina are the two members of the subgenus Perkinsidacus in the most recent treatment of these species (Hancock and Drew 2017b), and both may belong in the genus Zeugodacus. They have the shallow emargination of sternite V and the long surstylus lobes of the male genitalia that fit with Zeugodacus, but lack a medial vitta on the scutum and the lateral vittae do not extend anteriorly beyond the transverse suture. Because there is, at present, no molecular data to support either placement and because it is unclear which, if any, of these morphological characters are apomorphic we tentatively leave both species in Bactrocera.
Notes: This is likely a junior synonym of B. pernigra. The only distinguishing character is in the width of the basal dark band on the scutellum, but this appears to be