The world Polleniidae (Diptera, Oestroidea): key to genera and checklist of species

Abstract A key to the world genera and a checklist of the world species for the family Polleniidae, including distributions, are provided. The following taxonomic and nomenclatural changes are proposed: Nitellia hermoniella Lehrer, 2007 = Pollenia mediterranea Grunin, 1966, syn. nov., Pollenia bentalia Lehrer, 2007 = Pollenia semicinerea Villeneuve, 1911, syn. nov., Dasypoda angustifrons Jacentkovský, 1941 = Pollenia tenuiforceps Séguy, 1928, syn. nov.; Anthracomyza Malloch, 1928, resurrected name (monotypic; type species Anthracomyia atratula Malloch) is considered a valid name and tentatively assigned to Polleniidae, giving Anthracomyza atratula (Malloch, 1927) as a resurrected combination; Morinia crassitarsis (Villeneuve, 1936), stat. rev. is considered a valid species, and Micronitellia Enderlein, 1936, stat. nov. is considered an available name.


Introduction
The family Polleniidae (cluster flies) is a small group of oestroid flies comprising 147 species (Cerretti et al. 2019 and present paper). The family group name was originally proposed by Brauer and Bergenstamm (1889) to include the single genus Pollenia Robineau-Desvoidy. Later, the Old World Pollenia sensu lato (i.e., including the morphologically diverging New Zealand Pollenia species), the Oriental genera Dexopollenia Townsend and Xanthotryxus Aldrich, and the Nearctic genus Melanodexia Williston were treated in Calliphoridae as composing the subfamily Polleniinae (or tribe Polleniini) (e.g., Hall 1965;Dear 1986;Schumann 1986;Kurahashi 1989). The group was then re-circumscribed by Rognes (1991a) to include Morinia Robineau-Desvoidy, Nesodexia Villeneuve and (tentatively) Wilhelmina Villeneuve, the latter being reassigned to the calliphorid subfamily Phumosiinae by Rognes (2011), which is followed here. More recently, Cerretti et al. (2019) elevated the group to full family rank and gave morphological and molecular evidence to support both the monophyly of the Polleniidae and the inclusion of six taxa previously assigned to the Rhinophoridae, namely Alvamaja chlorometallica Rognes and five Afrotropical species moved from the genus Phyto Robineau-Desvoidy into Morinia (Cerretti et al. 2020).
During the last few years molecular data consistently retrieved the Polleniidae (almost always represented only by Pollenia) as sister to the Tachinidae and phylogenetically distant from the 'core' Calliphoridae (e.g., Singh and Wells 2013;Winkler et al. 2015;Cerretti et al. 2017;Blaschke et al. 2018;Kutty et al. 2019;Stireman et al. 2019) but this sister group relationship has remained practically without support from morphological evidence: the currently most convincing non-molecular synapomorphy could well be the breeding habit as parasitoids of soil-dwelling invertebrates. Tachinids are parasitoids of other arthropods, and groups near the base of the family develop on soil-dwelling insect larvae (Cerretti et al. 2014;Stireman et al. 2019); the natural history and host range information of polleniids is limited to a handful of Pollenia species (Keilin 1915;Tawfik and El-Husseini 1971;Yahnke and George 1972;Rognes 1991a;Szpila 2003;Grzywacz et al. 2012;El Husseini 2019), and all of these develop as endoparasitoids in earthworms. Marshall (2020) observed a native New Zealand Pollenia displaying what can only be interpreted as oviposition behaviour, extending the ovipositor into a mixture of loose soil and organic debris. Recent field observations of adults of an unidentified species of Melanodexia revealed that females have a similar behaviour to that observed for several Pollenia and other parasitoids of soil-dwelling organisms: they can be seen walking frenetically on bark lying on the ground, keeping wings folded on their back (SG pers. obs. 2019, California).
We recognise 147 species of Polleniidae classified into eight genera (Fig. 1). Pollenia is the most species-rich and widespread genus, with 95 species described from the Palaearctic, Oriental and Australasian regions [and seven species known from the Nearctic Region as introductions (Rognes 1991a;Whitworth 2006;Jewiss-Gaines et al. 2012, Bowser 2015] (Fig. 2). The remaining seven genera are considerably less diverse: the single species assigned to Alvamaja Rognes (A. chlorometallica Rognes) is recorded from a few localities in southeastern Europe, Dexopollenia comprises 21 species from the southeastern Palaearctic and the Oriental Region, the Nearctic endemic genus Melanodexia includes eight species, Morinia contains 13 species from the Afro- tropical and Palaearctic regions and Xanthotryxus includes seven species from southeastern Palaearctic and the Oriental Region. The remaining genera, the monotypic Australian Anthracomyza Malloch and the Palaearctic Nesodexia Villeneuve, are here tentatively assigned to the family. Remarkably, there are no Polleniidae recorded from the Neotropical Region, neither native nor introduced.
Many regional catalogues and keys to polleniid genera (and species) have been published in the recent decades (e.g., Hall 1948;James 1955James , 1970James , 1977Lehrer 1972;Pont 1980;Shewell 1987;Rognes 1991aRognes , 1998Kurahashi 1995;Jewiss-Gaines et al. 2012;Pape et al. 2015), and an incomplete key to genera is available from Peris (2004). The aim of the present paper is to lay the foundation for future taxonomic and phylogenetic studies by producing a key to the world genera of the family Polleniidae and a checklist of the world species.

Key to genera
A dichotomous key to the adults of both sexes was constructed to contain the genera here considered members of the Polleniidae. This means that we are excluding the monotypic genera Wilhelmina Schmitz &Villeneuve, 1932 andNepenthomyia Kurahashi &Beaver, 1979, both associated with pitcher plants of the genus Nepenthes and not considered polleniids. Wilhelmina was considered a possible member of Polleniidae (or Polleniinae, or Polleniini) by Schmitz and Villeneuve (1932), Fan (1965Fan ( , 1992  1997) and Rognes (1991a), but we follow Rognes (2011) in suggesting its reassignment to Phumosiinae. Nepenthomyia was considered "closely related" to Wilhelmina by Kurahashi and Beaver (1979), although with no indication of subfamily assignment, and a position within the Polleniinae was not accepted by Rognes (1991a). The key was constructed through direct examination of available material and from literature data. No specimens of Anthracomyza were examined and characters for it were derived from the original description by Malloch (1927).
Digital images of the specimens shown in Fig. 1C, D were taken using a Canon EOS 6D camera equipped with Canon Photo lens MP-E 65 mm 1.2.8 and processed by Canon Digital Photo Professional (Canon: Ōta, Tokyo, Japan), Combine ZM by Alan Hadley and GIMP 2.10.4 by Alexandre Prokoudine.

World checklist
The world checklist is based on original literature, though following the papers by Rognes (1987bRognes ( , 1991aRognes ( , 1998Rognes ( , 2010, Evenhuis et al. (2004Evenhuis et al. ( , 2010Evenhuis et al. ( , 2015Evenhuis et al. ( , 2016, O'Hara et al. (2011 and literature therein. It lists all currently valid nominal genera and species of the family Polleniidae including their synonyms and associated nomina nuda. Incorrect subsequent spellings have been entered to the extent they have come to our knowledge. Valid names of taxa are arranged alphabetically according to genus and species (no subfamilial or tribal classification is recognised here). Each genus-group name is listed with author, year, page, type species with author and date, and form of type fixation with author and date. Each type species is given in its original binomen (ICZN 1999;recommendation 67B), and if that name is a junior synonym it is followed by the valid name of the species in square brackets. Each species-group name is listed with author, year, page, type locality in standardised modern spelling (original quote in parenthesis if needed to avoid confusion), and relevant nomenclatural details (homonymy, lectotype designations, etc.). Unavailable names are listed with an explanation as to their unavailability, and incorrect subsequent spellings are given with the relevant reference. New specific synonyms are based on comparisons of the original descriptions of the nominal species in question with material (specimens, photos, illustrations) available to the authors. Additional information may be given under "Remarks". Distributional data are based on the literature and online databases (Fauna Europaea [Pape et al. 2015] for Palaearctic species) but we do not refer all records to their original sources. Records have been entered to the extent they have come to our knowledge and they are reported for countries and major islands, except that larger countries are recorded at the level of state or province.

Caveat for key users
Notwithstanding strong molecular evidence suggesting the monophyly of Polleniidae, members of the group apparently do not share any unique morphological apomorphy. For this reason, it is not possible to provide a simple and clear-cut diagnosis of this oestroid family since several exceptions have to be taken into account once given a common set of characters. Therefore, we refer to Cerretti et al.'s (2019) diagnosis of the Polleniidae, though highlighting all the uncertainties related to it.
Diagnosis. Small to medium-sized oestroid flies varying from yellow to black in ground colour. Facial sclerite at least weakly carinate [with few exceptions, e.g. Pollenia griseotomentosa (Jacentkovský)]. Stem vein bare dorsally. Anal vein not reaching wing margin. Posterodorsal margin of hind coxa bare. Prosternum and proepisternal depression bare. Postalar wall setose (occasionally bare in small specimens of Morinia). Female: ovipositor sclerite length moderate; sternite 8 of ovipositor elongate with posterior margin entire; cerci long and narrow. Male: ventral and ventrolateral surface of distalmost parts of distiphallus smooth.
A comprehensive phylogeny including all the eight genera recognised as cluster flies is still awaiting, as well as a thorough revision of the family, therefore generic boundaries within this family are still labile due to the absence of molecular evidence and strong morphological characters. For these reasons, we are here applying a traditional generic division without proposing any subfamilial or tribal classification nor any new generic synonyms. We anticipate future rearrangements, such as Dexopollenia synonymised with Pollenia, or the exclusion of Nesodexia. Moreover, the New Zealand Pollenia are still entirely untouched in a phylogenetic context.

Distribution. Palaearctic -China (Sichuan).
Remarks. The name-bearing type of M. crassitarsis has not been located, but unpublished studies (by TP) of Chinese specimens matching the original description would seem to support a status for this nominal species as valid.
Distribution. Australasian -Australia (New South Wales). Remarks. The Australian Faunal Directory lists the species as Anthracomyia atratula Malloch, 1927(Elliot 2007, while it is listed as Morinia atratula Malloch, 1927 in the Catalogue of Life (Roskov et al. 2019). Malloch (1928: 360) proposed Anthracomyza as a new replacement name for his own Anthracomyia, correctly arguing that the latter is "preoccupied by Anthracomyia Rondani". Anthracomyza was later listed as an unnecessary new name in the catalogue of Australasian Diptera (Kurahashi 1989), probably because Rondani (1856) originally gave the spelling Anthracomya, which differs by one letter and therefore does not enter into homonymy (ICZN 1999;article 56.2). However, as given by O'Hara et al. (2011), Rondani (1868 later emended his own spelling to Anthracomyia, and although this is now recognised as an unjustified emendation, it is an available name with separate authorship and therefore preoccupies Anthracomyia of Malloch (1927).
We here maintain Anthracomyza as a valid, monotypic genus; however, a careful examination of male and female terminalia is necessary to ascertain whether Anthracomyza belongs to Polleniidae.

Distribution. Palaearctic -France (Corsica).
Remarks. According to Rognes (1991a) Nesodexia corsicana has the ventral and lateroventral surface of distalmost parts of acrophallus provided with scale-like spinules (Rognes 1991a), thus the species does not share a key synapomorphic character state supporting monophyly of Polleniidae. Moreover, the general habitus and, in particular, the head profile, characterised by a prominent lower facial margin, of N. corsicana are reminiscent of many phumosiine calliphorids. However, unlike all phumosiines, the katatergite of Nesodexia is bare (Rognes 1997), and more data are needed to resolve its phylogenetic position.