Synopsis of the parasitoid wasp genus Cotesia Cameron, 1891 (Hymenoptera: Braconidae: Microgastrinae) in Australia, with the description of seven new species

The genus Cotesia Cameron, 1891 is one of the most diverse of the Microgastrinae, a subfamily of wasps that are exclusively endoparasitic on lepidopteran larvae. Species of Cotesia are widely utilised as biological control agents across the world. In Australia, there are currently 10 confirmed native species as well as four species introduced for the management of lepidopteran pests. The genus is morphologically conserved and has not been studied in the Australasian region for many decades. In this study, we use both comparative morphology and sequence data from the COI gene to delineate species, and in so doing describe seven new species from Australia: C. lasallei sp. nov., C. medusae sp. nov., C. ocellata sp. nov., C. reidarum sp. nov., C. scripta sp. nov., C. tjapekki sp. nov. and C. wonboynensis sp. nov., raising the number of species of Cotesia formally recorded in Australia to 21. We also provide updated descriptions of the previously described native species, diagnoses for the introduced species and a key to all currently described species found on the continent and from Papua New Guinea. This study treats only a fraction of the likely diversity of Cotesia, but provides a solid framework for future work.


Introduction
Microgastrinae are an extremely diverse, cosmopolitan subfamily of braconid parasitoid wasps. All species in the subfamily are endoparasitoids of lepidopteran larvae and often extremely host specific (Smith et al. 2008). Cotesia Cameron, 1891 is one of the most utilised genera of Microgastrinae Foerster, 1862 for the biological control of lepidopteran pests, with numerous species translocated around the world (Whitfield 1997). There are approximately 320 species described worldwide (Yu et al. 2016;European Journal of Taxonomy 667: 1-70 ISSN 2118-9773 https://doi.org/10.5852/ejt.2020 www.europeanjournaloftaxonomy.eu 2020 · Fagan-Jeffries E.P. & Austin A.D. This work is licensed under a Creative Commons Attribution License (CC BY 4.0).
Generally, Cotesia is very conserved morphologically among species and also shows a large intraspecific variation in characters often used to delimit species of microgastrines in other genera, such as the intensity of propodeal and tergite sculpturing. This causes difficulties in delimiting potentially new species from historically described species where no molecular or host information are available. In this study, we use a molecular dataset partly developed for a recent large-scale DNA barcoding study on Australian microgastrines (Fagan-Jeffries et al. 2018) to inform and direct species discovery and delimitation, and also to place some of the described species in relation to the new species, using publically available cytochrome oxidase subunit 1 (COI) data. We describe seven new species of Cotesia from Australia and provide a key to all described species found on the continent and from Papua New Guinea. We update the descriptions of native species with additional measurements and characters not included in original descriptions and, for the introduced species, include a diagnosis against other Australian species. Whilst both molecular data and examination of material in collections suggest that many more species of Cotesia exist on the continent, this study provides a firm starting point from which further work on the genus can be conducted.

Material and methods
Terms for general morphology follow Fernández-Triana et al. (2014) who combined traditional microgastrine morphological terms, such as those used by Mason (1981), with the standards introduced in the Hymenoptera Anatomy Ontology (HAO) project (Yoder et al. 2010;Seltmann et al. 2012) (Fig. 34). Terms for sculpture follow Eady (1968). For new species, measurements of holotypes are given, with those for paratypes (where measured) following in parentheses as ranges (see Fernández-Triana et al. (2014) for measurement terminology and appendix 1 in the same paper for discussion on characters prone to variable results when measuring). We define colour as either pale (white, cream or pale yellow), orange, light brown or dark (dark brown or black).
Distribution data presented in Austin & Dangerfield (1992) is included, but no attempts were made in this study to verify these data with specimens in collections and it should therefore be treated with caution. We use the Australian Faunal Directory (https://biodiversity.org.au/afd/home) for the current accepted names of lepidopteran hosts. Host data are only provided for the region of Australasia, several of the introduced species of Cotesia have wider host ranges in other part of the world. Images of specimens of Cotesia and characters were generated using a Visionary Digital BK+ imaging system with a Canon EOS 7D 18 megapixel camera, compiled in Zerene Stacker, Zerene Systems LLC, PMax software, and cropped and resized in Adobe Photoshop CS6 (Adobe Systems Inc., San Jose, CA, USA). Scanning Electron Microscope (SEM) images were generated on a FEI Quanta 450 FEG Environmental SEM. Nearly all specimens of the new species included in this study have had legs removed for DNA extraction, and thus nearly all type specimens are missing 1-3 legs. DNA extraction and sequencing methods for newly sequenced specimens follow the Sanger sequencing methods in Fagan-Jeffries et al. (2018), other than the specimen BOLD: AUMIC544-19, which was sequenced with the primers LepF (Hebert et al. 2004) and C113R (Smith et al. 2008). A Bayesian tree of the standard barcoding region for COI of the specimens of Cotesia and Glyptapanteles sequenced in Fagan-Jeffries et al. (2018), plus several newly generated sequences, and all sequences of species of Cotesia known to occur in Australasia publicly available on the Barcode of Life Database (BOLD), was constructed using the program MrBayes (Ronquist et al. 2012). COI was not partitioned and a GTR+I+G model of evolution was used. The tree was run for 20 000 000 generations and convergence was established using the program Tracer (Rambaut et al. 2018)

Results
From the large COI DNA barcoding dataset of Microgastrinae for Australia (Fagan-Jeffries et al. 2018), and additional specimens sequenced for this study, 16 potential new species of Cotesia from Australia were recognised using a 2% divergence level threshold approach, as validated for microgastrines in previous studies (Smith et al. 2008(Smith et al. , 2013 (Fig. 1). Seven of these species are not treated here due to being represented by single specimens. Two species were not able to be definitively separated from described species (C. icipe Fernández-Triana & Fiaboe, 2017 and C. radiantis), and further discussion on these cases is provided below. The remaining seven species are here described as new.

Type species
Cotesia flavipes Cameron, 1891: 185. For a complete list of synonyms and bibliography, see Fernández-Triana et al. (2020). of the seven newly described species of Cotesia as well as a Bayesian phylogeny of these species and the other Cotesia spp. known to occur in Australia with publically available COI data. Species listed in black are those treated in this study, whilst those in grey are not treated due to being represented by single specimens, or are BOLD sequences that fall outside the rest of that species clade and are therefore possible misidentifications. BOLD codes for these single specimens are given at the end of the label in the tree. Symbols on the map correspond to the symbols on the phylogeny. Bayesian posterior probability values of ≥ 95 are represented by *, whilst those of 90-94 inclusive are represented by °. The number of sequences in collapsed clades are given in brackets (n = x). Outgroups have been removed for simplicity.

Comments
Cotesia is one of the largest genera of Microgastrinae and was previously treated as the glomeratus group of Apanteles sensu lato (s.l.). A detailed description of the genus can be found in Mason (1981), whilst supplementary notes, a discussion of putative relationships and a list of world species can be found in Fernandez-Triana et al. (2020). Members of the genus can be diagnosed by the absence of a fore wing areolet, possessing a short inflexible hypopygium and a short ovipositor, T1 generally parallelsided or broadening posteriorly (occasionally in the Australian fauna with T1 narrowing posteriorly but never to the extent as in Glyptapanteles) and T2 normally broad and rectangular. Often species have the propodeum with a medial carina, and the propodeum, T1 and T2 coarsely sculptured.

Host information for Australian species of Cotesia
In Australia, species of Cotesia are known to parasitise at least nine different lepidopteran families (Table 1), with some species appearing highly specialised on individual host species or genera, whilst others (e.g., C. ruficrus) are recorded from numerous genera within a single family. The host information presented below, and in the following key, are restricted to records from Australia; for extra-limital host records see references listed under the treatment of each species.

Key to the described species of Cotesia in Australia and Papua New Guinea
This key is based on adult females and should be treated with some caution due to the variable nature of some characters in Cotesia, and the limited number of specimens available for some species during the construction of the key. Wherever possible, morphological identification should be supported by COI barcoding or host information.

Diagnosis
Cotesia anthelae can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: T2 ovoid shaped with very strong, distinct crenulate margins; scutoscutellar sulcus with 6-8 pits; scutellar disk mostly smooth with some shallow punctures; T1 mostly parallel sided.  Paratype AUSTRALIA • 1 ♀; same collection data as for holotype; NHMUK 3.c.002.

Female
From holotype; propodeum, T1 and metatibal spur/metabasitarsus ratio measured and imaged on paratype due to position of wings and legs in holotype.
Body length. Head to apex of metasoma: 2.6 mm.
Mesosoma. Anteromesoscutum irregularly punctate, punctures not perfectly round, some areas tending towards reticulate rugose; number of pits in scutoscutellar sulcus 7 or 8; scutellar disc mostly smooth with some shallow punctures; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum (unable to measure due to wing placement).
Propodeum. Medial carina present, propodeum smooth and shiny at anterior corners, posterior to this propodeum with irregular rugose sculpturing, posterior corner thirds with shiny area amongst sculpturing carinae.
Metasoma. T1 length / T1 width at posterior margin 1.7; T1 generally parallel sided, very slightly narrowing at centre, and more significantly curving inwards (narrowing) at posterior corners, smooth in anterior half, with large shallow punctures in posterior half; T2 width at posterior margin / T2 length 1.6, the sclerotised area of T2 very rounded, sculpturing irregular in anterior half, smoother in posterior half, with clear short crenulate furrow at posterior boundary and lateral edges; T2 length / T3 length 0.7; T3 sculpture smooth and shiny; ovipositor sheaths length/hind tibial length 0.2.

Host
Anthela ocellata (Walker, 1855) (record from type specimens) (Anthelidae), cocoons white. Opodiphthera eucalypti (Scott, 1864) (from Austin & Dangerfield 1992, this record remains unconfirmed as no corresponding specimens were located in collections, and is possibly erroneous). Wilkinson (1928a) notes in the description of this species that the manuscript name "Apanteles dardalae" given by Cameron on the holotype label appears to have not been published.

Diagnosis
Cotesia australiensis can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: anteromesoscutum punctate; fore wing r vein straight; antennal flagellomere 14 length/width > 2; scutellar disk mostly smooth with some shallow punctures and centre of the medial posterior band of the scutellum smooth; T1 parallel sided; T2 trapezoid with shallow crenulate border with T3; T3 with very sparse setae over most of tergite, mostly dark.

Material examined
Holotype (
Body length. Head to apex of metasoma: 2.5 mm.

FAGAN-JEFFRIES E.P. & AUSTIN A.D., Synopsis of Cotesia in Australia
Mesosoma. Anteromesoscutum with regular, shallow, small punctures, appearing smoother than most species treated; number of pits in scutoscutellar sulcus 10; scutellar disc mostly smooth with some shallow punctures; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum 0.6.
Metasoma. T1 length / T1 width at posterior margin 1.3; T1 parallel sided, curved at posterior corners, smooth in anterior half, with large shallow punctures in posterior half; T2 width at posterior margin / T2 length 2.3, the sclerotised area of T2 almost an isosceles trapezium, mostly smooth with some irregular sculpturing, wide, very shallow crenulate border at lateral edges; T3 sculpture smooth and shiny; ovipositor sheaths length/hind tibial length 0.2.

Male
Unknown.

Distribution
Type locality given only as Australia. Vic (record from examined material), NSW, Tas (records from Austin & Dangerfield 1992).

Remarks
We base the redescription and measurements of this species on the series of specimens in ANIC, with reference also to the images of the holotype available on the Smithsonian Entomology Collections database (http://n2t.net/ark:/65665/3639803ce-5025-4c10-bd7c-42bba8777cdf). (Bingham, 1906) Figs 2C, 15

Diagnosis
Cotesia deliadis can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: anteromesoscutum punctate; scutellar disk smooth with some shallow punctures associated with setae; centre of medial posterior band of the scutellum smooth; T1 mostly parallel sided; T3 dark with sparse setae only in posterior half; fore wing 2RS not creating a 'stub' at junction with r.
Body length. Head to apex of metasoma: 2.3 mm.
Mesosoma. Anteromesoscutum punctate, punctures reasonably regular but denser along notauli lines compared to lateral areas and very centre; number of pits in scutoscutellar sulcus 9 or 10; scutellar disc mostly smooth with some shallow punctures associated with setae; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum 0.5.
Propodeum. Medial carina present, irregularly rugose and punctate, posterior third smooth and shiny with sparse carinae.
Metasoma. T1 length / T1 width at posterior margin 1.8; T1 generally parallel sided, very slightly narrowing at posterior corners, very smooth, some very shallow punctures in posterior half; T2 width at posterior margin / T2 length 2.6, the sclerotised area of T2 not well distinguished at lateral edges, curved at anterior margin, T2 mostly smooth, some irregular pits by anterior and lateral edges; T2 length / T3 length 0.7; T3 sculpture smooth and shiny; ovipositor sheaths length/hind tibial length 0.18.
Male (from M.F. Braby specimens) As female, but with denser (still shallow) punctures on scutellar disc, T2 slightly longer in relation to posterior width, sclerotised area of T2 more sharply defined, closer in shape to subtriangular or hemispherical.

Remarks
We examined and sequenced specimens reared from Delias aganippe in Victoria by E. and J. Reid, which are morphologically very similar to C. deliadis, but much darker in colouration. With a 145 bp COI barcode available for a specimen identified as C. deliadis and reared from D. argenthona from the type locality in Qld (BOLD: AUMIC544-19), which differs by only 1 bp from the sequences of the Victorian specimens (BIN BOLD:ADL3255, Fig. 1), we assign the Victorian specimens to C. deliadis. Darker colouration in conspecific specimens further south in Australia is common, and this therefore expands the distribution of C. deliadis. Correspondingly, we assign the full-length COI barcode from the Victorian specimens to C. deliadis.
We also note that the species C. deliadis strongly resembles Glyptapanteles deliasa Austin & Dangerfield (1992) which parasitises D. aganippe in South Australia, but we were unsuccessful in obtaining DNA from paratypes of this species for comparison and therefore do not treat G. deliasa here.

Diagnosis
Cotesia erionotae can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: T3 pale; T1-2 mostly pale; antennae pale.

Remarks
We include this species in the key due to its presence in Papua New Guinea, but it is not known to occur in Australia.  (Wilkinson, 1928), holotype, ♀ (NHMUK 3.c.1037). B. Cotesia geometricae Austin, 2000, paratype (WINC).

Diagnosis
Cotesia flavipes can be separated from nearly all other species of Cotesia currently described from Australia and Papua New Guinea by the dorsoventrally flattened mesosoma, and can be separated from the closely related Cotesia nonagriae (described from Australia, presence in Papua New Guinea unconfirmed) most easily by molecular data (e.g., Genbank: DQ232336 from India, DQ232335 from Pakistan) (Muirhead et al. 2012; also see under C. nonagriae), and also by the shape of the face projection between the antennal bases. When viewed under SEM, the projection is more prominent and rounded with a relatively smooth margin, compared to C. nonagriae which has the margin tri-lobed and the projection less prominent.

Material examined
Other

Distribution
Cotesia flavipes is a cosmopolitan species which has been introduced into several countries (e.g., Barbados, Brazil, Colombia, Kenya, Mauritius, Peru, Rèunion, South Africa, USA) for biological control purposes (Muirhead et al. 2006). In the Australasian region it is recorded from Papua New Guinea (Muirhead et al. 2012).

Remarks
We include this species in the key due to its presence in Papua New Guinea. Until recently, C. flavipes was recorded as being present in Australia, but the Australian specimens have been identified as C. nonagriae, and this species raised from synonymy (Muirhead et al. 2008). Cotesia flavipes is therefore not currently recorded from Australia, but may possibly be introduced in the future for the control of stem-boring pests. Austin, 2000 Fig. 16B

Diagnosis
Cotesia geometricae can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: anteromesoscutum punctulate, much smoother than most other species; T3 only sparsley setose, mostly in posterior third; T1 parallel sided; fore wing vein 2RS not extending past junction with r, not creating a 'stub', hypopygium not emarginate.

Distribution
Currently known from Victoria and the Australian Capital Territory (from examined material and original description).

Remarks
As this species is only recently described and we feel that the description is relatively complete, we provide only new characters or measurements relevant to its diagnosis, and refer otherwise to the description in Schumacher et al. (2000).

Distribution
Found in all states and territories of Australia, see Fernández-Triana et al. (2020) for global distribution.

Diagnosis
Cotesia kazak can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: anteromesoscutum punctulate, smoother than most species; T3 densely setose; fore wing 2RS not creating a 'stub' at junction with r; T1 parallel sided.

Male
As female but hind leg all dark, body longer (2.5 mm body length), 14 pits in the scutoscutellar sulcus, some variation in measurements including T1 length / T1 width at posterior margin 1.6, and notably the antenna lengthened: antennal flagellomere 2 length/width 2.8; antennal flagellomere 14 length/width 3.6.

Distribution
Currently only known from arid South Australia, in regions surrounding Lake Torrens.

Host
Unknown.

Remarks
This species constitutes the BIN BOLD:ADL2795, and is greater than 3% divergent from any other sequences available on the database. Diagnosis Cotesia medusae sp. nov. can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: anteromesoscutum punctate reticulate; antennal flagellomere 14 length/width > 2; scutellar disk smooth with only shallow punctures associated with setae and with centre of medial posterior band of the scutellum smooth; fore wing r vein straight and fore wing 2RS not creating a 'stub' at junction with r; T2 semicircle shaped, border with T3 often shallowly crenulate; T3 mostly dark with some pale areas and sparse setae over most of tergite, notauli clearly differentiated from surrounding anteromesoscutum sculpturing.

Etymology
This species is named after Medusa, the 'Gorgon' or monster in Greek mythology. Gorgonphone, the species name of the host (Catopsilia gorgophone Boisduval, 1836) means 'Gorgon-Slayer'. Thus, there seems to be an appropriate amount of irony in the lethal parasitoid of this species being named for the original monster.

Male
As female, only slight variations in measurements.

Distribution
Currently only known from the type locality in Brisbane, although as the host is found through much of New South Wales and Queensland, this parasitoid also likely has a wider range than currently documented.

Remarks
Whilst only a 340 bp section of COI was obtained for this species, which is not long enough to be assigned a BIN on BOLD, it is clearly distinct morphologically and genetically from any publically available sequence of Cotesia, and as host data is available we felt the description was warranted despite not possessing a full COI barcode. (Olliff, 1893) Figs 5A, 6A-B, 21

Diagnosis
Cotesia nonagriae can be separated from nearly all other species of Cotesia currently described from Australia and Papua New Guinea by the dorsoventrally flattened mesosoma, and can be separated from the closely related C. flavipes (recorded from Papua New Guinea, presence in Australia unconfirmed) most easily by molecular data (see Muirhead et al. 2012) but also by the shape of the face projection between the antennal bases; when viewed under SEM, the projection is flatter (less rounded/prominent) than that of C. flavipes, and has a margin that appears tri-lobed, compared to the smooth margin of the face projection of C. flavipes.

Remarks
Cotesia nonagriae was recently removed from synonmy with C. flavipes (Muirhead et al. 2008) based on distinct molecular COI and 16S data and subtle morphological characters. Genbank numbers for verified C. nonagriae specimens are: JQ396711-13, JQ396704, JQ396747. The C. flavipes complex is a monophyletic group of seven described species (Fujie et al. 2018) that possess a dorsoventrally compressed mesosoma, most likely an adaptation for parasitising their stemboring hosts. Wilkinson synonymised Apanteles flavipes firstly with A. nonagriae Viereck, 1913 from Taiwan, and then later with A. nonagriae Olliff, 1893 from Australia. Muirhead et al. (2008) then raised C. nonagriae Olliff from synonymy with C. flavipes and redescribed the species. As this species is only recently redescribed and the description is relatively complete, we provide only new characters or measurements relevant to the diagnosis, and refer otherwise to the description in Muirhead et al. (2008). Diagnosis Cotesia ocellata sp. nov. can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea primarily by the small ocelli, OOL/posterior ocellus diameter > 2.5 (and normally greater) and T1 broadening consistently posteriorly, almost wedge shaped.

Etymology
The species name is from the Latin adjective 'ocellatus', meaning 'small eyed' and reflecting the small size of the ocelli of this species compared to other Australian Cotesia. It is a feminine nominative singular adjective.
Mesosoma. Anteromesoscutum reticulate rugose, in holotype relatively consistent across anteromesoscutum, in paratypes much smoother in centre and lateral sides; number of pits in scutoscutellar sulcus 9 (8); scutellar disc with only shallow pits associated with setae; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum 0.6 (0.7).

Male
Unknown.

Distribution
Currently only known from the wider Adelaide region, South Australia.

Remarks
This species constitutes the BIN BOLD:ADL5560, and has a maximum intraspecific distance of 0.16% and a distance of 2.59% to the nearest neighbour.

Diagnosis
Cotesia philoeampa can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea primarily by the fore wing vein 2RS extending past junction with r, creating a 'stub'; T1 parallel sided; anteromesoscutum punctate; fore wing vein r generally shorter than 2RS (occasionally of similar length).
Body length. Head to apex of metasoma: 2.9 mm.
Mesosoma. Anteromesoscutum punctate, punctures denser and tending towards rugose along notauli lines compared to lateral areas and centre area, where punctures are shallower and ateromesoscutum appears smoother; number of pits in scutoscutellar sulcus 9; scutellar disc mostly smooth with some shallow punctures associated with setae; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum 0.5.
Propodeum. Medial carina present, but carina not more prominent than other carinae on propodeum, all of propodeum irregularly rugose.
Metasoma. T1 length / T1 width at posterior margin 1.3; mostly parallel-sided, curving in right at posterior edge and bulging outwards a small amount in the posterior third, smooth in anterior half, posterior half with rugose and reticulate rugose sculpturing; T2 width at posterior margin / T2 length 2.3, wide and flat, covering most of metasoma width, slightly curved from anterior corners, very shallow indistinct sculpturing around edges -rugulose at anterior end, crenulate at posterior margin; T2 length / T3 length 0.7; T3 sculpture smooth and shiny; ovipositor sheaths length/hind tibial length 0.16.

Male
Unknown.

Distribution
Only known from the type locality, Wattle Flat, NSW.

Diagnosis
Cotesia radiantis can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: T1 broadening posteriorly, wedge shaped; fore wing vein r shorter than 2RS; anteromesoscutum sculpturing considerably stronger along  (2020) lines of notauli than elsewhere, almost smooth laterally -basal area of anteromesoscutum almost reticulate rugose; scutellar disk with shallow punctures; mesosoma not dorsoventrally flattened; hind coxa and metasoma only slightly paler than mesosoma; ocelli normal sized, OOL/posterior ocellus diameter < 2.4.

Male
Not examined, male included in the original description of Wilkinson as not remarkably different from female type.

Distribution
Only known from type locality in south-eastern Queensland.

Host
Agrotis munda (Walker, 1857) (Noctuidae). Record from type specimens. Note that the host name given in the original description is Euxoa radians Guénée, 1852, but that the actual host is not conspecific with Agrotis radians Guénée, 1852 due to confusion historically in the Australian literature. See Common (1958) for details. Gregarious (based on multiple specimens from description with identical label data).

Remarks
There are a large number of specimens sequenced as part of this project (BIN: BOLD:ABA6232, Fig. 1) that are broadly distributed across Australia and that strongly resemble C. radiantis. As this group of specimens cannot be separated definitively from the holotype of C. radiantis, we do not describe them as a new species, but also hesitate to confirm their identity as C. radiantis until a specimen reared from Agrotis munda is sequenced and the COI barcode compared (no attempt was made to extract DNA from the holotype of C. radiantis).  (Waterhouse, 1903) (Lepidoptera: Lycaenidae; BOLD AUMIC491-18; Genbank COI: MH138904); QM T246703.
Legs. Hind tibia inner spur length/metabasitarsus length 0.7 (hind tarsi broken off and lost in holotype after measurement taken).
Propodeum. Medial carina present, but smooth and indistinct in centre, rest of propodeum rugose, with some carinae more prominent than others.

Male
Unknown.

Distribution
Only known from type locality in the White Mountains National Park, Queensland, but the host extends westward throughout the northern parts of the Northern Territory and Western Australia, reflecting the possible distribution of this parasitoid.

Remarks
This species constitutes BIN BOLD:ADL4430, and is 2.31% divergent from the nearest neighbour on BOLD.

Diagnosis
T1 consistently broadening posteriorly, often with a middle patch of rugosity on posterior smooth band of scutellum (under scutellar triangular disk); scutellar disk with only shallow punctures; mesosoma not dorsoventrally flattened; hind coxa and metasoma dark; ocelli normal sized, OOL/posterior ocellus diameter < 2.4; fore wing vein r normally longer than, occasionally of similar length to 2RS; distal quarter of hind tibia darkened, including in lateral view.

Diagnosis
Cotesia ruficrus can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: T1 consistently broadening posteriorly, wedge shaped; scutellar disk with strong punctures; T3 with only a single row of setae in posterior half.   (Haliday, 1834), syntype of C. sydneyensis (Cameron, 1911) (Hufnagel, 1766), Helicoverpa armigera (Hübner, 1808) (Noctuidae): records from examined material. Gregarious. Also known from a wide range of noctuid genera in Australia and across the world, including the problematic quarantine pest Spodoptera frugiperda (Smith, 1797) which has recently been identified to have arrived in Australia (see Shenefelt 1972: 617;Nixon 1974: 495;Austin & Dangerfield 1992: 22;Gupta et al. 2019;Yu et al. 2016 for a list of host genera).

Remarks
Cotesia ruficrus represents an interesting case, in which strains were introduced from Pakistan to Australia and New Zealand (Cumber et al. 1977;Michael 1985) and possibly represent a different species to the native Apanteles sydneyensis, which was synonymised with the (also) native Apanteles antipoda by Wilkinson (1928a) and then with C. ruficrus (Wilkinson 1929). Until specimens are found which are distinct from the introduced C. ruficrus at a molecular level and morphologically identical to the types of A. sydneyensis or A. antipoda, we leave the synonymy with A. ruficrus intact, and assume the species to be cosmopolitan. We note that a similar situation occurred in the case of C. nonagriae and C. flavipes, where the two species were erroneously synonymised based on morphology and later discovered to be distinct species based on molecular evidence and subtle morphological differences (Muirhead et al. 2008), thus excluding C. flavipes from the Australian fauna. Due the uncertainty surrounding whether this species is cosmopolitan, or consists of both introduced and native unrelated lineages, we do not redescribe the species and provide only a diagnosis that encompasses both the examined types of A. sydneyensis and the sequenced specimens that are placed in a COI tree with northern hemisphere C. ruficrus specimens.
In the key of Nixon (1974), C. ruficrus is differentiated from C. vestalis by the character of the third tergite having setae restricted more or less to a single row on the anterior half of the segment (as opposed to third tergite being covered all over with fine setae, except for small mid-basal area). In the Australian specimens, the presence of only a single row of setae in C. ruficrus is valid, but it is not always in the anterior half of the tergite (more often in the posterior half). Additionally, Nixon (1974) notes that in C. ruficrus the scutellum has punctures, at least on anterior half, wide enough apart to leave smooth shiny interspaces fully equal to width of punctures, whist in C. vestalis there are no smooth interspaces anteriorly. In the specimens of C. vestalis examined for this study, there was variability in this character and some specimens had sculpturing resembling that of C. ruficrus -hence this character is not used in the key presented here. (Bingham, 1906) Figs 3B, 28
Propodeum. Medial carina distinct but only in posterior half, absent in centre of propodeum, rest of propodeum irregularly rugose, other than carinae around spiracles, rugosity much less defined than medial carina.

Male
Not examined (nor examined by Wilkinson 1928a), male described by Bingham (1906) as slightly shorter with abdomen shorter and more truncate posteriorly.

Distribution
Only known from Queensland.

Diagnosis
Cotesia scripta sp. nov. can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the following combination of characters: fore wing vein r subtly curved; fore wing 2RS not creating a 'stub' at junction with vein r; anteromesoscutum punctate reticulate; scutellar disk smooth, or with only shallow punctures associated with setae and centre of medial posterior band of the scutellum smooth; T1 parallel sided; border between T2 and T3 smoothly indented; T3 dark with sparse setae on posterior two thirds.

Etymology
This species is derived from the Latin for 'write' as an acknowledgement of The University of Adelaide Environment Institute funded Early-Mid Career Researcher writing retreat, on which this paper was partially compiled. It is a feminine nominative singular adjective.
Mesosoma. Anteromesoscutum sculpturing punctate reticulate, reasonably regular but less dense and shallower in centre and on lateral sides; number of pits in scutoscutellar sulcus 12 (10); scutellar disc with only shallow pits associated with setae and no basal rugostiy; maximum height of mesoscutellum lunules/maximum height of lateral face of mesoscutellum 0.6 (0.5).
Propodeum. Medial carina distinct in holotype but can be faint or nearly undiscernible in paratypes, rest of propodeum mostly smooth, some scattered rugosity and small carinae in posterior centre, posterior lateral corners, sometimes with anterior transverse carinae.

Male
As female, only slight variations in measurements, T2 thinner and stronger sculptured with strong indented border.

Distribution
Currently only known from the type locality on the Mornington Peninsula, Victoria. The host species is found throughout Australia, suggesting the wasp likely also has a wider distribution range than currently known.

Remarks
All specimens of this species, including the holotype, are not complete specimens, but the presence of DNA barcodes and host data warranted the description of this species regardless of the incompleteness of the description (missing the flagellomere 14 measurements) and poor condition of the type series. Measurements given of paratypes in description n = 2.
Body length. Head to apex of metasoma: 2.7 mm.
Propodeum. Medial carina indistinct in holotype, but clearly visible in male paratypes suggesting that female specimens may also have variable carina, holotype propodeum strongly rugose with some areas leaning towards reticulate rugose, but sculpturing less strong in some male paratypes.

Distribution
Currently, this species is only recorded from arid South Australia, in the Lake Torrens and Great Victoria Desert regions.

Remarks
This species constitutes the BIN BOLD:ADL5542, which has a maximum intraspecific distance of 1.28%, and a 2.41% distance to the nearest neighbour. Austin & Allen, 1989 Fig . 31A

Remarks
Introduced to Australia in 1951 from Italy as a biological control agent against P. xylostella, but not considered well established or a significant parasitoid of this lepidopteran pest (Furlong et al. 2008). This species forms a distinct clade in the COI phylogeny ( Fig. 1)

Etymology
This species is named for the town Wonboyn, from where the type series was collected and reared by Paul Whitington and Kerri-Lee Harris. Paratypes AUSTRALIA • 3 ♀♀, 1 ♂; same collection data as for holotype; ANIC 32 130295 to 130298 • 1 ♀, 1 ♂; same collection data as for holotype; WINC.

Male
As female, only slight variations in measurements.

Distribution
Currently known only from type locality on the southern NSW coastal region, although as the host occurs throughout the east coast of Australia and into Tasmania, this parasitoid is likely to have a wider range than is currently known.

Diagnosis
This species, which is likely C. icipe Fernández-Triana & Fiaboe, 2017, can be separated from all other species of Cotesia currently described from Australia and Papua New Guinea by the sculptured area on the centre of the medial band of the scutellar disk and T1 parallel sided.

Distribution
In Australia, known from Queensland, the Northern Territory and South Australia -see remarks.

Host
No host data available for specimens from Australia.

Remarks
This species, although clearly different from any currently described from Australia, is morphologically and genetically (1-1.5% divergent in the COI barcoding region) extremely similar to C. icipe. Cotesia icipe was described from specimens reared in Africa from Spodoptera littoralis (Boisduva, 1833) and Spodoptera exigua (Hübner, 1808) (Lepidoptera, Noctuidae). Spodoptera exigua is present in Australia as an invasive pest on a wide range of plants (Hill 2014).
The BOLD BIN Cluster IDs for this species are BOLD:ADL3257 (South Australian specimens) and BOLD:ABZ7318 (Queensland and Northern Territory specimens). BIN BOLD:ABZ7318 contains the DNA barcodes for the described C. icipe species as well as other specimens from Australia, Madagascar, Africa, Pakistan, French Polynesia, Bangladesh, Saudi Arabia, Colombia and Malaysia, and is less than 2% divergent from BIN BOLD:ADL3257.
In the original description of C. icipe (Fiaboe et al. 2017), the authors present a COI phylogeny showing the closely related specimens on BOLD from elsewhere in the world, which cluster by geographic location, but take the conservative approach of excluding those sequences from the species. Considering that the host species are widespread pests of agricultural crops, we believe it is likely that C. icipe is also therefore now a widespread species, and that these closely related DNA barcodes from outside Africa, including those from Australia, likely represent specimens of C. icipe.
However, without host data for the Australia specimens, we are reluctant to definitively confirm the presence of C. icipe in Australia until further specimens with host data become available. We do however present here the specimens, images and DNA data in the hope that it can assist applied research on the parasitoids of Spodoptera exigua.

Discussion
In this study we describe seven new species of Cotesia and place them in the context of historically described native species and introduced taxa found in the country. Whilst this raises the number of Cotesia species found in Australia from 14 to 21 (22 including the species suspected to be C. icipe), Cotesia in Australia is expected to be much larger than currently documented. In the COI phylogeny presented in this study, an additional seven species are identified but not treated due to lack of specimens, and broader geographical sampling will undoubtedly increase this number substantially. The purpose of this study is to provide a solid framework for the genus in Australia from which future work can build, and to create an identification resource that will be useful to workers in applied entomology.
Future work should aim to obtain DNA barcodes, where possible, from described species that do not currently have genetic data available, as Cotesia is a difficult genus to identify conclusively using only morphological information. Whilst many of these species are represented by single or very few specimens, non-destructive future technologies may increase the success rate, and lower the risk, of obtaining DNA from type material. The resolution of Cotesia will benefit significantly from having this data for all described species.
European Journal of Taxonomy 667: 1-70 (2020) In this study we make the assumption that the newly described Australian species are endemic and new species are therefore not compared or diagnosed against species found elsewhere in the world, other than by comparing COI barcodes with those available on BOLD. We feel this is generally a realistic assumption based on the isolation of Australia and the high rate of endemism of the insect fauna (Austin et al. 2004;Chapman 2009), but note that cosmopolitan species are theoretically possible. However, their recognition needs to be based on more than just comparative morphology. See for example the unresolved synonymy of C. ruficrus with C. sydneyensis, and the case of the synonymy of A. nonagriae with C. flavipes, which is now known to be erroneous. Species distributions can be further complicated by the undocumented introduction of species in conjunction with their hosts, such as the possible presence of C. icipe in Australia.