In the checklist below, a total of 81 genera and 2,999 extant species are recognized as valid, including 36 nominal species that are currently considered to be species inquirendae.

Two genera are synonymized under Apanteles: Cecidobracon Kieffer & Jörgensen, 1910 syn. nov., and Holcapanteles Cameron, 1905 syn. nov. Nine lectotypes are designated. A total of 318 new combinations, three new replacement names, three species name amendments, and seven species status revised are proposed. Additionally, three species names are treated as nomina dubia, and 52 species names are considered to be unavailable (including 14 as nomina nuda), listed at the end of the checklist.

Extinct taxa, only known as fossils (three genera and 12 species) are listed in a separate section below (Table 3).

The pace of species description in Microgastrinae has been steadily increasing since the first species was described in 1758 and has shown no signs of slowing down (Fig. 1). The total number of genera has also increased substantially, especially since 1965; the information is summarized in Whitfield et al. (2018), Fernandez-Triana and Boudreault (2018), and below.

Figure 1. 

Microgastrinae species described since 1758 based on data in present paper A Total numbers per decade B Cumulative number (1758–2019).

Primary types of Microgastrinae are deposited in 108 institutions worldwide, although 76% of those types are concentrated in seventeen collections (Table 1), seven of which have more than 100 primary types each. Localities of primary types are reported from 138 different countries.

Microgastrine wasps have been recorded in most countries and all continents except Antarctica. Only 16 countries do not yet have any recorded species of Microgastrinae: Bahrain, Botswana, Bhutan, Cambodia, Djibouti, Equatorial Guinea, Gabon, Gambia, Guinea, Guinea-Bissau, Kuwait, Laos, Liberia, Mauritania, Qatar, and Swaziland. This is of course just an artifact of insufficient collecting and/or lack of studies in those countries; each is expected to harbour many species.

The current data (Table 2) show two countries with 400+ Microgastrinae species each (China with 448 and Costa Rica with 427), another two with 300+ species each (Russia with 388 and Hungary with 327) and five with 200+ species each (USA, Germany, India, United Kingdom, and Canada). Overall, 34 countries have more than 100 described species recorded, although those numbers can be misleading. For example, the reason Hungary ranks so high is because of extensive studies in that country, done over many years by Jenö Papp while working in the Hungarian Natural History Museum. A similar situation applies to both the United Kingdom and Germany, where a long European tradition of experts on the group coupled with extensive collecting have provided figures that are close to the actual diversity in those countries. While the microgastrine fauna of those three countries is relatively well known, the opposite occurs in large and/or mostly tropical countries, where more species are still undescribed. For example, in Costa Rica, DNA barcoding has already identified more than 1,200 species just in ACG (Janzen and Hallwachs 2016). And the figures for China and India (which are considered to be “megadiverse countries”, sensu Myers et al. 2000), are still very far from being complete, as both countries should easily reach more than 1,000 species each. Other megadiverse countries such as Australia, Brazil, Colombia, Democratic Republic of Congo, Indonesia, Madagascar, Mexico, Peru, Malaysia, Papua New Guinea and USA are all likely to have similar (in some cases higher) totals, but studies thus far have been insufficient, leading to most of those countries having “only” a hundred species or fewer recorded at present.

There are three main limitations in our paper that we want to point out. The first relates to the coverage of primary types in our study. We were able to examine primary types for 1,394 species (46.5%), and for another 1,568 species (52.3%) we studied authenticated specimens, checked original descriptions, or read subsequent revisions. However, for 37 species (1.2%) we could not check any source of information, or it was considered inadequate, and they are left in the genus in which they were originally described (or as species inquirendae), with explanatory annotations. In future versions, we aim to increase the number of species for which we have examined primary types, but for the present paper the reader must consider the relatively large number of species still needing to be thoroughly studied. It is especially important to keep in mind that for some of those species for which we could only study descriptions (which may not be detailed or clear enough), the generic placement made in this paper might be incorrect.

A second limitation is the coverage of references concerning Microgastrinae. In the References section we tried to list all papers where original descriptions of Microgastrinae were published (those references in turn are cited under the corresponding treatment of every species in the checklist below). However, our list is not complete and we are aware of some omissions; in that sense, the latest versions of Taxapad (Yu et al. 2012, 2016) have more comprehensive lists of references than our paper. Especially important and comprehensive is Yu et al. (2016), which lists 6,200+ references related to Microgastrinae.

A third limitation of our paper is that we do not treat host records in detail. We expect to present host data for microgastrine species with verified information in a subsequent version of the world checklist, although it is improbable that we will be able to comment with reliability on all published records. The latest versions of Taxapad (Yu et al. 2012, 2016) provide the best coverage of references on hosts of Microgastrinae; however, that is only an uncritical compilation of literature, and many of those references report incorrect data. The reader is strongly advised to double check host references and be very cautious in interpreting information from secondary sources.

Extinct genera and species of Microgastrinae have been found in Eocene and Oligocene deposits, from 37–44 million years ago (MYA). Many specimens from the Miocene (20–30 MYA) are known from Dominican and Chiapas ambers, but most appear to be undescribed representatives of extant genera (Murphy et al. 2008). Belokobylskij (2014) revised the taxonomic status of all previously known taxa of fossil Microgastrinae and described one new genus as well as two new species. The origin of Microgastrinae has been estimated at ~ 54 MYA (Murphy et al. 2008).

Unlike previous work (Mason 1981, Yu et al. 2005, 2012, 2016), we exclude fossil genera or species from our world checklist. Instead, we tabulate in this section the three genera and 12 species of fossil Microgastrinae currently described (Table 3).

Microgastrinae was originally described at family rank, as ‘Microgasteroidae’, by Foerster (1863). At that time, it comprised only three genera: Microgaster Latreille, 1804 (the genus that provides the root for the subfamily name, meaning “small abdomen”, in reference to the relatively short length of the metasoma compared to other Braconidae), as well as two genera described by Foerster (1863): Microplitis (which means “small sword” or “small weapon”, referring to the generally short ovipositor in that genus) and Apanteles (meaning “incomplete”, in reference to the fore wing lacking the second intercubitus, leaving the second submarginal cell open or incomplete). Fornicia, although described by Brullé (1846) before Foerster’s work, was at the time considered to belong to other subfamilies in Braconidae (e.g., Dalla Torre (1898) placed the genus in Cheloninae; Ashmead (1900a) placed it in Sigalphinae; Granger (1949) placed it in Triaspidinae), and it was not recognized to be part of Microgastrinae until a century later (Baltazar 1962, Nixon 1965).

The high diversity of Microgastrinae quickly became evident, and so attempts to split the group into further genera started shortly after Foerster’s (1863) paper, e.g., by Reinhard (1880). Many additional genera (15 recognized in this paper) were described between 1882 and 1958, although some were not associated with Microgastrinae at the time, and others were not accepted as valid genera by some authors of the period, e.g., Muesebeck (1921) and Telenga (1955).

This view changed with two seminal works in 1965 and 1981. Nixon (1965) reclassified the subfamily limits and provided some structure to what was being recognized as a huge assemblage of parasitoids of Lepidoptera. He recognized 20 genera, eight of which were new, and reclassified the species within Apanteles sensu lato into a more practical and useful array of 44 species groups to facilitate identification. Mason (1981) fundamentally changed the taxonomy of Microgastrinae by recognizing 50 genera (23 of which he described as new), including numerous taxa that mostly corresponded to particular species groups of Nixon (1965, 1973), and additionally proposing new combinations for some 350 species.

Since Mason (1981) 32 genera have been described. Whitfield et al. (2018: fig. 2) graphically showed the increase in description of new genera during the past 150 years. Nevertheless, there are still many more genera of Microgastrinae that remain to be described, e.g., Fernandez-Triana and Boudreault (2018). Additionally, several genera, as currently understood, are probably polyphyletic and need to be split, e.g., Diolcogaster and Glyptapanteles. A comprehensive phylogenetic analysis of the subfamily is needed before we can achieve a clearer picture. However, just based on the material we have seen in collections, we estimate that the Microgastrinae is likely to comprise close to one hundred genera.

For the past few years the main problem with the generic concepts is that two different classifications of Microgastrinae have been proposed and are widely used: those based on Mason (1981) and on van Achterberg (2003). For a visual depiction of how the two classifications differ (based on the number of species assigned to each of the most speciose genera), see Figure 2.

Figure 2. 

Number of extant species per larger genera of Microgastrinae A Data from Taxapad 2016, which is mostly an update, with slight modifications, of van Achterberg (2003), total number of species: 2,710 B Data from present paper, which is mostly based on Mason (1981) but extensively updated, total number of species: 2,999.

The classification proposed by Mason (1981) had a narrower concept of Apanteles and Protapanteles, which resulted in a larger number of Microgastrinae genera treated as valid. Many of the new combinations resulting from that classification are in Mason (1981), although not all species have been properly transferred to the corresponding genus. Mason’s system has been followed by most researchers (see examples cited in the Introduction) and has remained largely stable for the past 38+ years, with a few exceptions: his genus Teremys was synonymized under Pholetesor (Whitfield 2006); and his arrangement of genera within tribes, largely based on phylogenetic grounds, has not been universally accepted (Austin 1990, Austin and Dangerfield 1992, Whitfield 1995a, Fernandez-Triana 2010; see also Walker et al. 1990 for further criticism of tribes within Microgastrinae). Mason (1981) based his paper on studies of the world fauna; however, a careful examination of the CNC collection (Mason’s base) and other material available to him at the time shows that specimens from the Afrotropical, Oriental, and Australasian regions were much more poorly represented than the remaining regions. Thus, most of the new genera from those regions described by Mason (1981) have later been found to have a wider distribution and greater morphological variation than originally thought, and some of those genera will need redefinition. Another consequence of the limited geographical coverage of the studied specimens is that the keys to tribes and genera in Mason (1981) work reasonably well for the temperate areas, but not as well for the tropical areas, especially the Old World tropics.

The classification proposed by van Achterberg (2003) reduced the number of genera by treating eleven genera recognised by Mason as synonyms or subgenera of Apanteles and Protapanteles. That system was later implemented in Taxapad (Yu et al. 2005, 2012, 2016) and other, mostly European, databases, e.g., Fauna Europaea (https://fauna-eu.org/) and Dyntaxa (https://www.dyntaxa.se/). Shortcomings of this approach have been pointed out by other authors, e.g., Broad et al. (2016) and Whitfield et al. (2018). The main issue with van Achterberg’s approach is that his classification was based mainly on the European species, a region with relatively little diversity in genera and species (see sections below), and is thus clearly insufficient to capture the rich fauna of Microgastrinae worldwide. Second, and more worrisome, van Achterberg’s generic concepts were applied in Taxapad to the entire world fauna, effectively producing numerous (perhaps hundreds) of new name combinations which have never been formally published, let alone critically assessed. The validity of those names may be questionable, but van Achterberg’s classification has been embraced uncritically by some users of Taxapad.

To complicate things further, generic concepts changed slightly in Taxapad from the 2012 to the 2016 version (Table 4). For example, Taxapad 2016 considers some taxa as subgenera that the 2012 version had listed as synonyms of Apanteles (Dolichogenidea, Exoryza, Iconella, Illidops, and Pholetesor) or as synonyms of Protapanteles (Nyereria, Rasivalva, Sathon, and Venanides). Other genera were treated differently, e.g., Distatrix is treated as a valid genus in the 2012 version but as a subgenus of Protapanteles in 2016, and Glyptapanteles is a synonym of Protapanteles in 2012 but a valid genus in 2016. Some of those decisions may have merit, but three are highly questionable:

a) Rasivalva should never have been considered to be part of Protapanteles as it has a complete areolet in the fore wing (a character not present in any Protapanteles or related genera);

b) Ectadiophatnus is listed as a genus of Microgastrinae in both the 2012 and 2016 versions, following Shenefelt (1973), despite having been published as belonging to the subfamily Blacinae since at least 1935 (Ferrière 1935, Mani 1938, Varshney 1976, Mason 1981) [van Achterberg (pers. comm.) has examined the type species and found that it is a new synonym of Eubazus Nees, in Brachistinae-Brachistini];

c) the species listed under Lissogaster have since 1988 been transferred back to Microgaster (see more details about that in Mason (1986) and in the checklist below, in the introductory comments to the genus Microgaster).

The rationale for the changes between versions of Taxapad is not always evident and, as far as we are aware, has never been explained in a published paper. As a result, it is difficult to follow the different arrangements of genera and subgenera, a problem which is further compounded by the use of tribes in the 2012 version, while the 2016 version added sub-tribes (Table 4).

We believe that the classification proposed by Mason (1981), although not entirely free from problems and shortcomings, provides the best framework currently available to deal with the world diversity of Microgastrinae and provides a solid and clear foundation from which to work towards future improvements. In this paper we largely follow that system, except for dividing the subfamily into tribes, as we do not think the tribes proposed by Mason properly reflect the phylogenetic relationships within the subfamily. We here classify the world species in 81 genera of Microgastrinae (Table 4 and checklist below).

The last two published keys to world genera of Microgastrinae were in Nixon (1965) and Mason (1981). Nixon (1965) recognized 19 genera in his key, whereas Mason (1981) included 50 genera (although Mason’s paper started with a key to tribes, and then genera within each tribe are keyed out and treated separately). Some regional generic keys have been published since, e.g., Tobias (1986) for the former Soviet Union, Austin and Dangerfield (1992) for the Australasian region, Whitfield (1997) for the New World, Chen and Song (2004) for China, and Kotenko (2007a) for the Russian Far East. However, with 81 genera considered in this paper, the information to recognize them in the aforementioned references is clearly outdated, and an updated key to world genera is badly needed.

Unfortunately, we still lack a robust phylogeny for the subfamily, which would be needed to provide a useful and comprehensive key. The limits of some genera at present are not well defined, and at times are contradictory; moreover, it is likely that future work will change many groups as currently understood. We anticipate that a few genera will end up as synonyms while several others, which are paraphyletic or polyphyletic as currently defined, will be split. This should likely result in an overall increase in the total number of genera as compared to present (e.g., see Fernandez-Triana and Boudreault 2018).

We divide the 81 genera recognized in this paper into four groups and characterize each group and singular genus with brief morphological diagnoses. We emphasize that these groups are not to be considered as monophyletic, and we caution that the discussion below is not to be taken as a new phylogeny for the subfamily, which is beyond the scope of the present paper. We do not present the information below as a surrogate key either; to key out Microgastrinae genera the reader is advised to initially consider the works mentioned at the beginning of this section. Our only intention here is to provide the reader with some basic information on the concepts we have followed when making decisions about generic placement of species, especially in the new combinations we propose in the checklist below. Besides comments on morphological diagnoses, we also provide illustrations for every Microgastrinae genus (at least one species per genus, usually more), the first time that has been done for the entire subfamily.

We separate Microgastrinae into four broadly defined groups:

a) unplaced genera, all of which have unique morphological characters that make them very distinctive, although they do not share any character in common per se, comprising 18 genera: Austinicotesia, Austrocotesia, Beyarslania, Billmasonius, Clarkinella, Exulonyx, Fornicia, Janhalacaste, Kiwigaster, Mariapanteles, Miropotes, Neoclarkinella, Pelicope, Prasmodon, Qrocodiledundee, Semionis, Xanthomicrogaster, and Zachterbergius;

b) Microplitis group, which includes the Microplitini (sensu Mason 1981) and four additional genera described by Fernandez-Triana and Boudreault (2018), for a total of eight genera: Alloplitis, Gilbertnixonius, Jenopappius, Microplitis, Philoplitis, Silvaspinosus, Snellenius, and Tobleronius;

c) Cotesia group, which includes most but not all of the Cotesiini (sensu Mason 1981), with 29 genera: Buluka, Carlmuesebeckius, Chaoa, Cotesia, Cuneogaster, Deuterixys, Diolcogaster, Distatrix, Eripnopelta, Exix, Glyptapanteles, Jimwhitfieldius, Keylimepie, Larissimus, Lathrapanteles, Markshawius, Nyereria, Ohenri, Parenion, Protapanteles, Protomicroplitis, Pseudofornicia, Pseudovenanides, Rasivalva, Sathon, Ungunicus, Venanides, Venanus, and Wilkinsonellus;

d) Apanteles group, which includes most but not all of the Apantelini + Microgastrini (sensu Mason 1981) with 26 genera: Agupta, Alphomelon, Apanteles, Choeras, Dasylagon, Dodogaster, Dolichogenidea, Exoryza, Hygroplitis, Hypomicrogaster, Iconella, Illidops, Kotenkosius, Microgaster, Napamus, Papanteles, Parapanteles, Paroplitis, Pholetesor, Promicrogaster, Pseudapanteles, Rhygoplitis, Sendaphne, Shireplitis, Xanthapanteles, and Ypsilonigaster.

a) Unplaced genera

Kiwigaster (Figs 136–137) is the only genus of Microgastrinae with sexual dimorphism in the number of antennal segments; females have 17 flagellomeres and males have 18 (Fernandez-Triana et al. 2011). All other known microgastrines have 16 flagellomeres in both sexes.

Only five genera of Microgastrinae, Austinicotesia, Austrocotesia, Miropotes, Pelicope, and Semionis, have hind wings without vein 2r-m (all other known Microgastrinae have that vein present, although often weakly pigmented).

Pelicope and Semionis can be recognized within this group because both have the fore wing areolet very large (while the other three genera are without an areolet or have a very small areolet). Pelicope (Fig. 181) has the propodeum unsculptured, notauli at least partially marked, and eyes in frontal view slightly divergent ventrally. Semionis (Figs 221, 222) has the propodeum with a partial transverse carina and many fine striations radiating from the nucha, the notauli not marked, and the eyes in frontal view are not divergent ventrally (Nixon 1965, Mason 1981).

Miropotes (Figs 157–159) differs from the other genera by the ovipositor sheaths and ovipositor with a unique shape, in most species strongly bent; eyes enlarged and strongly convergent with malar space totally or almost totally obliterated; metacoxa small and metatibial spurs very short (Fernandez-Triana et al. 2014d).

Austinicotesia (Figs 27, 28) and Austrocotesia (Figs 29–32) are similar to each other in several features (Austin and Dangerfield 1992, Fernandez-Triana and Boudreault 2018) but differ as follows: Austinicotesia has the fore wing without areolet (with areolet in Austrocotesia); fore wing with pterostigma relatively thin and long, 3.5 × as long as wide (pterostigma much less than 3.0 × as long as wide in Austrocotesia); fore wing vein 2RS much longer, ca. 1.5 ×, than vein r (fore wing vein 2RS much shorter, ca. 0.5 ×, than vein r in Austrocotesia); metafemur relatively thick and stout (of more normal proportions in Austrocotesia); T1 widening towards posterior margin and with strong hump followed by deeply excavated area and strong carinae (T1 more or less parallel-sided or narrowing towards posterior margin and without hump or excavate area in Austrocotesia); and T2 mostly smooth (usually mostly sculptured in Austrocotesia).

Only six genera of Microgastrinae have the propodeum mostly smooth except for complete longitudinal and transverse carinae: Beyarslania, Clarkinella, Janhalacaste, Neoclarkinella, Mariapanteles, and Prasmodon. We place them together because of the diagnostic value of that unique carination pattern, but it is clear that these genera do not constitute a monophyletic group.

Prasmodon (Figs 191–193) is the only genus in this subgroup with notauli strongly marked and fore wing areolet relatively large (Fernandez-Triana et al. 2014f).

Clarkinella and Janhalacaste also have a fore wing areolet (although very small, almost obliterated) and can be distinguished from each other as follows. Clarkinella (Figs 46, 47) has the scutellar disc with a smooth posteromedian band, T1 without a median longitudinal carina, and hypopygium mostly inflexible with only a sharp fold posteriorly (Mason 1981), whereas Janhalacaste (Figs 128, 129) has the scutellar disc with a coarse posteromedian band, T1 with a longitudinal sulcus on the anterior 0.6–0.7 of its length and posterior 0.3 with two short carinae centrally delimiting a slightly raised area, and hypopygium folded medially and with several pleats (Fernandez-Triana and Boudreault 2018).

Neoclarkinella (Figs 161–165), Mariapanteles, and Beyarslania all lack a fore wing areolet. Neoclarkinella can be recognized because it has a very distinctive T1 which sharply narrows towards posterior margin and has a wide depression on the anterior half, and a hypopygium with multiple pleats (Chen and Song 2004, Veena et al. 2014).

Mariapanteles and Beyarslania have the hypopygium mostly inflexible, with a posteromedian translucent fold where only a few or no pleats are visible; and T1 has a sharply defined median, longitudinal sulcus, at least on the anterior half. Mariapanteles (Figs 143, 144) has the ovipositor sheaths much longer (0.7 × as long as the metatibia length), and ovipositor mostly straight to slightly curved (Whitfield et al. 2012), whereas Beyarslania (Fig. 33) has the ovipositor sheaths relatively very short (less than 0.3 × metatibial length), and the ovipositor strongly downcurved (Mason 1981, at the time referring to the genus as Xenogaster). Mariapanteles is also the only genus in this group with the propodeum having some additional, small and short transverse carinae that radiate from the median carina (but, nevertheless, the propodeum still appears as if it is crossed by the median and transverse carinae, the defining trait of this group).

The remaining six genera in this group cannot easily be associated with any other genus and are discussed below in alphabetical order.

Billmasonius (Fig. 34) is recognized by T1 with a unique shape and desclerotization, with a relatively wide anterior 0.6 and very narrow posterior 0.4, so that widest part of tergite, near anterior margin, is around 4.0 × the narrowest width, along posterior 0.4, and with anterior 0.6 mostly desclerotized, only with lateral margins and narrow central strip sclerotized; T2 is also diagnostic, with a partially sclerotized area surrounding each spiracle on laterotergite 2 the same colour as T2, giving the impression of T2 having three peaks, the largest and central one being the actual T2, the two smaller lateral ones being the area surrounding the spiracles on each laterotergite (Fernandez-Triana and Boudreault 2018).

Exulonyx (Fig. 95) has a unique combination of features within Microgastrinae: propodeum with a partial median, longitudinal carina on anterior 0.6 and complete areola on posterior 0.4, hypopygium inflexible, ovipositor curving downwards on posterior 0.3, and T1 and T2 coarsely sculptured (Mason 1981).

Fornicia (Figs 96–98) is the only Microgastrinae genus with the epicnemial carina complete and the fore wing areolet absent; also, the head in lateral view is relatively small (compared to the mesosoma) (Austin and Dangerfield 1992), and T1–T3 form a carapace covering the entire dorsal surface of the metasoma. Only a few species in the Microplitis group (see below) have a partial to complete epicnemial carina, but all those genera have the fore wing with an areolet (usually relatively large), and the head of normal proportions.

Qrocodiledundee (Fig. 212) can easily be recognized by its propodeal apophysis, unique among Microgastrinae, as well as the flattened mesosoma, metafemur short and stout, pronotum dorsally enlarged, and the propodeum with a median carina and a partially defined areola (Fernandez-Triana and Boudreault 2018).

Xanthomicrogaster (Figs 246–249) is unique because of the following combination of features: hind wing with vein 1cu-a strongly sinuous and first submarginal cell tall (height at least 2 × its width), fore wing with a very small areolet, metacoxa very large (almost as large as the metasoma length), propodeum mostly smooth but with a strong and sharp median longitudinal carina, T1 very wide and with a strong median longitudinal sulcus, T2 rectangular and usually sculptured, hypopygium inflexible, and ovipositor sheaths relatively long (more than 0.5 × metatibia length) and with numerous setae. Some of these morphological features would suggest this genus could be placed within the Cotesia group, contrary to Mason’s (1981) opinion when he grouped it within his Microgastrini. However, Xanthomicrogaster has many other features that are so different to both our Cotesia group and Mason’s Microgastrini that we prefer to maintain it as an unplaced genus.

Zachterbergius (Figs 253, 254) has the longest and thinnest T2 among all known Microgastrinae, with T2 length 4.0 × its width at base and apex, 0.7–0.8 × as long as T1 length, and around 1.5× as long as T3 length. Also, the propodeum has a clearly defined median carina, partially defined transverse carina, and the posterior part of an areola; the antennal scape is very transverse, and the labial palpi are very long, extending to the mesopleuron (Fernandez-Triana and Boudreault 2018).

b) Microplitis group

This is one of the best-defined groups of genera within Microgastrinae (see Mason 1918), and most likely to be monophyletic. It is characterized by: tentorial pits relatively large, head mostly coarsely sculptured, stemmaticum usually very well defined and slightly to strongly raised from the surrounding areas, anteromesoscutum and scutellar disc usually coarsely sculptured, notauli almost always defined (often very clearly), propodeum always sculptured and with several strongly defined carinae, fore wing with areolet usually large, metacoxa relatively small, metatibial spurs short, T1 with median longitudinal sulcus, hypopygium inflexible and almost always relatively short, ovipositor sheaths with few setae that are mostly limited to the apex, and ovipositor almost always very short (much shorter than 0.5 × metatibia length).

Philoplitis (Figs 182, 183) has a unique combination of features including an enormous scutellum conically prolonged posteriorly over the propodeum (Mason 1981, Fernandez-Triana and Goulet 2009, Ranjith et al. 2019). It also has an occipital carina, and ocelli forming a very low triangle, to the point that the anterior ocellus seems almost on the same line as the posterior ones.

Silvaspinosus (Fig. 227) has the clypeus extremely long and thin, the malar line extremely short (almost absent), the mandible base separated from the rest of the head by a desclerotized area that looks almost like an opening, and mandibles relatively stout and large. The shape of the clypeus, and the separation of the mandible from the rest of the head by a desclerotized area are unique among Microgastrinae (Fernandez-Triana and Boudreault 2018). It also has the fore tarsus with a spine-like seta, and the scutellar disc with the posteromedian band smooth; both of which are unique and distinctive among the Microplitis group.

Gilbertnixonius (Fig. 99), is the only genus in this group that has the propodeum with both longitudinal and transverse carinae but without an areola (Alloplitis and Tobleronius have those carinae, although sometimes incomplete, but they also have a complete areola on the propodeum). Gilbertnixonius also has an epicnemial carina (otherwise only present in some species of Snellenius and in all species of the unrelated genus Fornicia) and an incomplete occipital carina (otherwise only present in Alloplitis, Philoplitis, and Tobleronius) (Fernandez-Triana and Boudreault 2018).

Alloplitis and Tobleronius are somewhat similar morphologically and distinguished from the other six genera in this group by the propodeum with a complete areola (in addition to partial longitudinal and transverse carinae). Alloplitis (Figs 7, 8) has T1 more or less parallel-sided or slightly widening towards the posterior margin, and T2 more or less rectangular; whereas Tobleronius (Fig. 233) has T1 strongly narrowing towards the posterior margin (width at posterior margin 0.3 × or less of width at anterior margin) and T2 very long and thin (although slightly widening towards the posterior margin) and with the area surrounding the spiracles on laterotergite 2 partially sclerotized and the same colour as T2 giving the impression of T2 having three peaks, the largest and central one being the actual T2, the two smaller lateral ones being the area surrounding the spiracle on each laterotergite (Fernandez-Triana and Boudreault 2018).

Microplitis (Figs 151–156) and Snellenius (Figs 228–232) are very similar and form one of the most morphologically distinct groups of Microgastrinae (Nixon 1965, Mason 1981, Walker et al. 1990, Shaw and Huddleston 1991, Austin and Dangerfield 1993, Fernandez-Triana et al. 2015b) with the following shared diagnostic features: propodeum with coarse sculpturing and a strong median carina and T2 and T3 with a poorly defined separation between them. Most species of Snellenius are easily distinguished by having the notauli and the scutellar disc strongly excavated and sculptured, and by having the scutoscutellar sulcus very wide and deep; both cases represent the most extreme examples within Microgastrinae. Additionally, the propodeum is divided into two distinct areas (faces) clearly marked by a strong angulation (observed in lateral view) and a transverse carina (observed in dorsal view). The main difficulty when trying to distinguish both genera is that those features appear to grade, from strongly excavated and sculptured notauli and scutellar disc (most Snellenius) to less excavated and less sculptured (a few Snellenius, most Microplitis), to basically smooth and unexcavated (some Microplitis). The only reliable feature to separate the two genera is the presence of an epicnemial carina in Snellenius, which is absent in Microplitis (Mason 1981, Austin and Dangerfield 1992, 1993, Fernandez-Triana et al. 2015b), although in practice it may be difficult to distinguish the epicnemial carina due to setae and/or sculpture on the epicnemium and mesopleuron.

Jenopappius (Figs 130–131) resembles Microplitis but with T2 strongly sculptured and rectangular, and T1 mostly sculptured and with a median depression anteriorly. Some Alloplitis may also have a somewhat similar sculpture on either T1 or T2 but the shape of those tergites is very different, and Alloplitis always has the propodeum with a complete areola, defined by strongly raised carinae. The combination of the sculptured propodeum without an areola, T1 with an anteromedian depression, and T1 and T2 with strong sculpture are very unusual and will separate Jenopappius from any other genus of Microgastrinae (Fernandez-Triana and Boudreault 2018).

c) Cotesia group

We place here genera with a completely inflexible hypopygium, ovipositor sheaths relatively short (less than 0.5 × metatibial length, usually much less) and mostly without setae (except apically in some cases). Most of the 29 genera considered here also have the propodeum without a complete areola (although some have it, and others have a complex arrangement of carinae and sculpture where a partial to complete areola can sometimes be defined). Although these features work well to recognize most members of the group, a few species of Sathon, Lathrapanteles, Glyptapanteles, and Ohenri have relatively long ovipositor sheaths, but in these cases the hypopygium is still always inflexible. Most or perhaps all the species within the Cotesia group posses a suite of characters indicative of parasitism of “macrolepidoptera” (sensu Mason 1981: 25), but the group is probably not monophyletic. From the Cotesiini (sensu Mason 1981) we exclude here Parapanteles and instead transfer it to the Apanteles group (see details under that group); the main reason being that this genus, as it had been understood, apparently includes two different sets of taxa: one that seems to be Cotesia species misidentified as Parapanteles (Valerio et al. 2009, Parks 2018, Freitas et al. 2019), and another (representing the majority of the genus, as currently understood, including the type species) that are more related to Dolichogenidea and Apanteles than to any genus in the Cotesia group. We also add here Sathon, which we consider to be closer to Glyptapanteles and related genera, unlike Mason (1981), who considered it to be part of his Microgastrini group.

The Cotesia group can be broadly split into two subgroups, based on whether the fore wing has an areolet (Buluka, Cuneogaster, Diolcogaster, Eripnopelta, Exix, Jimwhitfieldius, Keylimepie, Larissimus, Markshawius, Parenion, Protomicroplitis, Rasivalva, Ungunicus, Venanus) or does not have an areolet (Carlmuesebeckius, Chaoa, Cotesia, Deuterixys, Distatrix, Glyptapanteles, Lathrapanteles, Nyereria, Ohenri, Protapanteles, Pseudofornicia, Pseudovenanides, Sathon, Venanides, Wilkinsonellus).

Among the genera with a fore wing areolet, Jimwhitfieldius (Figs 132, 133) has the metatrochantellus with a unique shape (Fig. 133), the head with a strong depression behind the occiput, the metatibia with a very long and thick inner spur, and the ovipositor and ovipositor sheaths extremely short, probably the shortest in the entire subfamily (Fernandez-Triana and Boudreault 2018).

Venanus (Figs 237–240) is quite distinctive, and comprises small species, often with the body slightly depressed, face with a triangular flange between the antennal sockets, fore wing with a relatively large areolet, T2 with strongly defined lateral sulci, and ovipositor sheaths with very few and minute setae (Mason 1981).

The remaining genera in the subgroup seem to share one or several morphological features with Diolcogaster (whether those features are homoplastic or not). Diolcogaster (Figs 66–77), as currently understood, is most likely a polyphyletic genus that will need to be split into several genera. Until then, it is difficult to define unequivocally. Instead, we discuss the remaining genera in this subgroup in alphabetical order, with the features that distinguish them from Diolcogaster.

Buluka (Figs 35–37) has T1–T3 forming a carapace and occupying the entire dorsal surface of the metasoma, the fore wing has a complete areolet, and females have part of the ventral surface of the distal six or seven flagellomeres without longitudinal placodes, instead having an oblique groove bounded on one side by a row of bent-tipped sensilla (Austin 1989). The carapace is shared with Fornicia and very few species of other genera, e.g., Deuterixys, Pholetesor, none of which have a fore wing areolet. The basimacula species group of Diolcogaster (sensu Saeed et al. 1999) have both the carapace and areolet, but the antenna does not have the special groove and sensilla.

Cuneogaster (Figs 61, 62) resembles Diolcogaster but it has the glossa long and apically bilobed, T1 wedge-shaped, and the scutellar disc with the medioposterior band smooth (Choi and Whitfield 2006) whereas in Diolcogaster the glossa is not elongated, T1 is usually not wedge-shaped, and the scutellar disc has a medioposterior band of rugosity in most species.

Eripnopelta (Figs 87, 88) could be considered an atypical Diolcogaster, but the pronotal lateral surface does not have distinct furrows, the scutellar disc has a smooth and protruding medioposterior band, T1 does not have a distinct median groove on the basal half, and the fore wing areolet is very small, almost obliterated (Xiong et al. 2017).

Exix (Figs 89, 90) also seems morphologically related to Diolcogaster, but it is defined by T2 large and smooth, without submedian grooves, the hind wing has the vannal lobe concave and lacking setae, and the hind wing nervellus is externally concave (Mason 1981).

Keylimepie (Figs 134, 135) can be recognized by the reduced wings in females, relatively small eyes and long malar space. The shape and sculpture of the head, mesosoma sculpture, shape and sculpture of T2, and ovipositor are all similar to some Diolcogaster, but Keylimepie has a T1 without a median sulcus and instead it has the anterior 0.5 rather depressed and concave, and the posterior 0.5 with strong transversal striations (Fernandez-Triana and Boudreault 2016).

Larissimus (Figs 139, 140) is another genus related to Diolcogaster but it can be recognized by the greatly reduced vannal lobe in the hind wing with, almost entirely smooth body, and the only described species is the largest known species of Microgastrinae, with a body and fore wing length of 7–8 mm (Nixon 1965, Mason 1981).

Markshawius (Figs 145, 146) has a unique set of features (Fernandez-Triana and Boudreault 2018) which together are very distinctive (although some, but not all, are shared with other genera). The female head is elongated and strongly concave posteriorly, modified to be tightly appressed to the anterior margin of the pronotum (following its contour); the face has its upper margin produced dorsally between the antennal insertions into a triangular flange; the frons is very elongated, with ocelli clearly much higher than normal; the antenna is very short (much shorter than body length, usually shorter than the combined length of the head and mesosoma), with all flagellomeres except the first having a single row of placodes; the propodeum has a median carina (defined posteriorly) and transverse rugosity which includes a poorly and partially defined transverse carina; and T1 is either extremely long and thin, with length at least 6.0× its width centrally, or very thin on the anterior 0.3–0.4, then strongly widening posteriorly, its width at the posterior margin around 3.0 × its width centrally.

Parenion (Figs 176, 177) can only be confused with some Diolcogaster, but is distinguished by having T2 and T3 smooth and barely or not separated, scutellar disc with the medioposterior band smooth and very small lunules on its lateral surface (Mason 1981).

Protomicroplitis (Figs 201, 202) is closely related to Diolcogaster, both morphologically and molecularly, and some of the criteria used to define it may need revision. The genus is defined by some flagellomeres having three rows of placodes, relatively large fore wing areolet, and T1 very long and narrow (Mason 1981, Fernandez-Triana 2015), although the last two features are also present in a few Diolcogaster species.

Rasivalva (Figs 213, 214) is characterized by the ovipositor sheaths lacking setae, or with very few and minute setae (Mason 1981, Chen and Song 2004, Kotenko 2007b). This separates it from Diolcogaster, which has relatively long setae on the ovipositor sheaths, including a few strong and thickened setae in many species. Other distinguishing features that appear in some species are the scutellar disc with the medioposterior band smooth, body sculpture smoother overall than in Diolcogaster, and propodeum with a median, longitudinal carina that is sometimes reduced or absent.

Ungunicus (Fig. 234) has remarkable and very distinctive tarsal claws, with a very large basal tooth longer than the apex of the tarsal claw, and a median lobe with setae arising from its margin, which seems slightly bilobate. These claws are unique within Microgastrinae (Fernandez-Triana and Boudreault 2018).

Among the genera without the fore wing areolet, Chaoa (Fig. 39) was described from a single specimen (Luo et al. 2004), with little information provided. Based on the original description and illustrations of the holotype, this genus might just represent a species of Glyptapanteles, or perhaps Nyereria but without examining the type we cannot conclude and therefore retain it as a valid genus for the time being.

Carlmuesebeckius (Fig. 38) has the ovipositor and ovipositor sheaths relatively long, and the propodeum with a complete areola, unlike most other genera in this subgroup. Other unique features are T1 with a strong and raised median carina for most of its length, and the ovipositor bulging near apex and with two subapical serrate teeth on the lower (first) valvulae (Fernandez-Triana and Boudreault 2018).

Cotesia (Figs 48–60) is a relatively uniform genus morphologically, long considered the easiest group to recognize among all segregates from Apanteles sensu lato (Mason 1981: 113). Defining characters are: fore wing without areolet; T1 and T2 usually mostly to entirely sculptured, T3 also often at least partially sculptured or, more rarely, completely sculptured; T1 either widening towards its posterior margin (very often), more or less parallel-sided or barrel-shaped (often), slightly widening towards the posterior 0.7–0.8 of the tergite length and from that point slightly narrowing towards the posterior margin which is more or less rounded (rarely), or medially constricted (extremely rare), but never completely narrowing towards the posterior margin; ovipositor and ovipositor sheaths are very short to short, very rarely moderately long. The propodeum varies considerably but has a well defined median longitudinal carina (very often), although the median carina may be difficult to distinguish on its own in species with the propodeum strongly sculptured with an irregular pattern of carinae (often), or the median carina may be partially absent (rarely), or the median carina may be combined with a partial to complete areola partially defined by a transverse carina (rarely), or the median carina is absent and/or the propodeal surface is shiny overall and almost without any sculpture (rarely). The only other genus that could be confused here would be Protapanteles, which may eventually be considered as just a species group within Cotesia, with smoother propodeum and T1–T3.

Protapanteles (Figs 198–200) usually has T1 either slightly widening towards the posterior 0.7–0.8 of the tergite length and then slightly narrowing towards the posterior margin which is more or less rounded (often), more or less parallel-sided or barrel-shaped (rarely), or slightly widening towards the posterior margin (rarely). The propodeum is variously sculptured, usually having a median longitudinal carina that may be partially or completely defined, and rarely lacking the median carina. A character commonly used to define this genus, a modified spine on the fore tarsus (Nixon 1965, 1972, 1973, 1976, Mason 1981), is present in some species of many related genera, e.g., Cotesia, Glyptapanteles, Distatrix, Nyereria, and even in some non-related genera such as Silvaspinosus, and thus does not have the same diagnostic value as expressed by Mason (1981). Some species may be considered as borderline between Cotesia and Protapanteles, and others may be considered as borderline between Glyptapanteles and Protapanteles; thus, it is difficult to clearly define these three genera. Differences between Protapanteles and Cotesia were given in the previous paragraph. Differences with Glyptapanteles are mostly related to the shape of T1. In Glyptapanteles, T1 is either parallel-sided anteriorly and then strongly narrowing posteriorly, or its sides are gradually to strongly converging posteriorly when compared to Protapanteles which has T1 parallel-sided throughout, except for a strongly rounded apex, and propodeum sculpture that is usually, but not always, more rugose and carinated than in Glyptapanteles. Additionally, Protapanteles larvae have mandibles with a row of 12 or fewer large teeth concentrated distally on the blade, and its species distribution is almost completely confined to the Holarctic region (Mason 1981). However, the morphological features mentioned above vary considerably among different species (Arias-Penna et al. 2019).

Glyptapanteles (Figs 100–110) is most likely a polyphyletic assemblage, and may eventually be split into several genera. As a result, it is difficult to define (Arias-Penna et al. 2019). Some of its species may be confused with Protapanteles, Sathon, Lathrapanteles and, to a lesser extent, also Distatrix, Venanides, and Nyereria. The main features defining Glyptapanteles are: fore wing without an areolet; propodeum that is either completely smooth (often) to more or less rugose (more rarely), with a median longitudinal carina that is entirely absent (often), partially defined posteriorly (often) to complete and strong (rarely), or no median carina but instead a series of very short carinae radiating from the nucha (rarely); T1 narrows towards the posterior margin, usually strongly (almost always), or more parallel-sided, or rounded at apex, as in some species of Protapanteles (rarely); T2 is almost always subtriangular or trapezoidal (rarely shaped differently); ovipositor and ovipositor sheaths are relatively short (usually) to moderately long (rarely); setae at apex of ovipositor sheaths relatively long (as long or longer than setae on hypopygium). The differences from Protapanteles were given in the previous paragraph. Sathon has the ovipositor sheaths longer and male specimens have enlarged external genitalia; however, a few Glyptapanteles species have females with longer ovipositor sheaths, and a very few other species have males with external genitalia similarly enlarged; whether those species should be transferred to Sathon requires further study. Lathrapanteles has similar characters to Sathon (see more about those two genera below) and can be separated in the same manner from Glyptapanteles. Distatrix has the pronotum with only one furrow laterally, eyes enlarged and ovipositor sheaths without setae or with very few minute setae, whereas Glyptapanteles has the pronotum with two furrows, eyes that are almost never enlarged (but see Fernandez-Triana 2018, for one exception) and the ovipositor sheaths have much longer setae. Venanides can in turn be separated from Glyptapanteles based on having similar ovipositor sheaths to Distatrix (Mason 1981).

Distatrix (Figs 78, 79) is similar to Venanides, but it has two rows of placodes in the flagellomeres in females, and T2 has a characteristic shape, with the lateral margins widely diverging (Mason 1981, Grinter et al. 2009).

Venanides (Figs 235, 236) can be differentiated from Distatrix because it has only a single row of placodes in the flagellomeres in females, and T2 has less diverging lateral margins (Mason 1981). Additionally, Venanides specimens tend to be smaller and have a dorsoventrally compressed body that is also generally mostly smooth and shiny.

Sathon (Figs 218–220) is distinguished mainly by the enlarged external genitalia in males and relatively long ovipositor sheaths in females; some species probably have the longest sheaths among the entire Cotesiini (sensu Mason 1981). However, these features are not unique: a few Glyptapanteles species have similarly enlarged male genitalia, and all described Lathrapanteles species (Figs 141, 142) are also very similar to Sathon (e.g., Williams 1985, 1988). The limits of Lathrapanteles and Sathon need revision and it is possible that one will eventually be placed in synonymy with the other.

Deuterixys (Figs 64, 65) is a very distinctive genus on account of its T1–T3 sculpture and shape (there appears to be a second constriction between T2 and T3), the propodeum being smooth and shiny and with a complete and strong median, longitudinal carina, and the relatively small body length (Mason 1981, Whitfield 1985, Zeng et al. 2011).

Nyereria (Figs 166–169) has T2 divided into three sections by two deep, usually crenulated, longitudinal grooves delimiting a raised, median area that is not wider than long (Mason 1981). This genus can only be confused with a few species of Cotesia and Glyptapanteles that have their T2 with a similar raised, median area, although in those cases T2 is never as strongly defined by grooves.

Pseudovenanides (Fig. 211) has very scarce information available, but from the original description (Xiao and You 2002) it is clear that it is related to Glyptapanteles and, to a lesser extent, to Venanides. Apparently, T1 with a strongly marked longitudinal sulcus on most of the tergite is the defining feature of this genus.

Ohenri (Fig. 170) has many unique features and is only tentatively considered to be part of this subgroup lacking the fore wing areolet. The pronotum is considerably enlarged dorsally, the ovipositor has its lower valvulae with four subapical teeth, the tarsal claws have large teeth, and the propodeum has a median carina with a partially defined areola (Fernandez-Triana and Boudreault 2018).

Pseudofornicia (Figs 208–210) superficially resembles the (probably) unrelated Fornicia because its metasoma mostly forms a dorsal carapace, but it differs in lacking the epicnemial carina, the fore wing does not have an areolet, and T1 is movably joined to T2, whereas Fornicia has an epicnemial carina, fore wing with an areolet, and T1 and T2 are immovably joined (van Achterberg et al. 2015).

Wilkinsonellus (Figs 241–244) is a very recognizable genus, with T1 very long and thin, propodeum with distinctive sculpture and carination pattern, and fore wing with veins r and 2RS strongly angled (Mason 1981, Long & van Achterberg 2011, Arias-Penna et al. 2013, 2014).

d) Apanteles group

Mason (1981) proposed the tribes Apantelini and Microgastrini to accommodate species with ovipositor sheaths mostly setose and relatively long (at least 0.5 × metatibial length), hypopygium with ventral margin usually flexible and either with one (rarely) or several (commonly) pleats. The latter is the most diagnostic feature for this group; however, there are exceptions (all Alphomelon, most Hygroplitis, and a few species of Apanteles and Microgaster) where the hypopygium is mostly to entirely inflexible. In this paper we combine most of the genera included in the two tribes into a single Apanteles group composed of 26 genera. The group is clearly not monophyletic. Most, if not all, of the species included here have the “microlepidoptera suite of characters” sensu Mason (see further discussion in Mason 1981, Walker et al. 1990). Here we separate the group into several subgroups that can be recognized on simple morphological features, although the genera included in each subgroup are not necessarily related.

The largest subgroup includes 13 genera that lack a fore wing areolet: Alphomelon, Apanteles, Dolichogenidea, Exoryza, Iconella, Illidops, Napamus, Parapanteles, Pholetesor, Pseudapanteles, Rhygoplitis, Shireplitis, and Xanthapanteles. Another two genera could be placed here, at least partially: some species of Choeras lack a fore wing areolet; however, most of the species have a complete or partial areolet so we consider Choeras to be better placed with the subgroup of genera with a complete or partial fore wing areolet; and a similar situation occurs with Promicrogaster, where smaller species tend to lack the areolet whereas the larger species have a complete areolet, and we similarly place that genus in the subgroup with an areolet. These two genera exemplify the challenges of delimiting precise groups in Microgastrinae (a frustration also shared by Mason 1981: 77).

Among the genera without a fore wing areolet, four have the propodeum either with a median longitudinal carina (Iconella, Pseudapanteles, Rhygoplitis) or with a complex pattern that includes full sculpturing and a series of short carinae radiating medially on the posterior 0.2–0.3 near the nucha (Illidops). A fifth genus, Napamus, could also be included in this subgroup, as one of its two described species has the propodeum with a median, longitudinal carina; however, the other species does not (Papp 1993: 170). Nevertheless, Napamus (Fig. 160) can be characterized by its mouth parts elongate, fore wing vein R1 very short (shorter than pterostigma length), inner metatibial spur much longer (1.3 ×) than the outer spur, body and legs black, and wings strongly infumate.

Iconella (Figs 122–124) was described by Mason (1981) as a new genus based on the hind wing with a sinuous vein cu-a as a plesiomorphic character that suggests its unique status among similar genera. Fernandez-Triana et al. (2013a, 2014e) also considered the presence of a median longitudinal carina on the propodeum as strong support for its generic status. However, some Oriental species (with large body size and large and bilobate glossae) currently assigned to Iconella may eventually be placed in a different genus.

Illidops (Figs 125–127) includes species that have the scutellar disc with a medioposterior band of rugosity, fore wing vein R1 shortened, and the propodeum with a series of short carinae medially on its posterior 0.2–0.3, near the nucha (Fernandez-Triana et al. 2014e). In some, but not all species the lower margins of the eyes converge, and T3–T7 are weakly sclerotized (Mason 1981).

Pseudapanteles (Figs 203–207) is characterized by the glossa elongate and strongly bilobed apically, propodeum with a strongly defined median longitudinal carina but no transverse carina (traces of a transverse carina are very rarely present in a few Neotropical species), and T1 with a sharp median sulcus (Mason 1981, Whitfield 1997, Fernandez-Triana et al. 2014a).

Rhygoplitis (Figs 215–217) is the only genus in this subgroup with notauli relatively well defined. It also has the propodeum coarsely sculptured (in addition to a median, longitudinal carina), and fore wing with very short vein R1 (Mason 1981, Whitfield 1997).

The other eight genera without a fore wing areolet have the propodeum with a complete to partial areola, although in large genera such as Apanteles, Dolichogenidea, and Pholetesor, some species have lost all carinae and the propodeum is mostly smooth.

Shireplitis (Figs 225, 226) has the propodeum entirely sculptured, without median or transverse carina, but with the areola defined on the posterior 0.5 by two lateral carinae, ovipositor sheaths relatively short (0.4–0.5 × metatibia length), and legs short and robust – with the metafemur usually less than 3.0 × as long as wide (Fernandez-Triana et al. 2013b).

Alphomelon (Figs 9–11) has the gena with a white/pale spot that is relatively large and very distinctive (Mason 1981, Deans et al. 2003). A few other Microgastrinae genera have some species with a similar pale spot, but it is usually much smaller. Alphomelon is distinguished from the other Microgastrinae with white/pale spot on gena by its ovipositor sheaths being relatively long (much shorter in Cotesia, Glyptapanteles, Protapanteles), mesoscutum anteriorly without strong notauli (strong notauli in Prasmodon), propodeum without a median, longitudinal carina (strong median, longitudinal carina in Pseudapanteles), and the hypopygium inflexible and unpleated (almost always flexible and with several pleats in Apanteles).

Apanteles (Figs 12–26) is currently the most speciose genus in Microgastrinae and has some morphological variability. It usually has the propodeum fully to partially areolated, rarely smooth and never with a median longitudinal carina; fore wing without an areolet; hind wing with the vannal lobe usually strongly concave or straight (see next paragraph for more details on that); ovipositor sheaths relatively long; and the hypopygium almost always flexible and pleated. This genus could only be confused with Pholetesor or Dolichogenidea (which seem to be related to one another, see below) and Parapanteles. Most Apanteles species can be distinguished from both Pholetesor and Parapanteles by the flexible, pleated hypopygium and relatively long ovipositor sheaths (usually at least 0.5 × length of metatibia). In contrast, Parapanteles and Pholetesor have the hypopygium either entirely inflexible or at most with a small, translucent area near the posterior margin (which may look like a pleat in a few species); and the ovipositor sheaths are relatively short (less than half the metatibia length, usually much less). However, a few species of Apanteles have relatively short ovipositor sheaths, and very few species may even have an inflexible hypopygium (e.g., Fernandez-Triana et al. 2014e); the generic placement of those species may be revisited, but at present those exceptions make for a more difficult separation of these three genera.

Dolichogenidea (Figs 81–86) is even more difficult to distinguish from Apanteles, as there is some overlap in some species groups of both genera (e.g., Mason 1981: 53, 54). The differences are frequently subtle and, at times it is very difficult to assign a species to one or other genus depending on the interpretation of morphological features alone. Apanteles has the hind wing with the vannal lobe usually strongly concave or, more rarely, straight to very slightly convex; the central part of the vannal lobe lacks any setae or has few, sparse setae that are often minute and not continuous. In contrast, Dolichogenidea has the vannal lobe convex to slightly straight; the central part of the vannal lobe is more or less entirely setose so that a continuous fringe of setae is almost always visible (although setae may be small in a few species). The fringe of setae (or lack of them) is the only morphological character that almost always seems to work in distinguishing these genera from each other; we are aware of very few species currently assigned to Dolichogenidea where the fringe is not complete and could lead to the species being placed within Apanteles, despite molecular data strongly suggesting the best generic placement is Dolichogenidea. Other features function only partially and seem to represent trends that are far from being universally present in one genus or the other. For example, the anteromesoscutum punctures (when present) tend to be partially or completely fused near the scutoscutellar sulcus in Apanteles, whereas in Dolichogenidea, which usually does not have punctures on the anteromesoscutum anteriorly and very rarely has them near scutoscutellar sulcus, the punctures never fuse. The scutoscutellar sulcus in many Dolichogenidea species tends to be very narrow and sometimes looks almost obliterated, whereas the sulcus in Apanteles is usually wider. Despite the rather subtle morphological differences, DNA barcodes tend to cluster both genera clearly apart (Smith et al. 2013, Fernandez-Triana et al. 2014e).

Dolichogenidea tends to cluster near Pholetesor (Figs 184–190) and these genera seem to be closer to each other than either is to Apanteles. Dolichogenidea has a flexible, pleated hypopygium and relatively long ovipositor sheaths (usually at least 0.5 × metatibia length) whereas Pholetesor has the hypopygium entirely inflexible or with a small, translucent area near the posterior margin that could look like a pleat in a few species, and the ovipositor sheaths are relatively short, less than half the metatibia length (Mason 1981, Whitfield 2006).

The status of Exoryza (Figs 92–94) as a valid genus has been questioned by many authors (Valerio et al. 2004, Rousse and Gupta 2013, Fernandez-Triana et al. 2014e, 2016c). Mason (1981) characterized it as having T1 and T2 heavily sculptured, and the propodeum coarsely rugose, with an areola present but obscured by heavy sculpture. However, the distinction between Exoryza and Dolichogenidea may be particularly difficult because many species of the latter genus have the propodeum sculptured, with or without an areola, and T1 is occasionally sculptured, although not as strongly as in Exoryza (Fernandez-Triana et al. 2014e, 2016c).

Parapanteles (Figs 172–175) is a very difficult genus to understand at present. Parks (2018) found it to be paraphyletic. Some species of “Parapanteles” with available DNA barcodes cluster within Dolichogenidea and could just be considered as species within that genus, with short ovipositor sheaths and an inflexible hypopygium (similar to Pholetesor and the few borderline species of Apanteles mentioned above). Another group of Parapanteles seems to represent misidentifications of Cotesia (e.g., Valerio et al. 2009, Freitas et al. 2019). Whether a group of species that could be considered true Parapanteles actually exists remains to be seen. For the present, the genus can be defined as having the propodeum completely to mostly areolated (usually with well defined carinae), ovipositor sheaths short, and an inflexible hypopygium.

Xanthapanteles (Fig. 245) is a very distinctive genus, on the basis of the propodeum fully areolated with strongly defined and raised carinae, T1 very large and wide, T1–T3 sculpture like a finely pebble-grained surface (unlike any other Microgastrinae), flagellomere placodes arranged irregularly and fore wing relatively slender and much longer than body length (Whitfield 1995b).

Another subgroup within the Apanteles group includes six genera, Agupta, Dasylagon, Hypomicrogaster, Papanteles, Promicrogaster, and Sendaphne, that can be recognized by the fore wing with a very small areolet, sometimes almost obliterated. They also share (except for Agupta, see below) having the scutellum with lunules relatively high, more than 0.5 × the height of its lateral face. These genera are separated from each other based on different propodeal carination patterns, and T1 and T2 shapes and sculptures. Some described species of Choeras, almost exclusively from the Oriental region, have a very small areolet and thus could be included in this group. However, these are exceptions and are very likely to be transferred elsewhere or classified separately. For now, we place Choeras (see below) within the subgroup with a large fore wing areolet.

Agupta (Figs 5, 6) does not have enlarged lunules; however, it can be recognized by several unusual features: in males (and sometimes in females) the antenna has the first few flagellomeres with placodes irregularly distributed in three rows or no row can be clearly defined; the propodeum has a strongly raised median carina with small radiating carinae across its length; T1 shape (narrowing for first half, then parallel-sided) and T1 sculpture (anterior half mostly smooth, strongly concave and with central sulcus, posterior half punctured and with a polished area on posterior margin) are distinctive; and the body length is among the largest in Microgastrinae (second only to the unrelated genus Larissimus) (Fernandez-Triana and Boudreault 2018). Some large specimens of Choeras in the Oriental region (see previous paragraph) might end being placed within Agupta when more studies are done in the future.

Promicrogaster and Sendaphne can be recognized by the following combination of features: glossa elongate and bilobate, metacoxa very long (0.8–1.0 × metafemur length and 0.6–0.8 × metatibia length), and ovipositor and ovipositor sheaths very long – among the longest in Microgastrinae usually 2.0 × as long as the metatibia or even longer. Most species have the body length longer than the fore wing length, usually by 0.2–0.4 mm (the majority of Microgastrinae species have the fore wing slightly longer than the body length). These two genera are very closely related and may eventually be treated as a single genus. Promicrogaster (Figs 194–197) has the ovipositor apically sinuate; propodeum sculptured and usually with some carination (which may include a complete or partial median longitudinal carina, or an indication of a partial areola posteriorly); T1 parallel-sided to slightly narrowing towards the posterior margin; and T2 transverse, its width at the posterior margin 3.0–4.5 × (rarely 2.0 ×) its length medially (Fernandez-Triana et al. 2016b, Fernandez-Triana 2019). Sendaphne (Figs 223, 224) has the ovipositor straight apically, propodeum mostly smooth and without carina (with the rare exception of having sparse punctures and a few rugae near the nucha), T1 strongly narrowing towards the posterior margin, and T2 subtriangular (Fernandez-Triana et al. 2014h).

Dasylagon (Fig. 63) has the propodeum fully areolated (defined by strong carinae), T1 comparatively very wide and large (in dorsal view more than 0.3 of entire metasoma), T2 very transverse, metasomal terga entirely smooth, hind wing with a sinuous vein cu-a, and ovipositor and ovipositor sheaths relatively very long (more than 1.5 × metatibia length) (Mason 1981).

Hypomicrogaster (Figs 113–121) has the propodeum with a complex carination pattern, which includes a median carina and a more or less complete areola, although some species have all carinae reduced, but still the propodeum would be mostly sculptured. The head is relatively transverse, i.e., wider than in most other genera of Microgastrinae, and T1 and T2 are mostly to entirely smooth (Mason 1981, Valerio and Whitfield 2015).

Papanteles (Fig. 171) has the propodeum fully areolated, T1 relatively long (ca. 2.0 × its width at posterior margin), T1 and T2 strongly sculptured, T2 and T3 comparatively narrow and not occupying the entire dorsal surface of the segment (dorsal width of T2 and T3 half the width of T5 and following terga), and the ovipositor sheaths are approximately the same length as the metatibia length (Mason 1981).

The remaining eight genera in the Apanteles group all have the fore wing areolet relatively large; even when some species may have a relatively smaller areolet, it never appears almost obliterated.

Ypsilonigaster (Figs 250–252) has a very characteristic T1, with a median sulcus shaped like an inverted Y, a unique feature to recognize the genus (Fernandez-Triana and Boudreault 2018).

Hygroplitis and Microgaster have the propodeum with a median carina, fore wing areolet relatively large, anteromesoscutum anteriorly mostly smooth, T1 and T2 heavily sculptured (also T3, partially or entirely), T1 relatively large and wide (width at posterior margin greater than width at anterior margin), and T2 mostly rectangular. The two genera are very closely related and DNA barcodes suggest Hygroplitis may eventually be synonymized under Microgaster. Hygroplitis (Figs 111, 112) has the body somewhat depressed dorsoventrally, notauli more strongly impressed, flagellomeres with three rows of placodes, and the hypopygium usually inflexible although in some cases it is weakly but distinctly pleated (Mason 1981); whereas Microgaster (Figs 147–150) does not have the body dorsoventrally depressed, the notauli are barely visible, flagellomeres are usually (but not always) with two rows of placodes, and the hypopygium is usually (but not always) flexible and pleated (Mason 1981).

Paroplitis (Figs 178–180) species are relatively small, with a body length of 2.5 mm or less; legs, especially the metafemur, short and robust; antenna short, with flagellomeres in females having only a single row of placodes; hypopygium almost entirely sclerotized but with a sharp fold medially; propodeum rarely entirely sculptured but almost always with a median longitudinal carina, at least on the anterior 0.5, and sometimes also with a complete or partial transverse carina; and T2 usually smooth, rarely sculptured (Mason 1981, Fernandez-Triana et al. 2013b).

Kotenkosius (Fig. 138) has a unique propodeal carination pattern that includes three complete longitudinal carinae, one medially, the other two sublaterally, and a complete transverse carina near posterior 0.6, with additional small striae radiating from the median and sublateral longitudinal carinae, and most carinae strongly defined and raised (Fernandez-Triana and Boudreault 2018).

Choeras (Figs 40–45), as presently understood, is clearly a polyphyletic assemblage of species, some of which may eventually be placed in different genera. It is one of the few Microgastrinae genera that has some species without a fore wing areolet (although the shape of the remaining veins r, 2RS, and 2M usually indicate a partially defined areolet), and other species with a complete areolet that can vary from very small in some species to large in others (van Achterberg 2002, Fagan-Jeffries and Austin 2018b). The propodeum also varies, from having a complete longitudinal median carina to having a partial one, to not having any visible median carina, or having just minute carinae radiating from the nucha. T1 is mostly rectangular (slightly narrowing towards the posterior margin in some species), but never much wider on the posterior margin than on the anterior margin, and T2 is mostly transverse. Many Oriental species of “Choeras” most likely represent different lineages from the temperate species and may warrant placement in different genera, e.g., some of the species may be better placed in Agupta.

Dodogaster (Fig. 80) has a unique set of features in the Apanteles group. The propodeum has a more or less complete areola and a partial median carina, the fore wing has a relatively large areolet, and T1–T3 are heavily sculptured and almost form a carapace (Rousse and Gupta 2013).

Microgastrinae are present in all continents except Antarctica. Specimens can be found in all major terrestrial ecosystems, from 82°30'N (Canada, Nunavut, Ellesmere Island, Alert) to 55°S (Argentina and Chile, Tierra del Fuego) in the New World and 50°S (New Zealand, Auckland Islands) in the Old World, and from sea level up to at least 4,500 m (Fernandez-Triana 2018). The information currently available allows us to make preliminary comments on species diversity and distribution at the generic level (Table 5 and Fig. 2).