Three new genera of Mymaridae (Hymenoptera) from the Neotropical region

Three new genera and species of Mymaridae from the Neotropical region are described: Megamymar waorani Huber, gen. and sp. nov. ; Neopolynemoidea chilensis Huber, gen. and sp. nov. ; and Porcepicus herison Huber, gen. and sp. nov. Their possible relationships are discussed to place them in context among the previously described genera of Mymaridae.


Introduction
Mymaridae are a relatively large family of Hymenoptera, at present including about 100 valid genera, 90 of them extant and 10 extinct, including 2 fossils that may be misplaced in Mymaridae. This is almost double the number treated in the only world key (Annecke and Doutt 1961). Over the past 60 years but especially during the last 40 large numbers of specimens have accumulated in major institutions. These were sometimes collected using Malaise or yellow pan traps in previously poorly collected countries or areas, or in unusual habitats such as in tree canopies or soil. As a result, many new species continue to be discovered and publications that attempt to define generic limits more precisely must to take them into account. Nevertheless, the limits for many previously described valid genera still remain unclear and may become even more unclear because of the additional species discovered. Their limits could perhaps be redefined to include or exclude species that are near the edge of a given generic concept. If excluded, such species could be placed, perhaps doubtfully, into another genus where they appear to fit better. However, specimens occasionally are found that are clearly so unusual that they cannot be placed reasonably into any existing generic concept. These specimens not only represent new species but also new genera. Three are described below to make their names available for a planned generic key and species catalogue of Mymaridae in the Neotropical region excluding Mexico.

Methods
Specimens were point or card mounted and photographed. Because of its large size, the specimen representing the first genus and species was retained on a point mount and photographed using a ProgRes C14 plus digital camera mounted on a Nikon SMZ1500 microscope; its description and, especially, measurements are therefore not as accurate as for the other two species. One specimen of each of the other two species was dissected and slide mounted in Canada balsam. These specimens were photographed at different focal planes using the same camera as above mounted on a Nikon Eclipse E800 compound microscope, and the resulting layers combined and refined using Zerene Stacker TM . The layers were edited from the top of the focus stack down to produce one image showing surface structures. For certain views, such as posterior or ventral, the layers were combined and edited from the bottom up. These views were flipped either horizontally or vertically in Adobe Photoshop to appear as if the specimen was photographed from that surface. A unique image number is given to different structures but if different planes of the same structure (from the same individual) are illustrated, e.g., of the metasoma (Fig. 10), then the different planes are given the same number with different letters. Body length of a card-mounted paratype of each of two species was measured with an ocular micrometer mounted in a Leitz-Wetzlar binocular microscope at 64× or 160× magnification, as needed. Morphological terms follow Gibson (1997), with some terms from Huber (2015). Measurements are given in micrometres (μm) except body and ovipositor length of the largest species are in millimetres (mm). Because of rounding errors, it appears that the ratios are slightly incorrect compared to those calculated using the absolute measurements (micrometres) but, in fact, those calculated from the ocular micrometer readings before converting to micrometers are more accurate, so are given in the descriptions.
Abbreviations used fl = flagellar segment (in males), fu = funicular segment (in females), gt = gastral tergum, LOL = distance between a lateral ocellus and median ocellus, mps = multiporous plate sensillum, OOL =distance between a lateral ocellus and eye margin, POL = distance between lateral ocelli. The specimen depositories are: long as mesoscutum as measured from their junction with anterior margin. Scutellum ~4.0× as long as poorly defined frenum; axilla barely advanced, each ~1.0× as wide as long. Metanotum slightly longer than frenum, without obvious dorsellum. Propodeum slightly longer than scutellum + frenum, with a shallow and narrow longitudinal median depression, in lateral view almost horizontal; spiracle in a shallow wide depression. Wings. Fore wing fairly wide, with apex rounded and slightly asymmetrical, and with almost straight margin behind venation; venation ~0.2× as long as wing length; parastigma with proximal but apparently without distal macrochaetae, with hypochaeta next to proximal macrochaeta. Hind wing narrow and almost straight. Legs. Legs long; tarsi 4-segmented, with tarsomere 1 of all legs longer than tibiae. Metasoma. Metasoma ~2.3× as long as mesosoma (Fig. 1). Petiole ~3.4× as long as wide, in dorsal view hidden by anterior extension of gaster. Gaster (measured to anterior apex of anteriorly truncate dorsal horn) ~6.4× as long as wide and ~1.2× as high as wide; gt 5 apparently the longest tergum ( Fig. 1). Ovipositor with exserted portion extremely long.
Male. Unknown. Derivation of genus name. A euphonious combination from Greek: megas, meaning large, and Mymar, the name of the type genus of Mymaridae. The genus name is neuter. Mega refers to the large body of the only known species of the genus, which is over 1 mm longer than the next longest Neotropical species, Erdosiella mira (Annecke & Doutt), whose body length is 3.7 mm.
Relationships. Megamymar is best placed in Mymarini sensu Annecke and Doutt (1961) because of the combination of 6-segmented funicle, 4-segmented tarsi and distinctly petiolate gaster. Among the genera of Mymarini, the supraorbital trabecula extending posteriorly only to the level of the median ocellus, the fore wing shape, short fringe setae and venation with relatively long parastigma indicate that Megamymar is most closely related to Erdosiella Soyka and Tanyostethium Yoshimoto in the New World, and perhaps Narayanella Subba Rao in the Old World. Megamymar is clearly separated from these three genera by the petiole much shorter than the gaster and the anterodorsal extension of the base of the gaster over the apex of the propodeum; none of their species have this combination of features. Instead, in these genera the supraorbital trabecula extends to about the level of the lateral ocelli, the petiole is usually as long or longer than the gaster, and gt 1 does not project anteriorly over the propodeum though occasionally the base of gt 1 may slant slightly anteriorly before receding uniformly and smoothly posteriorly to the anterior margin of gt 2 . It could perhaps be argued that the features of Megamymar are just one of degree rather than substance and therefore Megamymar should be treated as just an extreme representative of Erdosiella, but then many of the numerous genera of Mymarini should be placed in synonymy under one another as well. Whether such an approach would clarify generic relationships within Mymarini is debatable.
Megamymar superficially resembles several of the largest species of Australomymar Girault and the extralimital genera Borneomymar Huber, Neotriadomerus Huber and Paranaphoidea Girault. At least some species in all these genera, none of which are mophologically closely related to Megamymar, have a long ovipositor often greatly exserted posteriorly (Australomymar, Borneomymar, Polynemoidea Girault) or anteriorly (Neotriadomerus). In the Neotropical region, Australomymar appears superficially to be the most similar looking genus but the body of the largest species is at most ~3.0 mm long and the base of the female gaster does not extend anteriorly in a dorsal horn (the gastral sac). In other genera of Mymaridae the ovipositor may extend anteriorly but it is always ventral to the mesosoma. Diagnosis. Megamymar waorani is the only described species in the genus. Its diagnosis is therefore the same as for the generic description, i.e., ovipositor extending anteriorly in a short, truncate gastral sac dorsal to propodeum (Fig. 1).
Derivation of species name. The species is named after the indigenous Waorani people of Ecuador in whose reserve M. waorani was collected. Their way of life has been seriously affected by resource extraction and settlement by colonists. The species name is treated as a noun in apposition.
Biology. Unknown, but because of its size, likely a solitary parasitoid in large insect eggs. We suggest that the host is most likely a species of Orthoptera. First, despite M. waorani being the third longest fairyfly species known worldwide, after specimens of a species of Neotriadomerus from Australia and specimens of one species of Australomymar from New Zealand, it has apparently never been collected at ground level, despite considerable Malaise or pan trapping in equatorial rain forests of the Neotropical region. Second, the host egg must be at least 4.8 mm long and eggs of this size are mostly likely to be found among species of Orthoptera. Third, although hosts of any species of Mymaridae with body length over 3.0 mm are unknown, one relatively small (~1.3 mm) species of Australomymar has been reared from Tettigoniidae (Orthoptera) and one large (2.5 mm) species of Acmopolynema Ogloblin has been reared from Oecanthus spp. (Orthoptera: Gryllidae). Diagnosis. Female with the following combination of features: toruli almost touching transverse trabecula (Fig. 2a); scape over 10× as long as greatest width and 1.4× as long as head width, funicle 6-segmented and clava 3-segmented (Fig. 3); fore wing venation extending almost 0.7× wing length (Fig. 7).
Male. Unknown. Derivation of genus name. From Greek: Neo-meaning new, referring to its occurrence in the New World and Polynemoidea, a monotypic genus known so far only from Tasmania. The name Neopolynemoidea is given to draw attention to the general similarity the two genera, one from the Old World and one from the New World.
Neopolynemoidea differs from Polynemoidea by: toruli almost in contact with transverse trabecula (separated from transverse trabecula by almost its own height in Polynemoidea), scape over 10× as long as wide (no more than 3× as long as wide in Polynemoidea); fore wing venation at least 0.6× as long as fore wing length (no more than 0.4 as long as fore wing length in Polynemoidea).
Relationships The majority of these genera occur in the southern hemisphere, particularly in the southernmost areas, though several extend well into the northern hemisphere. Two other genera appear to have the clava with 3-segments in at least one species: one Eustochus (Eustochus) from China was described as having 3-segmented clava (the other described species have a 2-segmented clava) but this may be an artefact of partial antennal collapse, giving the appearance of a third segment; and Kompsomymar Huber, with a single described species from Australia, appears to have only partial divisions separating the claval segments. The hosts and biology of all but one (Stethynium) of the above genera are unknown. Among these, Krokella, Paracmotemnus and Polynemoidea have a fore wing venation longer than half the wing length, as in Neopolynemoidea, but none have the extremely long scape in females. Females of the only described species of Polynemoidea, however, have a long, exerted ovipositor. Therefore, based wing venation and ovipositor features a close relationship of Neopolynemoidea to Polynemoidea is proposed as being the most probable. Diagnosis. Neopolynemoidea chilensis is the only described species in the genus so its diagnosis the same as for the generic description.
Description. Female. Head. Head slightly narrower than mesosoma (14:17), ~2.0× as wide as long, ~1.7× as wide as high and ~1.15× as high as long, measured laterally; transverse and supraorbital trabeculae with short dark sections alternating with light sections. (Fig. 12a). Face slightly wider than high; torulus ~1.4× its own height from transverse trabecula; preorbital groove bulging laterally ventral to eye. Compound eye ~1.7× as long as malar space and with a few short blunt setae among ommatidia. Vertex posteriorly with sharp, slightly concave margin at junction with occiput (Fig 12b); median ocellus well separated from transverse trabecula; ocelli in a low triangle, with lateral ocelli at posterolateral angle of vertex, and POL ~2× LOL and ~23× OOL. Back of head with vertical occipital groove, complete transverse occipital groove obtusely angled medially dorsal to occipital foramen, and oral cavity posteriorly almost confluent with occipital foramen (Fig. 12b). Antenna. Scape with radicle barely differentiated; funicle 6-segmented; clava 1-segmented (Figs 14a, b). Mouthparts. Mandibles without teeth, with rounded apex shorter than maxilla, presumably not meeting when closed (Fig. 12a). Mesosoma. Mesosoma ~0.8× as long as wide, ~1.2× as long as high, and ~1.4× as wide as high. Pronotum in lateral view almost vertical, apparently longitudinally divided medially, in dorsal view barely visible except laterally. Prosternum almost 2.0× as wide as long and longitudinally divided medially (Fig. 19). Mesoscutum as long as scutellum + frenum, in lateral view flat; notauli incomplete, barely indicated anteriorly (Fig. 18a). Scutellum ~0.25× as long as frenum, without campaniform sensilla or fenestra; axilla barely advanced, about as long as wide; second phragma extending to apex of propodeum, widely truncate apically. Metanotum linear and barely visible, without defined dorsellum. Propodeum medially ~5.0× as long as metanotum. Wings. Fore wing narrow and curved ( Fig. 15) with distinct lobe posterior to parastigma; venation ~0.3× as long as wing length; parastigma with distal macrochaeta but without proximal macrochaeta or hypochaeta. Hind wing ( Fig. 16) with distinct bend in basal third beyond venation, with anterior margin concave and posterior margin convex. Legs. legs short; tarsi 5-segmented, with protarsomere 1 the longest segment, and meso-and metararsomere 1 the shortest segments. Metasoma. Metasoma ~0.8× as long as mesosoma (Figs 18a, 20). Petiole vertical, not visible in dorsal view, ~10× as wide as long (high). Gaster ~1.2× as wide as long, and ~1.1× as wide as high; gt 2 and gt 3 the longest terga, gt 1 planoconvex, the almost straight anterior margin and curved posterior margin meeting laterally to form an acute angle (Fig. 18a); cercus apparently with only 2 cercal setae. Ovipositor arising in apical half of gaster, slightly less than 0.5× gaster length and slightly less than 0.5× metatibia length.
Male. Unknown. Derivation of genus name. An arbitrary combination of letters based on the French word for porcupine, porc-épique, referring to the long and strong setae distributed on the antenna, body, and legs.
Relationships. Porcepicus belongs to the Camptoptera group of genera. It appears to be most similar to Camptoptera by the back of the head having a vertical occipital groove, transverse occipital groove, and narrow and curved fore wing and lack of a hypochaeta. The 6-segmented funicle in females, slightly dorsoventrally flattened mesosoma, gaster wider than long, and apparent absence of a petiole distinguishes it from Camptoptera Foerster as well as the other genera in the genus group. Diagnosis. Porcepicus herison is the only described species in the genus so its diagnosis is the same as for the generic description.
Derivation of species name. The name is an arbitrary combination of letters similar to the French word for hedgehog, hérisson. The name is treated as a noun in apposition.

Biology.
Unknown. As with most species in the Camptoptera group of genera the hosts are unknown. We suggest the hosts are Coleoptera based on at least one record from that order for Camptoptera and one for Litus.

Discussion
The fairyfly fauna of the Neotropical region is one of the most diverse in the world, with about 50 genera, taking into account synonymies and genera added since Huber (1995Huber ( , 2006; it is about equal to or second to that of the Australian region (Lin et al. 2007). Generically, the fauna is certainly much more diverse than the Nearctic region including Mexico (Guzmán-Larralde et al. 2017;Huber et al. 2020) and the Afrotropical region (Huber et al. 2020). The species described above do not fit in any previously described genus of Mymaridae in the Neotropical region. While one would expect most new genera to be found among the smaller members (less than 0.5 mm in length) of Mymaridae, it is a surprise that the largest member represents a distinct new genus. This shows how little is really known about the Neotropical fauna, especially in relatively uncollected habitats such as tree canopies of equatorial forests or soil.