Introduction to the Mymaridae (Hymenoptera) of Bangladesh

Abstract An identification key to the 15 genera of Mymaridae found so far in Bangladesh is given, based on about 520 specimens collected using yellow pan traps placed in agricultural habitats and at the edge of ponds, mainly at Bangabandhu Sheikh Mujibur Rahman Agricultural University, Gazipur. Species already reported from Bangladesh are listed and three more are added: Acmopolynema orientale (Narayanan, Subba Rao & Kaur), Himopolynema hishimonus Taguchi, and Mymar pulchellum Curtis.


Introduction to the Mymaridae (Hymenoptera) of Bangladesh Introduction
Ten named species of Mymaridae (Hymenoptera), representing four genera, have been recorded from Bangladesh: Anagrus flaveolus Waterhouse (Kamal et al. 1993, Sahad andHirashima 1984), almost certainly a misidentification of A. nilaparvatae Pang & Wang (Triapitsyn 2014); Anagrus incarnatus Haliday (Sahad and Hirashima 1984, Gurr et al. 2011), this is likely a misidentification of A. nilaparvatae (Chiappini 2002: 236); the ethanol, critical-point dried, and mounted on cards. Representative specimens of all but two of the genera were slide mounted in Canada balsam, using the method described in Huber (2015). Photographs of the head, antenna, and wings were taken with a ProgRes C14 plus digital camera attached to a Nikon Eclipse E800 compound microscope, and a selection of the resulting layers combined electronically and edited in Zerene Stacker TM . Specimens are deposited in the Canadian National Collection of Insects, Arachnids and Nematodes, Ottawa, Ontario, Canada, and the University of Rajshahi, Motihar, Rajshahi, Bangladesh. Abbreviations used in the key are: fl x for funicle segment, and mps for multiporous plate sensilla.

Results
The breakdown of the ≈ 520 card-and slide-mounted specimens is approximately as follows (some specimens of the two most commonly collected genera,  (Fig. 30); hind wing extremely narrow, the distance between anterior and posterior margins at most about the length of a setae on the wing membrane (Fig. 32)  Fore wing bare or almost so behind venation; subantennal grooves almost in contact with each other but if not then with distance between them at junction with mouth margin much less than half the distance from a groove to preorbital groove at lateral margin of face .............Cosmocomoidea Howard -Fore wing with at least one row of microtrichia, but usually with numerous scattered microtrichia behind venation; subantennal grooves with distance between them at junction with mouth margin at least half distance from a groove to preorbital groove at Head in anterior view quite wide ventrally, the genae only slightly converging; mandibles directed medially, their apices crossing each other, the head not appearing beaklike; antenna with fl 2 ringlike (Fig. 9); fore wing evenly wide along its length distal to venation and distinctly curved near apex (Fig. 10); gaster separated from propodeum by a distinctly narrow petiole, the mesophragma thus not extending posteriorly into gaster ...... Camptoptera Förster -Head in anterior view quite narrow ventrally, the genae strongly converging; mandibles directed ventrally and narrowing apically, their apices usually not crossing each other and giving head a beak-like appearance; antenna with fl 2 about as long as preceding and following segments; fore wing much narrower medially along much of its length distal to venation then distinctly widening near apex; gaster widely joined to propodeum by a wide petiole barely distinguishable from propodeum or gaster, so mesophragma projecting posteriorly well into gaster . Face with distinct subantennal groove extending from each torulus to mouth margin (Fig. 33); antenna with clava 3-segmented (Fig. 33); fore wing with distinct lobe posterior to and just distal to apex of stigmal vein (Fig. 34) (Fig. 7); fore wing usually without or with only a slight lobe (Figs 5, 8), rarely with a more distinct lobe posterior to and just distal to apex of stigmal vein (Fig. 27)  Mandibles greatly reduced to minute stubs without teeth, and maxilla elongate; head in lateral view with eye in contact with back of head, the gena almost entirely absent; fore wing membrane rather sparsely and unevenly covered with microtrichia concentrated mainly in apical half of wing beyond venation apex (Fig. 12) (10) Vertex with ocellar triangle surrounded by a stemmaticum (seen as white lines) (Fig. 4); clava in lateral view usually asymmetrical, with its dorsal margin strongly curved and ventral margin straight (Fig. 4); fore wing narrow, without marginal and medial spaces and without socketed seta at apex of retinaculum .. Fore wing with posterior margin behind venation distinctly lobed (Fig. 27) (12) Face with a small pit medial to each torulus (Fig. 17); antenna usually with funicle segments short (Fig. 17) (Fig. 1); antenna usually with funicle segments, especially fl 2 , longer (Fig. 2); fore wing with at least a few, thickened blunt microtrichia mainly on the dark areas (Fig. 3)

Biology
Published host records exist for at least one species in most of the genera keyed above (Huber 1986). Cicadellidae and related families of Auchenorrhyncha (Hemiptera) are hosts for ten genera: Acmopolynema, Anagrus (Anagrus) + Anagrus (Paranagrus), Cosmocomoidea, Gonatocerus, Himopolynema, Lymaenon, Mymar, Palaeoneura, Polynema (Polynema) and Stethynium (Triapitsyn 2002). The remainder parasitize a variety of hosts. The principal host families only are listed here: Anaphes on Chrysomelidae and Curculionidae, Erythmelus on Tingidae and Miridae. Camptoptera and Dicopus have few or no published host records; they appear to parasitize Coleoptera and Psocoptera, respectively. Hosts are unknown for Ptilomymar which is closely associated with water and almost certainly parasitizes eggs of aquatic insects of some kind. The various genera were collected in the following crops or habitats (not all specimens had associated plant names): Palaeoneura bagicha (Narayanan, Subba Rao & Kaur). BANGLADESH. Dhaka: Kalni, 2-5.vii.2007, N. Islam, pan trap in lady's finger field (1 female, CNC).

Discussion
Features useful for generic identification of most Mymaridae occur on the head, female antenna, and wings of specimens. Careful study of these structures requires well-mounted specimens on slides and/or good photographs. When these are available most specimens from a given country may be identified correctly to genus on this basis alone, often without having to examine other body parts. The generic identification key was carefully and deliberately constructed to demonstrate this. Features of the antenna and wings are also relatively easy to study on card-or point-mounted specimens. Only a few features of the mesosoma and metasoma were added, where necessary. However, additional features are certainly also useful and are needed when the fauna of an entire region is treated. Those additional features are, of course, widely used in all generic keys to Mymaridae, e.g., Ramesh Kumar et al. (2013) and are essential to define a genus properly. At the generic level the almost unknown fauna of Mymaridae of Bangladesh, with 15 genera recorded here, is about 40% of the much better studied fauna of India. Ramesh Kumar et al. (2013) recorded about 140 species classified into 31 genera. Since then, Huber (2015) reclassified the species groups of Gonatocerus into separate genera and other genera (new to India, not yet recorded from Bangladesh) have been recorded (e.g., Triapitsyn 2014). bringing the number of Indian genera to almost 40. Many of the genera found in India almost certainly occur also in Bangladesh, as further collecting will undoubtedly reveal. The number of species in Bangladesh will be fewer than in India simply because it does not have the variety of ecosystems and elevational range of its far larger neighbour. At the species level, much more work would be needed to sort out and identify correctly the specimens collected in our study. This can only be done meaningfully in the context of more regional studies that include not only India but preferable the entire Oriental region and Palaearctic areas of eastern Asia.
The greatest number of genera collected was at pond edges. This is perhaps not surprising because it is a much more natural habitat, presumably with many more plant species and potential insect hosts (both terrestrial and aquatic) than experimental field plots planted with a single crop. Triapitsyn SV (2015) Taxonomy of the genus Anagrus Haliday (Hymenoptera: Mymaridae) of the world: an annotated key to the described species, discussion of the remaining problems, and a checklist. Acta Zoológica Lilloana 59 (1-2): 3-50. Triapitsyn SV, Berezovskiy VV (2004) Review of the genus Anagrus Haliday, 1833 (Hymenoptera: Mymaridae) in Russia, with notes on some extralimital species. Far Eastern Entomologist 139: 22. Trjapitzin SV (1996) Taxonomic notes on the Australian species of Anagrus (Hymenoptera Mymaridae). Russian Entomological Journal 4(1-4): 106.