The Sepsidae of the Mitaraka expedition, French Guiana (Diptera)

ABSTRACT This paper documents on the sepsid fauna (Diptera: Sepsidae) encountered during the Mitaraka 2015 expedition. These are the first published records of Sepsidae Walker, 1833 from French Guiana. During this survey, six genera with 11 species were recorded from French Guiana for the first time: Archisepsis armata (Schiner, 1868), Archisepsis diversiformis (Ozerov, 1993), Archisepsis peruana (Ozerov, 1994), Meropliosepsis sexsetosa Duda, 1926, Microsepsis armillata (Melander & Spuler, 1917), Microsepsis furcata (Melander & Spuler, 1917), Microsepsis mitis (Curran, 1927), Palaeosepsioides erythromyrmus (Silva, 1991), Palaeosepsioides mitarakensis Silva, n. sp., described herein, Palaeosepsis Duda, 1926 and Pseudopalaeosepsis Ozerov, 1992. The actual state of Sepsidae taxonomy, based mostly on the male fore femur morphology and genitalic characters, has led to a case where sometimes females cannot be properly assigned to a specific taxon – this was the case here for most of the Microsepsis Silva, 1993 females which corresponded to 25% of the collected specimens, one female assigned to Palaeosepsis dentata (Becker, 1919)/dentatiformis (Duda, 1926), and two females at present identified as Pseudopalaeosepsis sp. A checklist of Sepsidae of French Guiana is provided also including doubtful and unidentified female specimens.


INTRODUCTION
Sepsidae Walker, 1833 are a small family of acalyptrate Diptera. Most members have an ant-like appearance, due to a basal constriction of the abdomen; they are small to medium-sized (body length 2-7 mm), with hyaline wings, which feature a dark spot near the tip in some species. The larvae are closely associated with various decaying organic materials, but occur most frequently on mammalian faeces (Silva 2016).
The family is well represented in all biogeographic regions, with 361 valid species in 38 valid genera (Ozerov 2005;Iwasa 2008;Ozerov & Iwasa 2008;Ang et al. 2008Ang et al. , 2017Ozerov & Krivosheina 2011, 2014Ozerov 2012Ozerov , 2014Ozerov , 2018aIwasa & Thinh 2012). Silva (2010) presented a key to the Neotropical genera and a synopsis of the fauna of Central America. At present, the Neotropical region includes 45 described valid species in 10 genera (Silva 2016). There are no published records of Sepsidae from French Guiana and only 13 species in five genera are reported from the neighbouring countries of Guyana and Brazil (states of Amapá and Pará) (Ozerov 2005). Because many species are morphologically very similar which might mask subtle differences between species, the revision of current species concepts in combination with molecular analyses might well reveal a much higher species richness and more distinct ecological preferences.
During the 2015 expedition in southwestern French Guiana, invertebrates including Diptera were collected with a wide array of methods. The Museum national d'Histoire naturelle, Paris and the NGO Pro-Natura International selected French Guiana as study area for this expedition because the knowledge on the arthropod fauna of this French overseas department was still very fragmentary (Brûlé & Touroult 2014;Pascal et al. 2015;Touroult et al. 2018). The objective of the present paper is to document on the sepsid fauna encountered during the above-mentioned survey, to present a checklist of Sepsidae of French Guiana and to describe a new species of Palaeosepsioides Ozerov, 1992 discovered during the expedition.

MATERIAL AND METHODS
French Guiana is a French overseas department of about 85 000 km² located on the eastern Guiana Shield. It is bordered to the east and south by Brazil and to the west by Suriname. The climate is equatorial, characterized by a mean annual temperature of 26°C and annual rainfall varying from 2000 mm in the south and west to 4000 mm in the northeast. The rainy season usually lasts from May till August and the main dry season from December till January, with a smaller one in April (Guitet et al. 2014).
In 2015 the "Our Planet Reviewed" or "La Planète revisitée" Guyane 2014-2015 expedition, also known as the "Mitaraka 2015 survey", was conducted in French Guiana (Pollet et al. 2014Pascal et al. 2015;Touroult et al. 2018) as the fifth edition of a large-scale biodiversity survey undertaken by the Museum national d'Histoire naturelle, Paris and the NGO Pro-Natura international (both in France). Basic arthropod taxonomy and species discovery were at the heart of the survey, although forest ecology and biodiversity distribution modelling were also project topics. The expedition was conducted in the Mitaraka Mountains, a largely unknown and uninhabited area in the southwesternmost corner of French Guiana, directly bordering Suriname and Brazil (Pollet et al. 2014;Krolow et al. 2017). It is part of the Tumuc-Humac mountain chain, extending east in Amapá region (Brazil) and west in southern Suriname. The area consists primarily of tropical lowland rain forest with scattered inselbergs, i.e., isolated hills that stand above the forest plains. The rain forest itself consists of numerous small hills of about 300-400 m a.s.l., separated by small to larger streams or palm swamps.
Between 22 February and 27 March 2015, two consecutive equal-sized teams (of about 30 researchers) explored the area, including more than 10 invertebrate experts. A third smaller team returned to the site from 12 to 20 August 2015. The second author (MP) was coordinator of the collected Diptera, and was also the only Diptera worker actively involved in ZOOSYSTEMA • 2020 • 42 (14) this survey. Invertebrate sampling was carried out near the base camp, on the drop zone (an area near the base camp that had been clear-cut entirely to allow helicopters to land) and, in particular, along four trails of about 3.5 km that started from the base camp in four different directions (see Krolow et al. 2017;Pollet et al. 2018;Touroult et al. 2018). During the first period (22 February-11 March 2015) over 21 different collecting methods were applied, with a total of 401 traps operational within a perimeter of 1 km². This array consisted primarily of pan traps (n = 280), Charax butterfly traps (n = 50), square Malaise traps (SLAM) (n = 32), Flight Intercept Traps (FIT, n = 13) and Butterfly banana traps (n = 12), but also a light trap. In the subsequent periods, pan traps were no longer operational. A total of 223 invertebrate samples (often pooled yields of different traps of the same type) were examined, including 94 sweepnet samples. However, only one of the latter contained sepsid flies as MP focused mainly on Dolichopodidae Latreille, 1809 during his active collecting.
Non-pan trap samples were sorted to insect orders and families at the SEAG offices (http://insectafgseag.myspecies. info/fr), while pan trap samples were treated similarly at MP's home lab. The biological sampling benefited from the access and benefit sharing agreement "APA973-1", which is included in each sample label. Dipteran subsamples (mostly per family) were subsequently disseminated among experts worldwide (see Pollet et al. 2018;Touroult et al. 2018), in the case of Sepsidae to the first author (VCS). The identification of the sepsid species was conducted using taxonomical reviews and identification keys (Ozerov 1993(Ozerov , 1994(Ozerov , 2004Silva 1993Silva , 2010, original descriptions (Silva 1991;Ozerov 1992), and via direct comparison to reliably identified species in collections of the Museu de Zoologia de São Paulo, São Paulo (MZUSP). All collected material was stored in 70% alcohol, while the holotype of the new species was dried using HMDS (Brown 2016

Collecting methods* BT FIT MT (6 m) SLAM SW BPT WPT YPT
No. species 1 9 3 4 1 1 2 3 No. specimens 5 54 7 7 1 3 12 32 Guyana, Suriname and Brazil (states of Amapá and Pará) as based on literature data (mainly Ozerov 2005) is also provided (see Table 2). During the Mitaraka 2015 survey, the largest number of species and specimens was encountered in FIT's. These numbers were about two and nearly eight times higher than those of SLAM traps respectively, though the latter method produced nearly four times as many samples. Other trap types collected four species at most. Yellow pan traps yielded considerably larger sepsid numbers as compared to white and blue traps, but the species richness was indifferently poor in all pan trap types. Nine species were encountered near the drop zone where 46% of the specimens were collected, whereas trail A and C provided only 3 and 6 species respectively. No Sepsidae were collected on the investigated inselbergs and the two species Archisepsis armata (Shiner, 1868), Microsepsis armillata (Melander & Spuler, 1917), that were gathered on one of the rocky outcrops ('savane roche 2') also occurred within the canopied rain forest habitats. Microsepsis sp. and Microsepsis armillata proved to be most widely spread, being collected in seven and six of the 37 different sampling sites respectively. Considering the actual state of Sepsidae taxonomy, based mostly on the male fore femur morphology and genitalic characters, sometimes females cannot be properly assigned to a specific taxon -this was the case here for most of the Microsepsis Silva, 1993 females, one female assigned to Palaeosepsis dentata (Becker, 1919)/dentatiformis (Duda, 1926), and two females at present identified as Pseudopalaeosepsis sp.
A new species of Palaeosepsioides Ozerov, 1992, Palaeosepsioides mitarakensis Silva, n. sp., was discovered during the survey and is described herein. reMArks

CHECKLIST OF SEPSIDAE OF FRENCH GUIANA
The record of this species from Puntarenas, Argentina, by Ozerov (2005) is considered an error. A locality with the name Punta Arenas is known from Chile (province Magallanes) and one called Puntarenas (province Puntarenas) from Costa Rica, but not from Argentina. As no Sepsidae have been recorded from Chile thus far, and as the species is already known from Costa Rica, it is believed that Ozerov's locality refers to Puntarenas in the latter country. New to French Guiana. (Ozerov, 1994) Phalacrosepsis peruana Ozerov, 1994: 92.  Color. Frons blackish brown in posterior third and yellowish brown in anterior two thirds; face and gena whitish yellow. Postcranium with occiput blackish brown and postgena yellowish brown; thorax and abdomen brownish yellow, medial postsutural scutum posterior to the anterior dorso-central seta and scutellum brownish. Legs yellow in the following parts: all coxae, anterior part of fore femur, fore tibia, apical two thirds of mid tibia, and all tarsi. Mid and hind femora brown, with basal part yellowish. Basal third of mid and hind tibiae blackish brown, and apical third of hind tibia brown. Wing without dark spot near apex. Basal costal cell, membrane area posterior to stem vein completely, and costal cell in basal half, blackish.
Legs ( Fig. 2A, D). Fore coxa long and simple, with one dorsal seta apically. Fore femur and tibia as in Fig. 2D. Mid coxa with a row of setulae in upper half. Mid femur with two strong anterior setae in center. Mid tibia with one strong ventral, one preapical anterodorsal, one apical posteroventral and one anteroventral setae. Hind femur without major setae. Hind tibia without osmeterium, and with one apicodorsal seta.
Wing (Fig. 2E). Hyaline, with only basal costal cell, costal cell and base of wing dark brown. Basal medial and basal radial cells separate. Anal vein long. Alula entirely covered with microtrichia, narrow. Upper calypter with setulae on margin. Halter white.

Female
Unknown. note The abdomen of the holotype was not dissected because the male specimen is unique, and the species is adequately defined by external characters. As the abdomen was not bended under the body it was possible to identify some of the features (see further).
reMArks Based on the current knowledge of Neotropical sepsids, and using the key of Silva (2010), this male specimen collected in the Mitaraka survey keyed out to Palaeosepsioides Ozerov, 1992. It did not match, though, two of the characters of the generic concept as defined by Ozerov (1992): "absence of hairs on the upper calypter margin" (the specimen shows long setulae on the upper calypter margin) and "the abdominal sternite 4 extremely long, with long lateral lobes, considerably longer than abdominal sternite 5" (our specimen shows only long setae on the posterior margin of tergite 4). However, its abdomen features a strong constriction after tergite 1 + 2 and lacks a postmetacoxal bridge. It certainly does not belong to Meropliosepsis Duda, 1926, Meroplius Rondani, 1874 or Themira Robineau-Desvoidy, 1830 due to the absence of a well-developed fronto-orbital seta; nor can it be assigned to Microsepsis Silva, 1993 as the male terminalia are symmetrical and the CuA + CuP vein is longer than the length of the cell cua. It does not belong to Archisepsis Silva, 1993 or Pseudopalaeosepsis Ozerov, 1992 either because it shows an elevation in the posterior margin of the syntergite 1 + 2 and the abdominal tergites are polished and shiny, with short setae. And finally, it does not fit into Lateosepsis Ozerov, 2004 nor Palaeosepsis Duda, 1926 because it does not have a postmetacoxal bridge. As a result, it seems most appropriate to assign it to Palaeosepsioides and to adjust the definition of the genus in anticipation of more material to be collected in order to carry out a phylogenetic analysis of the sepsid genera that include this taxon.
Sepsis dentatiformis Duda, 1926a: 46.  reMArks Both species, P. dentata and P. dentatiformis, were not recorded from French Guiana before, which renders this the first record of this genus for this part of South America, awaiting the discovery of male specimens that will allow an unequivocal species identification. This specimen has the dorsocentral setae arranged 0 + 2 (no presutural versus two postsutural dorsocentral setae), which excludes P. punctulata Ozerov, 2004 (with dorsocentral setae arranged 0 + 1). The wing is not spotted at apex, which renders it either P. dentata or P. dentatiformis (both feature a wing without an apical spot . Taking into account the species from neighbouring countries that currently seem to be lacking in French Guiana (Table 2), the use of additional collecting methods might well increase this French Department's species list to about 20. Considering the biology of sepsids, the best collecting method are traps, baited with mammalian faeces; cattle dung is the standard bait used (Pont & Meier 2002). In our study, FIT's were most productive in terms of both species (nine of 11 species) and specimens (over 2/ 5 of all specimens), though, over 1/4 of the specimens were retrieved from the yellow pan traps. An optimal sampling for Sepsidae in this kind of area would be by sweeping or hand-netting over substrates attractive to adults as well as the use of passive trapping devices like FIT's and Malaise trap types, together with traps baited with fresh mammalian dung. Sampling sites should be visited regularly during the day, during consecutive days, until no more sepsids are attracted to the baited traps. It is also important to install those traps in a diversity of microhabitats (e.g. shady versus open, sunny sites) in order to sample species that prefer either of these habitats.