Overview of the Black Fungus Gnats from the Tristan da Cunha archipelago , including a redescription of Hyperlasion viridiventris ( Frey ) ( Diptera : Sciaroidea : Sciaridae ) With 2 maps , 23 figures and 1 table

Black Fungus Gnats (Diptera: Sciaridae) were collected with Malaise traps on Tristan da Cunha and Nightingale islands in the South Atlantic. Four species were identified of which Lycoriella sativae (Johannsen, 1912) is already known from literature. For three others – Bradysia tilicola (Loew, 1850), Hyperlasion viridiventris (Frey, 1945) and Pseu­ dolycoriella setigera (Hardy, 1960) – this is the first report from this archipelago. To date, 8 species are known from this group of islands. They belong to the genera Bradysia Winnertz (3 species), Cosmosciara Frey (1), Hyperlasion Schmitz (1), Lycoriella Frey (2) and Pseudolycoriella Menzel & Mohrig (1). The male genitalia for 7 particularly well known sciarid species from the archipelago are already illustrated but until now only poorly preserved type specimens of Hyperlasion viridiventris (Frey) from the Azores were known. Both male and female have been redescribed from the new material. An overview is also given of the existing sciarid records from the islands. The published data were analysed and compared with the new records.


Introduction
The four main islands of the Tristan da Cunha archipelago, Tristan da Cunha, Nightingale, Inaccessible and Gough, all lie in the South Atlantic between 37°S 12°W and 40°S 09°W [see map and locality description in Hänel & Heyne (2008): 122, fig. 1]. The Sciaridae, commonly known as Black Fungus Gnats, were first recorded at the archipelago from a collection made in the three northern islands (Inaccessible, Nightingale, Tristan da Cunha) during the 'Norwegian Expedition to Tristan da Cunha in 1937-1938 . This brought to light two species, namely Cosmosciara perniciosa (Edwards, 1922) and Bradysia mycorum Frey, 1948[in Menzel & Mohrig (2000 a junior synonym of Lycoriella ingenua (Dufour, 1839)], both regarded as cosmopolitan species (Frey 1954). Subsequently two more species were recorded from a collection made during the Gough Island Scientific Survey of [1956][1957]. They were two undetermined species of Bradysia Winnertz (Freeman 1962). One of these was later identified with additional material collected during the 'Gough Island Terrestrial Invertebrate Survey of 1999-2001' , as Bradysia nocturna Tuomikoski, 1960[= B. forficulata (Bezzi, 1914] with Holarctic distribution was not mentioned in Jones et al. (2003). Also recorded from the 1999-2001 collection was another genus, namely Lycoriella Frey, 1942, represented by two species. To further investigate and update the situation in the northern islands, a brief invertebrate collecting project was carried out at Tristan da Cunha and Nightingale islands in early 2005 as a snippet of a larger program entitled 'Empowering the People of Tristan da Cunha to Implement the Convention on Biological Diversity (CBD)' (Hänel 2005). The authors, having pursued the findings beyond the scope of the projects means, here present the information that has come to light from that collection. Involved are four identifiable species, three of which had not been recorded from the archipelago before. These include the rare and poorly known species Pseudolycoriella seti gera (Hardy, 1960) and Hyperlasion viridiventris (Frey, 1945). The findings are presented, discussed and summarised together with existing information, so as to provide an overview of the 'status quo' with regards to the Sciaridae as currently known for the Archipelago.

Material and methods
The fieldwork was conducted by Christine Hänel on Tristan da Cunha and Nightingale Islands (hereafter referred to as 'Tristan' and 'Nightingale') between early February and the beginning of April, 2005. The majority of the time was spent on Tristan and a period of two weeks on Nightingale. Sampling was random and opportunistic, but focused on trying to cover the major habitat types represented. On Tristan, 23 sites at eight different locations were sampled. Most of the areas sampled (62.5 %) were in the lowlands, although a few (12.5 %) were on the slopes above, and a couple (2.5 %) were on the plateau. No higher altitudes were sampled. The habitats encompassed principally agricultural crop and pastures, some 'intruded' fern-bush communities, and a boggy wetland (Soggy Plain). On Nightingale, 10 sites were sampled in two main regions. The latter comprised the lowland areas dominated by tussock grasslands and the higher central wetland areas. Sciaridae were obtained at six of the eight Tristan localities (Map 1) and at both localities on Nightingale (Map 2). Flies were principally collected using Malaise traps. All specimens were preserved in 70 % ethanol. The specimens were prepared by P. Vilkamaa as permanent mounts for microscopic examination using Euparal as mountant. The nomenclature and systematics used are based on the revisions by Menzel & Mohrig (2000) and Menzel & Smith (2009). Representatives of each species from Tristan da Cunha archipelago are deposited in the following collections (Hänel, pers. comm

Results
In total four Sciaridae species were collected in 2005. These are Bradysia tilicola (Loew, 1850), Hyperlasion viri diventris (Frey, 1945), Lycoriella sativae (Johannsen, 1912), and Pseudolycoriella setigera (Hardy, 1960). The details for the collected specimens are summarised below with the data pertaining to the relevant sites (e.g. locality, habitat, and vegetation types) and a general overview of the species status.

Material from the archipelago:
Only known from the literature (Freeman 1962, Holdgate 1965.

Discussion:
Bradysia tilicola is perhaps the most common species of Sciaridae worldwide. Due to its cosmopolitan distribution it has been described under different names 19 times. For a complete list of synonymies see Menzel & Mohrig (2000). The species is potentially present in nearly every house in the world with indoor plants. The larvae develop in flowerpots and may occasionally cause damage to the plants. The species is also found outdoors, but is never highly abundant. It may spread out actively from human settlements and is most likely to having been introduced with plants to nearly any place with human activities and also to Tristan da Cunha. There are already records from other archipelagoes like the Azores and the Canary Islands.
For more detailed information about habitats see .

Material from the archipelago:
Only known from the literature (Freeman 1962, Holdgate 1965.

Distribution:
Recorded from Gough Island (2  in the BMNH collection).

Material from the archipelago:
Only known from the literature (Frey 1954).

Discussion:
Cosmosciara perniciosa has only occasionally been recorded in Europe from natural habitats like moor, dry grassland and deciduous woodland, but was mostly found near human settlements, in gardens and glasshouses, even as a pest on young plants (e.g. cucumbers). As the genus Cosmosciara is more tropical than Palaearctic, it can be assumed that this species was introduced into Europe by man. In that way it has most probably reached many distant islands, also the Tristan da Cunha archipelago.

Distribution:
The species was previously only known from the Azores (Pico, San Jorge and San Miguel Islands). Accordingly, its status was regarded as a Macaronesian endemic (Borges et al. 2005). A total of 36 specimens were collected on Tristan da Cunha and Nightingale Islands. New to Tristan da Cunha archipelago.

Description:
Male: Body length: ca 1.8 mm. Head: brown, roundish; eyebridge 2 facets wide, narrowed towards the middle into 1 facet, medially without facets. Antennae dark brown and short; 4th flagellomere 2.1-2.2 times as long as wide (Fig. 16); necks short, pale and well differentiated; basal part 1.7-1.8 times as long as wide, with long, coarse setae and very long sensillae; antennal setae slightly longer than segment width. Maxillary palpi yellowish, with one club-like segment (Fig. 12); palpal segment with 2-3 bristles (one long, distinctly curved outer bristle above sensory pit; 1-2 shorter bristles on apex of maxillary palp); sensory pit large, deep and on apex of maxillary palp; sensilla long. Thorax: dark brown. Anterior pronotum with 4-5 setae. Prothoracal episternum short, with 3-5 setae. Postpronotum bare. Katepisternum (Fig. 14) short, low and distinctly emarginated, with cranial hollow. Mesonotum with sparse, short setae and 2-4 longer lateral bristles; scutum short, with rather short and thick laterals, thick and curved dorsocentrals in two indistinct rows, acrocentrals in anterior part only, short and curved. Scutellum distinct with 2 strong marginal bristles and some shorter setae. Legs and coxae short, strong and pale brown; coxae with strong and long setae; legs distinctly compact with short tarsi; fore femur thickened; tibiae short, shorter than tarsi, thickened towards apex; all tibiae without spines among the setae; apex of fore tibia with sparse prolateral patch of setae subapically; tibial spurs 1+2+2, shorter than diameter of corresponding tibiae, tibial spurs short and pin-like, one spur on mid and hind legs shorter (slightly unequal in length, sometimes both spurs reduced). Claws narrow and straight, without teeth. Wing (Fig. 10) light brown with reduced, wedge shaped base (wing length 1.3 mm, width/length 0.45); C with sparse setosity; posterior veins weak and as the wing membrane without macrotrichia; stem of M longer than M-fork; fork of M short, triangular and very wide; x = y, both bare; CuA-stem short, about 1/3 to 1/2 x; R 1 = 0.75-0.85 R; R-R 1 short, merging with C well before base of M-fork; c = 0.65-0.80 w. Halteres short, strong and brown with short stalk. Abdomen: slightly paler than thorax; body setae sparse, short, fine and dark. Hypopygium (Fig. 8) about as high as wide, pale brown as abdomen; gonocoxites short; intercoxal area unmodified; ventral base widely rounded; inner margin with short, relatively dense setae. Gonostylus (Fig. 9) longish oval, about 2.2 times as long as wide (slightly shorter than gonocoxite); apex widely rounded, with short and strong bristles (no apical tooth); innerside with 7 hyaline megasetae [spines] in upper half arranged in 2 groups (4 subapical and 3 very closely adjacent megasetae in middle). Tegmen higher than wide, unsclerotized and apically roundish; area of teeth very small and round, with fine teeth. Aedeagus very short. Female: Similar to male but slightly larger. Body length: 2.5 mm. Head: Ocelli present. Eyebridge complete, 2 facets wide. Maxillary palpi with one segment (Fig. 13), with 2-4 bristles (one distinctly longer). Antennae with 14 flagellomeres; basal segments of antennae dark; pedicel enlarged and distinctly attenuated; 4th flagellomere 1.7-2.2 times as long as wide (Fig. 17); necks short and well differentiated; basal part 1.25-1.5 times as long as wide, with coarse, long, erect setae and very long sensillae; setae about as long as segment width. Thorax: Anterior pronotum with 4-5 setae. Prothoracal episternum with 7-8 setae. Katepisternum (Fig. 15) larger and longer than in males, wedge shaped with cranial hollow. Legs slightly thicker than in male, forelegs strong, with swollen femur (Fig. 11); tibial spurs short, one spur on mid and hind legs distinctly shorter (more strongly reduced than in males). Wings distinctly narrower than in males (wing length 1.4-1.6 mm, width/length 0.40-0.45); R 1 shorter, = 0.5-0.8 R; c = 0.85 w; necks of halteres slightly shorter. Abdomen: long and narrow, with fine short setae, with greenish tint. All other characteristics as in males.

Discussion:
Hyperlasion viridiventris (Frey) was described from eight specimens from the Azores, of which one male (lectotype; hypopygium missing) and five females (paralectotypes) are left in the collection of MZH. The new specimens from Tristan da Cunha are in general paler than Frey (1945)

Material from the archipelago:
Only known from the literature (Frey 1954.

Distribution:
Recorded from Gough, Inaccessible and Tristan da Cunha Islands (Frey 1954

Discussion:
Like Lycoriella sativae (Johannsen) [= L. castanescens], Lycoriella ingenua is a ubiquitous species associated with fungi. As such it is well known as a pest in cultures of fungi, but it is less frequent than L. sativae in natural ecosystems. It was probably introduced with fungi cultures throughout the world and has formerly been recorded from a couple of other islands. For more detailed information about habitats see .

Discussion:
Larvae are found in plant material in the soil and are involved in decomposition. They feed on fungi and in this connection have been regarded as a pest to mushroom growers (e.g. Tibbles et al. 2005). Due to this probable relation with fungal hyphae Lycoriella sativae (Johannsen) [= castanescens (Lengersdorf)] is common in nearly every kind of ecosystem throughout the Holarctic Region and most probably beyond. It is especially dominant in disturbed ecosystems like cultivated fields and temporarily flooded areas. For more detailed information about habitats see .

Discussion:
Although Pseudolycoriella setigera has currently been found only on three remote island localities, we suppose, that it is a widespread species. It only appears to be rare, because it does not occur in the Palearctic but possibly in the poorly studied Afrotropical Region. The occurrence on four different archipelagoes leads to the assumption, that it was also introduced by man.

Discussion
All the species of Black Fungus Gnats that have to date been found at the Tristan archipelago are considered to be alien introductions to the islands. The cosmopolitan Cos mosciara perniciosa (Edwards) and Lycoriella ingenua (Dufour) [in Frey (1954)  virtually no restrictions or quarantine controls. Tourists are welcome and landings on all three northern Islands are permissible. The abundance of spectacular and rare birds occurring at the Tristan Island is a source of great attraction that has in recent years brought an increasing number of visitors to the islands, including tour groups. The latter arrive mostly on vessels that do not operate with strict hygiene regulations, such as for instance are enforced by the supply ship 'S.A. Agulhas' when moored at a foreign port and transporting goods and passengers to the Tristan Territory. This is particularly problematic where tour operators run a route of island-hopping expeditions and no quarantine measures are in place. Since the occurrence of the two newly recorded sciarids Pseu dolycoriella setigera and Hyperlasion viridiventris is otherwise only known from one locality each on the globe, namely Hawaii and the Azores respectively, and Tristan does not conduct trade with these islands, it points strongly to their import having been via a tourism operation of an uncontrolled type. It is also probable, that these two species have a wider distributional range than only the respective archipelagoes and occur also in continental areas like Africa, South America or East Asia, which are only poorly known concerning their sciarid fauna. Interesting is that Pseudolycoriella setigera and Hyperla sion viridiventris were represented in various habitats and localities on Tristan and Nightingale, including the male form of H. viridiventris which had hitherto been poorly known. Clearly the Islands provide favourable conditions for them. Little to no information exists about the biology and habits of the species, and not much can at this stage be deduced from the 2005 findings alone, as no quantifiable type of collecting methods were employed nor could a comprehensive survey of all habitats be made. It is therefore somewhat curious that on Tristan, the male of Hyperlasion viridiventris was only found at sites higher than 1260 ft. a.s.l., despite the fact that collecting efforts were minimal in elevated areas compared to the more extensive ones carried out in the lowlands. However, on Nightingale the males were found at both the high level sites (ca. 660 ft. a.s.l.) and at sea level.
Only one of the mentioned species was possibly not introduced by human activities, namely Bradysia forficulata. This was recorded from Gough Island alone, the most isolated island of the archipelago. Adult Sciaridae are transported by atmospheric drift as aeroplankton over distances of more than 1,000 km (Harrell & Yoshimoto 1964), so it is possible that Bradysia forficulata may have reached Gough Island in that way. Another indication for this hypothesis is the existence of Brady sia watsoni Colless, 1962 from Macquarie Island in the South Pacific (Colless 1962). We have not studied the type specimen, but regarding the description it is not impossible, that it is identical to B. forficulata. There are currently no endemic species of Sciaridae known from Tristan da Cunha and it is questionable, if sciarid species existed there at all in earlier times before human influence. Endemism is still not proven for Sciaridae, because the distribution of most species is known only fragmentarily and because aerial drift is a common phenomenon in the family. The Tristan da Cunha archipelago is one of the most isolated on earth and there have not been any land bridges to other land masses. That is why there are not even local faunistic elements of Sciaridae, like for example on Campbell Island, with a clear predominance of species from New Zealand (Harrison 1955).