Fossils Shed a New Light on the Diversity and Disparity of the Family Limoniidae (Diptera, Nematocera)

Simple Summary A new established subfamily (family Limoniidae) from Cretaceous Burmese amber shows a unique reduction of radial veins combined with complete set of medial veins. The discovery of this new subfamily contributes to a better understanding of the diversity and disparity of the family, and is important for further research on the evolution of this group of insects. Abstract A new subfamily Drinosinae (Diptera, Limoniidae) is established with two fossil genera, Drinosa and Decessia gen. nov. with one new species, Decessia podenasi gen. et sp. nov. from Cretaceous Burmese amber. Additional description of Drinosa prisca is based on new material. A new subfamily shows unique reduction of radial veins combined with complete set of medial veins.


Introduction
Cretaceous ambers are of great value for paleoentomological research. We find there the inclusions of descendants of the Triassic and Jurassic fauna hitherto known only from imprints in sediments, as well as the first representatives of families and genera which dominate in recent fauna [1][2][3][4][5][6][7][8]. Especially important is Burmese amber (Figure 1), aged at the beginning of the Upper Cretaceous, Cenomanian (98.79 ± 0.62 Ma [9]). This amber comprises large number of perfectly preserved complete bodies of insects which enable us to carry out research at the same scientific precision as on recent entomofauna. The most abundant in Burmese amber among insect inclusions are specimens of flies, which are often almost perfectly preserved [10]. Wing venation, antennae, legs, and especially male genitalia are the base for systematic and phylogenetic research of this group of insects, and enable us to compare fossil and recent specimens [11]. This is especially true for groups like the Limoniidae (lower Diptera, =Nematocera) with externally exposed male genitalia, which provide the necessary set of characters diagnostic to the species.

Materials and Methods
The study was based on 10 inclusions from Burmese amber dated on 98.79 ± 0.62 Ma, (Upper Cretaceous, Cenomanian) [9] (Figure 1), two inclusions of Decessia podenasi gen. et sp. nov. and eight of Drinosa prisca [13]. The specimens were examined with the Nikon (SM Z25) stereomicroscope equipped with a Nikon digital camera (DS-Ri2), and with the Nikon SMZ 1500 stereomicroscope equipped with a Nikon DS-Fi1 camera. The measurements were taken with NIS-Elements D3.0. Drawings were completed by tracing the photographs. The wing venation nomenclature follows that of [11,39], and the terminology of male genitalia follows [40]. The measurements are given in millimeters (mm), the length of particular segments of antennae or palpus according to the pattern: antenna or palpus section number/length of this section in millimeters. The "MP" abbreviation is part of specimen traditional number from the collection of Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Kraków and it means "Muzeum Przyrodnicze" -Natural History Museum. Institutional Abbreviation: ISEA PAS-Institute of Systematics and Evolution of Animals, Polish Academy of Sciences, Kraków, Poland.
Description. Body brown, ca. 2 mm long; prothorax small, scutellum rather narrow; pleura of thorax and lateral margins of mediotergite in some specimens blackish due to oxidation. Head wider than long, with medium-sized, oval eyes widely separated; vertex covered with sparse, light setae longer than scape but shorter than head. Antenna longer than head, 13-14 segmented, flagellomeres with very long verticils longer than segments bearing them; scape rather short, cylindrical or nearly cylindrical; pedicel barrel-like, longer than wide, slightly longer than scape; first flagellomere expanded medially, shorter than pedicel, other flagellomeres rather thin and oval, last flagellomeres usually shortest. Palpus short, with last palpomere longer than penultimate one; palpomeres covered by dense, erect, short setae, with additional scarce long setae, usually longer than segments bearing them.    Thorax brown, wings approximately three times as long as wide without darkened spots except dark, oval pterostigma. Wing venation: Sc ending opposite 1/2 to 2/3 of Rs; sc-r positioned close to or at tip of Sc; arculus absent, Rs almost as long as R 2+3+4 ; R 1 ending just behind half length of R 2+3+4 ; cell r 2+3+4 widely expanded in distal part, distal part of cell r 5 very narrow; crossvein m-m present; crossvein m-cu in distal part of d-cell, rather short and straight; M 1 and M 2 short, petiole of m 1 present; d cell small, trapezoidal; crossvein m-cu in 2/3 of d cell; A 1 long, almost straight; A 2 long, straight, approximately as long as Mb. Legs long, slender, without tibial spurs.
Male genitalia differ between the two genera and are presented in the generic descriptions.       Comparison. In contrast to all other Limoniidae, in the Drinosinae, subfam. nov., the vein Rs branches only into two veins: R2+3+4 and R5; therefore, only three radial veins are present and reach the wing margin: R1, R2+3+4 and R5. The crossvein r-r (R2) is absent. The medial field is not reduced in the Drinosinae; all four veins M1, M2, M3 and M4 are present and reach the wing margin, and the discal cell (d cell) is closed. In other subfamilies of Limoniidae except genus Helius, Elephantomyia Osten-Sacken, 1860 [19] and Toxorhina Loew, 1850 [46], five radial veins are present, including the crossvein r-r (R2).  (Figures 2A and 3B-D) longer than head, but shorter than length of head and thorax combined. Flagellomeres 2-5 with very long setae, approximately as long as three or four flagellomeres combined. Palpus (Figures 2B and 3D) rather thin, palpomeres with few, not very long setae. Wing (Figures 2C and 3E,F) Figures 3A and 5A-E), semi-inverted, ninth tergite and sternite fused. Gonocoxite greatly elongated, central part cylindrical, covered with very long setae, gonocoxite bears a long dorsomesal process at distal 1/5. Two pairs of gonostyles present, positioned apically on gonocoxite and forming together C-shaped structure; inner gonostylus very short, with two apical branches, outer gonostylus long and thin, tapering to apex and curved inwards. Aedeagus very long, thin; distal 1/4 is bifid; paramere rod-shaped, short. Additional, probably semi-membraneous structure extends between both gonocoxites, and terminates in two divergent, semicircular horns which support the basal part of an extremely long, flexible aedeagus. This structure seems less sclerotized than parameres.

Additional description. Antenna
Female: ovipositor ( Figures 2F and 5E,F) with very long, thin cerci; hypogynal valvae long, thin and stiff. Remarks. The ovipositor of Drinosa prisca was incorrectly described and illustrated in original description [13]. Authors probably studied a destroyed female specimen, with missing or invisible distal part of ovipositor. A female studied by us (no. MP/4074; Figures 2F and 5F) is complete with a well visible ovipositor, which is slender, with very long both cerci and hypogynal valvae.
Diagnosis. Antenna 14-segmented. Venation: Sc longer than half wing length; Rs branches into two veins reaching wing margin: R 2+3+4 and R 5 ; d cell closed; petiole longer than upper margin of d cell and longer than M 1 ; four medial veins reach wing margin; m-cu positioned in distal half of d cell. Male hypopygium with gonocoxite not very long, with long mesobasal process half as long as gonocoxite; outer gonostylus long and thin, saber-shaped; inner gonostylus half as long as outer gonostylus, tapering to apex, with small lobe at base.
Comparison.Decessia gen. nov. and Drinosa differ by the number of flagellomeres: antennae of Decessia are 14-segmented while those of Drinosa, 13-segmented. The two genera differ mainly by morphology of male hypopygium. In Decessia gen. nov. the gonocoxite is much shorter than in Drinosa; also position of processes on gonocoxite is different in both genera. The aedeagus in Drinosa is extremely long and thin, extending far beyond the hypopygium and supported by a large, additional structure; in Decessia the aedeagus is very short, comprised between the parameres. There are also differences in wing shape and venation: in Decessia gen. nov. the wing is more rounded, and the vein Sc extends beyond half wing, while the wing of Drinosa is more rectangular, with a well-developed anal angle, and Sc ends about mid wing; there are also differences in proportions of d cell and distal section of vein M 1 and the petiole of m 1 , as illustrated in Figure 2C (Drinosa) and Figure 6C (Decessia).
Thorax ( Figure 7A): wing 1.80 mm long, 0.64 mm wide, approximately three as long as wide. Wing ( Figures 6C and 7C): vein Sc ending opposite 2/3 of vein Rs; sc-r just before before tip of Sc; arculus absent; Rs almost as long as R 2+3+4 ; vein R 1 ending just behind half the length of R 2+3+4 ; cell r 2+3+4 widely expanded distally, cell r 5 distally narrowed; crossvein m-m present, d cell closed; crossvein m-cu straight, positioned in distal part of d-cell; M 1 and M 2 short, petiole 0.23 mm long, d-cell short and wide, 0.21 mm long, trapezoidal, crossvein m-cu in 2/3 length of d-cell base; A 1 elongate, almost straight; A 2 elongate, slightly curved at wing margin, approximately as long as Mb.
Legs ( Figure 7A) slender, without tibial spurs. Abdomen: hypopygium ( Figure 6D,E and Figure 7D) 0.31 mm long; gonocoxite elongate and narrow, 0.16 mm long, with mesobasal long and thin process; outer gonostylus almost straight, c. as long as gonocoxite and twice as long as inner gonostylus; inner gonostylus tapering to the apex, tip curved; aedeagus short, barely extending beyond parameres which are massive, crescent shaped.
Female unknown.

Discussion
The Limoniidae, represented in recent fauna, are the oldest group within Tipulomorpha, and one of the oldest families of the Diptera in general [11]. Among the recent fauna, it is one of the most species-rich families, but also their fossil record is also one of the richest among the Diptera in the Mesozoic and Cenozoic. The oldest representatives of Limoniidae are known from the Upper Triassic [1,11,27,30], while the immediate ancestory lineage, the Archilimoniidae, were among the oldest known fossil flies from the Middle Triassic [11]. The first rapid radiation of Limoniidae occurred at the turn of the Triassic to Jurassic [47][48][49]. The first 70 million years of Limoniidae evolution are known only from imprints in sediments what limits our knowledge about their morphology mostly to the wing venation. Usually, male and female genitalia are only fragmentarily preserved in sediments which hinders a detailed study on morphology of these structures. Since the Lower Cretaceous the Limoniidae are documented also by inclusions in fossil resins, beginning from Lebanese amber (age 140 Ma [50,51]). The inclusions in fossil resins provide a full and certain information on morphology of entire bodies and allow comparison of fossil and recent species almost on the same level. Equipped with this knowledge we can better understand the ways of evolution and the phylogenetic relationships within family Limoniidae and all the infraorder Tipulomorpha. In this study we have decided not to perform a new phylogenetic analysis because previous analyses did not bring satisfactory results and we consider it better to wait until more similar new discoveries in the fossil record are published, to reach as broad taxon sampling as possible.
During over 210 Ma years of evolution numerous species of Limoniidae appeared and testify to the great diversity of this group in the past and today. Various evolutionary lineages (genera) show convergences in morphology, which impede tracking the processes of evolution within the family. Unfortunately, molecular research based on fossil specimens appeared unsuccessful due to breakdown of DNA during the fossilization, which concerns the sediments as well as the resins. Information contained in sensational publications proved to be premature, and till now no successful analysis was made, even when examining specimens perfectly preserved in amber ( [52]; for reviews see: [53][54][55]).
Even molecular research on recent taxa often do not provide us with sound basis for full phylogenetical research, as they do not represent extinct lineages. Therefore, fossil specimens are of fundamental value for such research, especially when they convey full information on the morphology of entire bodies, as the amber inclusions do.
The present discovery of new species in the beginning of Upper Cretaceous (Cenomanian) Burmese amber offers a new insight on taxonomy within Limoniidae. New materials enabled us a full reconstruction of male and female genitalia of a previously described species, Drinosa prisca. It was previously correctly observed that this species does not fit any known subfamily of limoniids [13]. The morphological features of newly discovered species Decessia podenasi gen. et sp. nov. inclined us to classify both species to a new subfamily and to designate a new genus.
Wing venation of the genus Drinosa was compared [13] with wing venation of representatives of the genera Teucholabis Osten-Sacken, 1859 [19] and Gonomyia in which a strong reduction of radial veins is visible (citation: "The wing venation of Drinosa resembles that of the genera Teucholabis and Gonomyia which are characterized by strongly reduced Rs branches with only two reaching the wing margin. The radial branches of Drinosa are even more reduced because of the lack of R 2 vein with only R 3 and R 4+5 present" [13]: 471). In fact, in these genera the vein R 4+5 does not exist, which is also true to all Limoniidae, save a few exceptions. In the Limoniidae, as well as in most Tipulomorpha, the vein Rs branches into veins R 2+3+4 and R 5 ( [27,38]: 145). The presence of the vein R 4+5 is characteristic only for the oldest evolutionary line of Tipulomorpha, the Archilimoniidae from the Middle Triassic [12,31], and to a majority of the Pediciidae, whose oldest fossil representatives are known from the Middle Jurassic ([25]: 56; [56]: 136; [57]: 240). Wing venation of representatives of all seven subfamilies of Limoniidae, and of the Archilimoniidae is shown in Figure 8. In the Limoniidae an exceptional, rare appearance of a very short section of vein R 4+5 occurs, for instance, in the genus Mesotipula (subfamily Architipulinae); among several fossil species of this genus there is only one, Mesotipula (Irenatipula) sigmoidea Lukashevich, 2009 [56], with a very short section of vein R 4+5 which seems to be an example of convergence in plesiomorphic furcation of the radial sector ([58]: Figure 3B).
In Gonomyia (subfamily Chioneinae) three branches of Rs reach wing margin, R3, R4, and R5; r-r is absent, probably fused with R3 ([36]: Figure 2; therein also extensive discussion on reduction in radial field in this genus).

Conclusions
The Drinosinae subfam. nov. shows a unique venation of wing, in which the extreme reduction of radial veins is combined with a complete set of four medial veins. This venation pattern results in an amazing appearance of the wing: the medial field is more densely set with veins than the radial field. The remaining six subfamilies of the Limoniidae show either the reverse combination, or both fields are more or less equally armoured with veins ( Figure 8). The weak support of the radial field with veins in Drinosinae may indicate some unknown properties of the biomechanics of their flight.

Conclusions
The Drinosinae subfam. nov. shows a unique venation of wing, in which the extreme reduction of radial veins is combined with a complete set of four medial veins. This venation pattern results in an amazing appearance of the wing: the medial field is more densely set with veins than the radial field. The remaining six subfamilies of the Limoniidae show either the reverse combination, or both fields are more or less equally armoured with veins ( Figure 8). The weak support of the radial field with veins in Drinosinae may indicate some unknown properties of the biomechanics of their flight.
The two genera, Drinosa and Decesia, that have exceptional wing venation deserve placement into a new subfamily, Drinosinae, and contribute to our understanding of the systematics and evolution of the Limoniidae. This particular venation is known only from one place and time: Cretaceous Burmese amber, and apparently did not survive to recent times.