A new species of Dasyhelea (Diptera: Ceratopogonidae), mining the leaves of the floating fern Salvinia minima Baker

ABSTRACT A newly described species, Dasyhelea mesophylla Dominiak and Borkent, from Costa Rica is the first species in this large and diverse family known to be a leaf-miner. The species is described as third and fourth instar larvae, pupa, and male and female adults. The biology of larvae mining the floating leaves of Salvinia minima Baker is also described. The pupa has a distinctive, sharply pointed respiratory organ, shared with some other species of Dasyhelea, which pierces the surface of the leaf to breathe. The species belongs in the grisea group and its close relatives within that group are noted.


Introduction
One of the problems of being taxonomists working on such a large group as biting midges, with 6274 species now recognised (Borkent and Dominiak 2020;Borkent et al. 2022) and so many more yet to be named, is that more observations and discoveries accumulate during field work than it is possible to publish on. In 1993 the second author discovered the first instance of leaf-mining in Ceratopogonidae, of larvae of a species of Dasyhelea Kieffer mining the leaves of the floating aquatic fern Salvinia minima Baker in Costa Rica. After 30 years of sitting in storage, this new Dasyhelea midge is finally being described; we also provide this new Dasyhelea midge and providing the first detailed report of its leaf-mining behaviour, which is unique within the family Ceratopogonidae. The species and its exceptional feeding habit was briefly mentioned by Borkent and Craig especially of the family Apiaceae (Szadziewski et al. 1997), and they are known as pollinators of the cacao tree Theobroma cacao L. and rubber tree Hevea brasiliensis Muell. Arg. (Lee et al. 1989). Mouthparts are similarly developed in both sexes (Carter et al. 1921;Glukhova 1981), suggesting that females and males can actually feed on the same substrates. Unlike most other Ceratopogonidae, there are no vertebrate biters, ectoparasites or predators among Dasyhelea and all females lack biting mouthparts. Various species feed on nectar, sweet secretions of plants, or honeydew Szadziewski et al. 1997). However, some laboratory observations indicate that imagines of Dasyhelea do not need to take a meal at all (Zilahi-Sebess 1931).
The intrageneric classification of the genus Dasyhelea is problematic. Since comprehensive phylogenetic studies on this group of biting midges have not been undertaken, currently it is impossible to propose a natural subgeneric system. However, five subgenera arranged by Remm (1962Remm ( , 1979 Remm). The New World's species are generally arranged in various species groups. Four of them, namely the cincta group, grisea group, mutabilis group and traverae group, were established by Wirth (1952). Later, many new species groups were proposed: the leptobranchia group , for some species included previously in the traverae group (see Wirth 1952); the borgmeieri group ; and the brevicornis group (Grogan and Wieners 2006). Recently, Díaz et al. (2010) proposed the patagonica group for a few southern Neotropical species.
Comparison of the two systems, subgenera vs species groups, shows that the cincta group and the holosericea group are equivalent to the subgenus D. (Sebessia) (proposed by Remm 1979), the grisea group corresponds to the subgenus D. (Dasyhelea), and the mutabilis group together with the borgmeieri group corresponds to D. (Pseudoculicoides) (Remm 1962;Waugh and Wirth 1976).

Materials and methods
Specimens were examined, studied, measured and drawn using a Wild M3 dissecting microscope and a Zeiss Jenaval compound microscope. Photomicrographs were taken with a Canon EOS Rebel T3i camera through the Zeiss Jenaval compound microscope.
Terms for adult structures follows those used in the Manual of Afrotropical Diptera (Borkent 2017); and for pupae, only features differing from the generic description by Borkent (2014) are given here. Larval terms follow Díaz et al. (2019) and Fürst von Lieven (1998).
The abbreviations used to represent the museums cited in this text are as follows: CNCI

Diagnosis
The only Dasyhelea in the Neotropical Region with the following characteristics. Male adult: with a single elongate radial cell, sternite 9 straight, paramere sinusoidal, twisted and tapering to a narrow apex, fused basally with gonocoxal apodemes and forming with them an asymmetrical structure, aedeagus symmetrical with dark, well-developed anterolateral club-like projections and 2 well separated and slender posterior projections with apices that are hooked dorsolaterally. Female adult: with 1 elongate radial cell, frontal sclerite broader than long, with sternite 9 elongate anteriorly, conical and with a rounded apex. Pupa: with the respiratory organ curved and coming to a sharp point, with a double row of about 7 circular pores each (so about 14 total) restricted to the apical 0.3 of the respiratory organ, and with a slightly elongate, pointed, gradually tapering terminal process bearing 2-3 pointed tubercles laterally near its base. Larva: not presently diagnosable, but see taxonomic discussion.

Distribution and bionomics
Dasyhelea mesophylla is known only from the type locality on the west coast of Costa Rica at an altitude of about 5 m. Adult specimens were reared from larvae and pupae present in S. minima leaves (Figure 8) floating in a lagoon on the northern margin of Carara National Park, about 1.5 km east from highway 34 along the hiking trail. The lagoon from which this species was collected in 1993 is periodically flooded by the abutting and extremely large Tárcoles River. The larvae and pupae were common but no specific data were taken in this regard. Virtually every clump of leaves had at least one leaf being mined (or that had been previously mined). Leaves were generally in clumps of 6-12 leaves. Borkent and Craig (2001) described the pupa of Stilobezzia rabelloi Lane, which have piercing respiratory organs to obtain oxygen from the dangling 'roots' (actually modified leaves), from the same habitat as D. mesophylla.
Observations of the mining larvae and of pupae in the S. minima leaves were made under the dissecting microscope as follows. Several fourth instar larvae were observed in    some detail, although many were otherwise present. The larvae actively mined the leaves, scraping at live fern tissue, had green material in their guts and produced green excrement. The narrow labrum and anterolateral margins of head capsule extending anteriorly likely facilitate the scraping of plant tissue. The mines of at least fourth instar larvae produced a bulge on the upper surface of the leaves. The frass produced by larvae was either in the mine or on the surface of the leaf. At least some larvae periodically leave the mine for the surface and then burrow back into the leaf. A few larvae were observed to move from one leaf to another. Possibly this is a means of ensuring that overcrowding does not limit any given larva which can probably move from one abutting leaf to another. One larva was observed feeding on the interior of the leaf but with about 2/3 of its body on the surface of the leaf. Another larva was seen crawling on the surface, then gnawing through the surface to the interior and, after feeding for some time, drawing the rest of the body into the leaf. Generally, when a leaf was mined, there were only 1-2 fourth instars present (aside from other earlier instar larvae), but one instance of three larvae in a single leaf was observed.
Several third instar larvae were observed mining the interior of the leaves, in an identical manner to the fourth instar larvae, but no third instars were seen on the surface of the leaves. One second instar larva was observed in a tunnel just slightly wider than its body, winding through the leaf. The mines of third and fourth instar larvae were more excavated in places, connected by swollen (leaf surface bulging at least dorsally) tunnels about the width of those larvae. First and second instar larvae seemed to be concentrated on the peripheral margins of the leaves, while third and fourth instar larvae were generally in the main body of the leaves.
On the other hand, several leaves were observed to be partially mined but without larvae, suggesting the larva had died or, more likely, had moved on to another leaf to mine there.
Two larvae were observed to pupate. One pupa was entirely encased within the mine with no direct opening to the surface, other than the insertion of its respiratory organs through the surface cuticle of the leaf. Another pupa had cut an opening near the anterior end of the pupa, but without the respiratory organs protruding. Observed again later, the pupa had stuck one respiratory organ up through the leaf cuticle, indicating the pupa only needs to obtain oxygen periodically from the surface (or can withdraw for some time if disturbed). Both pupae had their fourth instar exuviae present nearby (Figures 4(a), 8(c-d)). A number of other pupae were observed with either one or both respiratory organs protruding from the leaf.
Two emerging adults in the laboratory had free-floating pupal exuviae. The one pupa noted above that had been entirely encased by the surface of the leaf had made a jagged exit hole in the leaf, and the other appeared to escape through the hole already present. Some leaves were found with pupal exuviae with part of their abdomens still in the leaves.
It is worth noting that Forno and Bourne (1984) previously recognised an unnamed Dasyhelea as 'phytophagous' on Salvinia molesta D.S. Mitch. in Brazil. Another species, called Dasyhelea sp. 3 (grisea group), was reported by Torreias et al. (2013) from southeastern Brazil, where its 'immature stages were collected in S. auriculata' Aubl (6). However, considering the sampling method used, the immatures (pupae only?) were most probably found among rhizomes and submerged, root-like leaves of Salvinia. Three males and one female were obtained but no details regarding their morphology are given by the authors . Pelli and Barbosa (1998) mentioned D. paulistana Forattini and Rabello as present on S. molesta in Brazil, but they did not assign this species to taxa feeding on Salvinia or to taxa causing harm to it. The latter Dasyhelea species together with D. pseudopollinosa Díaz and Ronderos were collected from mats of S. auriculata and Azolla filiculoides Lam. in Brazil and Argentina (Díaz et al. 2014).

Taxonomic and phylogenetic discussion
This new species belongs to the grisea species group or subgenus D. (Dasyhelea) (Dominiak 2012) if following the subgeneric division proposed by Remm (1962Remm ( , 1979. The presence of the strongly modified pupal respiratory organ of D. mesophylla is shared by a number of species in the grisea species group, and because it is unique within the family (Borkent 2014), we consider this a synapomorphy of these species. As such, from the limited number of Dasyhelea species known as pupae (Borkent 2014 Remm (syn. D. lugensis Brodskaya) from Europe, and two unnamed species from Indonesia (Sumatra, listed as Holoconops sp. in Mayer 1934b) and Australia (A. Borkent, pers. obs.), respectively. All of them have virtually identical respiratory organs; these are thick at the base to about midlength and taper to a sharp apex. The respiratory organs of the pupae of D. mesophylla are somewhat more elongate than those of the other abovementioned species of Dasyhelea. Although not known as immatures, male D. unicolour Remm and D. stackelbergi Remm appear very similar to those in this group of species and may be closely related.
The pupae of D. mesophylla stick their respiratory organs through the upper surface of the mined S. minima leaves to obtain oxygen from the exposed surface of the leaf, suggesting that these other species also use their respiratory organs in a similar manner. However, because at least the larvae of D. traverae do not appear to mine leaves (Thomsen 1935(Thomsen , 1937Waugh and Wirth 1976) it is also possible that Dasyhelea species with such respiratory horns obtain air from water plants, similarly to S. rabelloi.
The piercing respiratory organs of pupae of a group of Stilobezzia Kieffer is clearly separately evolved, as evidenced by differences in the details of the respiratory organs compared with Dasyhelea as well as the phyletic distance between the two groups (Borkent and Craig 2001).
In spite of the fact that species of Dasyhelea are common throughout much of the world, only 62 out of 628 known species have been described as larvae (Díaz et al. 2013;Borkent 2014Borkent , 2016Borkent , 2018Borkent , 2019Duan et al. 2019;Borkent and Dominiak 2020;Lu et al. 2020;Borkent et al. 2022). One reason is probably that larvae of Dasyhelea are generally morphologically conservative. However, the very narrow protruding labrum and corresponding anterolateral projections of the head capsule, likely important for scraping leaf material free to ingest, may be distinctive among those larvae which have been described so far, keeping in mind that, of the species with a piercing respiratory organ, only the larvae of D. caesia (Brodskaya 1995), D. chani (Wirth and Linley 1990) and D. traverae (Thomsen 1937) have been described. It is important to recognise that most descriptions of Dasyhelea larvae are rather superficial. Detailed comparative studies are needed of the significant differences in larvae between the numerous species of this genus. There are at least some marked differences exhibited by some taxa (eg those species living in Nepenthes pitcher plants in Southeast Asia which have extremely long, narrow head capsules ;Tokunaga 1961;Wirth and Beaver 1979). None of the Dasyhelea species described from larvae or pupae other than D. mesophylla are known to be leaf-miners.

Derivation of specific epithet
The name mesophylla refers to the parenchyma between the epidermal layers of a leaf, the tissue where the larva and pupa of the new species have been found. Forno and Bourne (1984) recorded an unidentified species of Dasyhelea as 'phytophagous' on Salvinia. They noted its presence only on S. molesta that they studied in southeastern Brazil. Torreias et al. (2013) recorded the presence of an unnamed species of Dasyhelea in the grisea group from Lago Grande, Ilha da Marchantaria, Iranduba, Amazonas, Brazil, as immature stages were collected in S. auriculata. This suggests that there may be more leaf-mining Dasyhelea species or that perhaps these two records concern D. mesophylla which possibly is more broadly distributed in the Neotropical Region.

Conclusions
The discovery here of leaf-mining by a species of Dasyhelea, combined with the wide array of other, generally small and restricted habitats inhabited by the other 105 species known as immatures (Borkent 2014;Brahma et al. 2016;Díaz et al. 2018;Duan et al. 2019;Lu et al. 2020), suggests that there are likely further interesting, specialised habitats yet to be discovered among the remaining named and many unnamed species in this genus.
Because species of Salvinia can be pestiferous, occurring in huge numbers in aquatic habitats, a number of authors have examined the possibilities for biological control but have focused on larger insects (Bennett 1966;Forno and Bourne 1984;Julien 2012;Martin et al. 2018;Maseko et al. 2019;Wahl and Diaz 2019). Although their impact did not seem great, it may be valuable to study the effect of D. mesophylla on the productivity of floating ferns.