New taxa, including three new genera show uniqueness of Neotropical Nepticulidae (Lepidoptera)

Abstract After finding distinct clades in a molecular phylogeny for Nepticulidae that could not be placed in any known genera and discovering clear apomorphic characters that define these clades, as well as a number of Neotropical species that could be placed in known genera but were undescribed, three new genera and nine new species are here described from the Neotropics: Stigmella gallicola van Nieukerken & Nishida, sp. n. reared from galls on Hampea appendiculata (Malvaceae) in Costa Rica, representing the first example of a gall making Stigmella; Stigmella schinivora van Nieukerken, sp. n. reared from leafmines on Schinus terebinthifolia (Anacardiaceae) in Argentina, Misiones; Stigmella costaricensis van Nieukerken & Nishida, sp. n. and Stigmella intronia van Nieukerken & Nishida, sp. n. each from a single specimen collected the same night in Costa Rica, Parque Nacional Chirripó; Stigmella molinensis van Nieukerken & Snyers, sp. n. reared from leafmines on Salix humboldtiana, Peru, Lima, the first Neotropical species of the Stigmella salicis group sensu stricto; Ozadelpha van Nieukerken, gen. n. with type species Ozadelpha conostegiae van Nieukerken & Nishida, sp. n., reared from leafmines on Conostegia oerstediana (Melastomataceae) from Costa Rica; Neotrifurcula van Nieukerken, gen. n. with type species Neotrifurcula gielisorum van Nieukerken, sp. n. from Chile; Hesperolyra van Nieukerken, gen. n.. with type species Fomoria diskusi Puplesis & Robinson, 2000; Hesperolyra saopaulensis van Nieukerken, sp. n., reared from an unidentified Myrtaceae, Sao Paulo, Brasil; and Acalyptris janzeni van Nieukerken & Nishida, sp. n. from Costa Rica, Guanacaste. Five new combinations are made: Ozadelpha ovata (Puplesis & Robinson, 2000), comb. n. and Ozadelpha guajavae (Puplesis & Diškus, 2002), comb. n., Hesperolyra diskusi (Puplesis & Robinson, 2000), comb. n., Hesperolyra molybditis (Zeller, 1877), comb. n. and Hesperolyra repanda (Puplesis & Diškus, 2002), comb. n. Three specimens are briefly described, but left unnamed: Ozadelpha specimen EvN4680, Neotrifurcula specimen EvN4504 and Neotrifurcula specimen RH2.


Introduction
The nature of Latin America is generally known to be both very diverse and peculiar in its composition and Lepidoptera are no exception to this phenomenon. Biogeographically the region is usually known as the Neotropical region, or the Neotropics, although especially botanists often exclude the southern -non tropical -part (Antonelli and Sanmartín 2011;Morrone 2014). Probably more than one third of the globally named Lepidoptera species can be found in the Neotropics (Heppner 1991), although the number given by Heppner very likely is still a huge underestimate (Kristensen et al. 2007). The peculiarity is evident from many of the phylogenetically lower Lepidoptera, the non-Ditrysia, for example the endemic jaw moth family Heterobathmiidae in Patagonia, the endemic Andesianidae, the Palaephatidae, shared only with Australia, amongst which the supposed sister group to the Ditrysia is to be found and other rare southern families such as Cecidosidae and Neopseustidae (Nielsen 1985;Regier et al. 2015;Bazinet et al. 2016). Most of the families with typically leafmining larvae also include endemic New World or Neotropical groups or radiations: in Heliozelidae the basal Patagonian genus Plesiozela Kristensen, 2003 (Karsholt andKristensen 2003), in Tischeriidae a radiation of the New World endemic Astrotischeria Puplesis & Diškus, 2003, specialised on Asteraceae (Diškus and Puplesis 2003) and in Opostegidae the endemic genera Notiopostega Davis, 1989 and Neopostega Davis & Stonis, 2007(Davis 1989Davis and Stonis 2007). Only the leafmining family Nepticulidae did not have endemic Neotropical genera.
Until fairly recently the nepticulid fauna of the Neotropics was almost completely unstudied, and twelve of the species described between 1877 and 1962 were tentatively placed in the large global genus Stigmella Schrank, 1802, with just a single species in the -at the time -monotypic genus Enteucha Meyrick, 1915 (Davis 1984). Collecting and rearing leafmines was rarely done in Latin America, with the notable exception of Bourquin (1961) who described the first two South American Asteraceae feeding nepticulids. During their expeditions to Patagonia in the 1970's and 80's both the Danish researchers Ebbe Nielsen and Ole Karsholt and Don Davis from the USA collected Nepticulidae, mostly adults, and other non-Ditrysian moth families (Nielsen 1985), but this material remained unstudied for some time. A research team led by the Lithuanian researcher Rimantas Puplesis (later named Jonas R. Stonis) started fieldwork in Latin America in the late 1990's and published a revision of Neotropical nepticulids early this century, bringing the total to 74 species in seven genera by 2002 (Puplesis and Robinson 2000;Puplesis et al. 2002aPuplesis et al. , 2002b, including several species only known from southern Florida. Since then they continued with papers describing 40 new species, bringing the total to 114 in July 2016 (Šimkevičiūtė et al. 2009;Stonis et al. 2013aStonis et al. , 2013bStonis et al. , 2013cStonis et al. , 2014Diškus and Stonis 2014;Remeikis et al. 2014;Remeikis and Stonis 2015;Remeikis 2015, 2016). These revisions continue, fieldwork continues and collections in Copenhagen, Washington and Vilnius still contain unnamed material (J. R. Stonis personal communication).
Even though almost all Neotropical species up to now have been placed in known nepticulid genera, which have been described from other parts of the world, several species show unique combinations of characters, and their placement is therefore disputable. Only the genera Enteucha Meyrick, 1915 and Manoneura Davis, 1979 were described originally from the Neotropics (respectively British Guyana and southern Florida), but Enteucha is not confined to the Neotropics, and also occurs in the Palearctic and Oriental regions ( van Nieukerken 1986;). The monotypic Manoneura was synonymised with Enteucha by van Nieukerken (1986). Puplesis and Robinson (2000) reinstated Manoneura as a good genus on the basis of autapomorphic characters, and described a second species, but it was synonymised again after it was placed with high confidence inside a monophyletic Enteucha clade in a molecular phylogeny with eight genes ; . Peculiar taxa include the species placed in Glaucolepis Braun, 1917, Fomoria Beirne, 1945 and several in Enteucha; also some species in Acalyptris Meyrick, 1921, particularly the A. latipennata group, show combinations of characters unknown in this genus elsewhere. Also the single Neotropical species in Ectoedemia Busck, 1907 (E. fuscivittata Puplesis & Robinson, 2000) has an aberrant morphology.
While working on the molecular phylogeny of global Nepticulidae, to be published almost simultaneously , the sampling initially lacked Neotropical material and EvN and CD thus attempted to obtain fresh material for DNA analysis through several sources. This comprised very limited collecting by EvN in 2000 and some material received from others, further KN had been actively collecting and rearing nepticulids in Costa Rica the last decades, and CS collected leafmines during his annual visits to Lima since 2010.
Even though we could still only include a limited set of material compared to the described diversity, the analyses based on DNA sequences yielded several surprises. Three clades with distinctly long branches in the phylogeny could not be placed in any known genus, and other Neotropical species that were studied and could be placed in known genera also appeared to be hitherto unnamed. Taxonomists working on Nepticulidae fortunately have the tradition to be reluctant in describing new genera, preventing the systematic chaos that has been evident in various other groups of Lepidoptera. It is only after careful consideration of the morphological and molecular evidence that we here name these genera as well as several species, of which the common denominators are Neotropics and availability of DNA sequences (minimally a DNA barcode), and often provide additional interesting information on biology.
We describe three new genera and nine new species, including the first nepticulid species from Brazil and Costa Rica and the first known gall-forming Stigmella species.

Material
Material was either collected by one of the authors or loaned from various institutions. The Costa Rica material was collected by Kenji Nishida, and made available as loan from the Museo de Zoología, Escuela de Biología, Universidad de Costa Rica (MZUCR), allowing for further study, including DNA analysis. Some of the specimens, including all holotypes were later donated to Naturalis (RMNH), because of its large collection of global Nepticulidae. Holotypes of the species from Brazil, collected by Erik van Nieukerken and Peru, collected by Chris Snyers are deposited respectively in DZUP and UNALM. On the basis of DNA barcodes, analysed by us, we could match several specimens from the global Malaise trapping program to three of our new species, for two of these we borrowed some material from BIOUG, stored in ethanol 96%, to check the genitalia. We also cite the other specimens from the Malaise trappings with matching DNA barcodes that we did not study. The Argentinean samples from this program are returned to MACN and the Costa Rican samples are donated to RMNH. We select a Holotype of each species, but refrain from designating and labelling paratypes, as these have no name-bearing function in Zoological Nomenclature (International Commission on Zoological Nomenclature 1999). All material is also listed in Suppl. material 1.

Methods
Collecting and rearing methods varied depending on the collector, but generally followed commonly used methods for the collection of leaf miners.
Genitalia were prepared according to our standard procedures, usually including DNA extraction, and described earlier in detail ( van Nieukerken 1985;van Nieukerken et al. 2010). Wings were denuded in ethanol 70% and stained with phenosafranin before embedding in euparal.
Measurements. Measurements of genitalia were obtained from digital images, using calibrated scaling in the Zeiss AxioVision software, see below, 20× objective for male genitalia and 10× or 20× for females. Capsule length was measured from vinculum to middle of uncus; valva length from tip of posterior process to ventral edge, excluding the sublateral process; phallus length was measured from the sclerotized tube, from tip of ventral process/carina, excluding any protruding vesica parts. Total bursa length includes all of the internal genitalia from cloaca to anterior edge of bursa. Genitalia measurements are usually rounded off to nearest 5µm. Forewing length was measured from tip of fringe to attachment on thorax, usually at magnification of 20×, also preferably on photographs. Antennal segment counts include scape and pedicel; they were counted on photographs or directly using the stereo microscope.
Morphological terms. The terminology used largely follows our earlier treatments and other recent papers on Nepticulidae ( van Nieukerken 1986;van Nieukerken et al. 1990), but is slightly adapted to follow Kristensen's (2003) detailed morphological treatment. We thus use phallus rather than aedeagus for the male intromittent organ, and adopt Wootton's (1979) venation nomenclature, meaning that R2-5 become Rs1-4 and Cu changes into CuA. See also Doorenweerd et al. (2016).
Photographs of mounted moths were made with an AxioCam digital camera attached to a motorized Zeiss SteREO Discovery V12, using the Module Extended Focus, Zeiss AxioVision software, to prepare a picture in full focus from a Z-stack of 10 to 35 individual photos. Leafmines were photographed by EvN with a similar camera on a Zeiss Stemi SV11 stereo-microscope, without extended focus; other illustrated leafmines are from field photographs. Genitalia and wing slides were photographed with a similar camera on a manually operated Zeiss Axioskop H, usually with just a single exposure. Photographs were edited with Adobe Photoshop, avoiding any change to the real object, but backgrounds are cleaned from excess debris and artefacts by using healing brush and clone tools; tone and contrast are adjusted and a little sharpening is used in some photos. Photographs of venation, some mines and some female genitalia were taken in sections, and combined with the photomerge tool in Photoshop.
The photographs in this paper were taken by: Figs  Molecular methods were described in detail in previous studies (Doorenweerd et al. 2015;Doorenweerd et al. 2016). The DNA barcode data and GENBANK numbers are given both in Suppl. material 1 and in BOLD dataset DS-NEONEP (doi: 10.5883/DS-NEONEP).  chamber narrow, located longitudinally in central part of gall, and line with compacted dark brown frass on lower part of chamber, part of the frass reaching towards nectary gland (Figs 26,27). Galls were found on ca. 2 meter-tall treelets or large trees of 8-25 m (n=7), however we were unable to examine higher parts of large trees. Most of the collected galls were empty, having a more or less rectangular-shaped exit slit of less than 1 mm wide (Fig. 30). We found inquiline phorid fly larvae (Diptera: Phoridae) inhabiting old gall chambers. The gall was already recorded by Hanson et al. (2014).
Cocoon (Fig. 31). Pale brown to brown, oval, exit slit side is broad and opposite side tapered (apple seed-shaped), 2.8-3.2 mm long, 1.7-1.9 mm wide (n=2). Under rearing conditions inside plastic bags, cocoons were constructed on either host leaf surface or on paper towels (n=5). Under natural conditions in the field, no cocoons were found on host plant leaves or stems near the galls.
Voltinism and habits. Larvae were found inside the galls in March, May and June. Adult emergence has been recorded in January, April, May, June and October. Parasitoids. Braconidae: Adelinae: Adelius sp., endo-parasitoid, koinobiont of host larva and pupating inside host cocoon. It was reared from the Alto Masis Station site (n=2). Adelinae are specialised parasitoids of Nepticulidae (Yu et al. 2011).
Distribution. Costa Rica: Alajuela, Heredia and Puntarenas provinces. Galls have been recorded mostly from Caribbean/Atlantic slope in lowland Costa Rica. The type locality is near the continental divide on the Pacific slope.
DNA barcode. The Holotype (RMNH.INS.23672) and one female (RMNH. INS.23739) are barcoded, with less than 1% difference. Barcode Identification Number BOLD:ACG9386. The female was also sequenced for other genes and used in the molecular phylogeny . Sequences may be retrieved in BOLD and Genbank under voucher/sample ID RMNH.INS.23739.
Remarks. The genitalia of Stigmella gallicola resemble somewhat those of the S. betulicola group, but in our multi-gene molecular analysis it groups in the clade which contains the S. prunifoliella and S. ultima groups, without a strongly supported position.
There are only few cases of galling in Nepticulidae known, and this species is the first known in the genus Stigmella. Galling has evolved independently several times in Nepticulidae: the North American Ectoedemia populella Busck, 1907 makes a petiole gall (Busck 1907;Wilkinson and Scoble 1979), but starts with a short leafmine along the midrib. Its behaviour caused Busck to name the new (then monotypic) genus Ectoedemia (= making an external swelling), but only a few species feeding on Populus have similar habits and all other species in the genus are leafminers. The European Trifurcula pallidella (Duponchel, 1843) makes a spindle shaped gall in the stem of Cytisus species (Fabaceae) (van Nieukerken et al. 2004), but this gall growth appears to be caused by the intense mining activity in a small portion of the stem. Muhabettana nigrifasciata (Walsingham, 1908) of the Canary Islands starts with a small gall on a vein, and later the larva makes a long leafmine in the leaf of Periploca laevigata (Apocynaceae) (Klimesch 1972); Klimesch uses the German word "Gallenmine". Ectoedemia castaneae (Busck 1913) was according to Busck reared from small galls on young twigs, but this is probably not a real gall-former, but an effect of a spiral barkmine in a thin branch: the species is synonymised with the barkminer Zimmermannia bosquella (Chambers, 1878) (van Nieukerken et al. 2016). Of all these species, Stigmella gallicola induces and forms the most conspicuous swellings, and its biology is adapted to live inside the gall chamber. Gall-forming nepticulids could have been overlooked, and further studies in gall-inducing Lepidoptera may reveal more species. It is worth mentioning that larvae of Rhodoneura cf. terminalis (Walker, 1865) (Thyrididae) induce spindle-shaped galls on the stem apex of Hampea appendiculata. They were reared at Estación Biológica Monteverde and OTS La Selva Biological Station (K. Nishida, personal observation).
Etymology. The specific name is a noun in apposition, derived from the Latin noun galla (= gall) and suffix -cola, "dweller in".
Other material examined. 6♂, 3♀, 5 adults, galls, larvae. Differential diagnosis. A completely leaden coloured species, with orange head and leaden edged scape, collar also leaden. The genitalia do not resemble any other Neotropical Stigmella species closely.
Female genitalia . Length of bursa ca 340 µm. T8 apically pointed, with ca 15-20 setae total. Posterior apophyses longer than anterior ones. No sclerotisations or signa observed in the single relatively poor genitalia slide.
Leafmines . The mine is a much contorted upper surface gallery, track often doubling back, usually confined to the small space between two lateral veins and the midrib; mine filled with black frass; larval exit hole on upperside. Mine poorly visible in transmitted light, or from underside, due to leaf thickness.
Egg. The egg is always deposited on leaf upperside, frequently against a lateral vein. Larva: green. Voltinism and habits. Larvae were collected in late August, adults emerged indoors in September, and were caught in malaise traps from late September to late October. We collected the mines on planted trees, together with large numbers of the then still undescribed Leurocephala schinusae Davis & McKay (Davis et al. 2011).
Distribution. Argentina: Misiones. DNA barcode. We barcoded two specimens from our reared series (not the Holotype), that appeared to match with a large number of records from the Malaise traps in  Misiones, giving a total of 26 barcodes, all in BIN BOLD:ACN0764. One specimen was also sequenced for other genes and will be used in a forthcoming analytical paper. Sequences may be retrieved in BOLD and later also in Genbank under voucher/sample ID RMNH.INS.24681.
Remarks. Molecular analysis suggest that S. schinivora is closely related to the North American Anacardiaceae feeding species in the Stigmella prunifoliella group, which suggests a single host-shift from Rosaceae or Rhamnaceae, but in some analyses it also groups with the European S. diniensis (Klimesch, 1975) that feeds on Cistaceae. It is currently the only known Neotropical species of this group of species.  Davis et al. (2011) described four gracillariid leafminers of the Brazilian pepper tree, partly aiming at finding suitable candidates to release as biological control against the invading pepper tree in Florida. They did not report on any nepticulid, but we assume that S. schinivora is also widespread with the host. It could be added to the list of potential control candidates, but we doubt its effectiveness, given the small size. Also the fact that several Anacardiaceae miners are not very host specific is a risk, as released control species might shift to native North American Rhus and Toxicodendron species.
Other DNA barcode. Holotype BIN: BOLD:ACG8765. The holotype was also sequenced for three other genes and will be used in a forthcoming analytical paper. Sequences may be retrieved in BOLD and later also in Genbank under voucher/sample ID RMNH.INS.24037.
Remarks. This species and the next were collected the same night. They both belong in core Stigmella, in the clade with the S. lapponica, S. rhamnella and S. sanguisorbae groups, but do not belong to any of these groups.
Etymology. Costaricensis: an adjective, derived from the country name Costa Rica and the suffix -ensis, indicating geographical origin. Differential diagnosis. Externally S. intronia and S. costaricensis are very similar, but the fascia in intronia seems a bit wider and is placed more anteriorly. Both species resemble somewhat the North American S. slingerlandella (Kearfott, 1908). The species are best separated by the shape and spinosity of the valva and form of uncus and gnathos.
Biology. Host plants. Unknown. Voltinism and habits. The moth was collected in February at a light sheet. Distribution. Costa Rica: San José Province: Chirripó National Park: Llano Bonito area, a cloud forest surrounded by large oak trees.
DNA barcode. Holotype BIN: BOLD:ACG8514. The holotype was also sequenced for other genes and used in the molecular phylogeny ); here we discovered the presence of several introns in a copy of the gene Elongation Factor 1α. Sequences may be retrieved in BOLD and Genbank under voucher/ sample ID RMNH.INS.24036.
Remarks. See under S. costaricensis. Etymology. Intronia, a noun in apposition, arbitrarily derived from the word Intron (based on English: intragenic region), because of the presence of several introns in a copy of the gene Elongation Factor 1α .
Female genitalia (Figs 54-57). Total length bursa ca 730 µm. Abdominal tip rather narrow, but not pointed; T8 with 5-7 setae on either side, and a small patch of many small setae anteriorly; anal papillae without setae. Bursa covered with small pectinations, with a band of larger and stronger sclerotised ones, in 4 rows, running around bursa longitudinally; ductus spermathecae straight, without convolutions.
Biology. Host plant (Figs 58, 59). Salicaceae: Salix humboldtiana Willd., a small tree. Leafmines (Figs 60-63). Mine first a narrow gallery, usually running towards leaf edge and leaf apex and not doubling back, occasionally running to leaf base and then often doubling back and ending towards apex. Later mine a wide irregular gallery, almost full depth. Frass black, in a narrow central line. larval exit hole on leaf upperside.
Egg. Always deposited on leaf upperside close to the midrib. Larva yellow. Voltinism and habits. Larvae were always present between December and late January. Adults usually emerged after two to three weeks after the cocoon was spun, suggesting multiple generations. The species has only been searched for between early December and late January, we have no information on other seasons. Mines were always found at the shady side of the trees.  Remarks. This species clearly belongs to the Stigmella salicis group s.str., that is widespread in the Holarctic region and of which all but one species feed on Salicaceae (sensu stricto). However, morphologically it differs by the presence of numerous small cornuti in the phallus, whereas all other species have a reduced number of differently shaped cornuti; the latter thus is a good apomorphy for the Holarctic members of the group. The female shares the characteristic apomorphy: a band of signa around the bursa copulatrix. In our molecular phylogeny ) S. molinensis is sister to all Holarctic species of the group. The other Neotropical species, included in the salicis group by Puplesis and Robinson (2000) are excluded by us and placed in the Stigmella epicosma group of which probably many feed on Asteraceae. Since the host plant, Salix humboldtiana, and ten other native Salix species (Alford and Belyaeva 2009) are widespread in South and Central America, we expect that S. molinensis and possibly related species are more widespread. Male genitalia. (Figs 70-73). Vinculum lateral arms articulate with sides of tegumen; ventral plate expanded, not bilobed. Tegumen band shaped. Uncus variously bilobed. Gnathos an inverted V. Valva more or less triangular, transtilla with transverse bar present or weakly sclerotised, sublateral processes absent or short. Phallus relatively short, without distinct carinal processes; cathrema a normal striate thickening around base of ejaculatory duct, vesica with variable number of cornuti.
Female genitalia (Figs 74-79). T8 with or without row of setae; no setose anal papillae. Anterior apophyses often broadened, posterior apophyses usually narrow, straight, and longer than anterior ones. Vestibulum folded, more strongly staining in Chlorazol Black, with indistinct or no sclerotizations; corpus bursae asymmetric, curved; wall completely devoid of spines or pectinations or with a group of numerous, large, blunt pectinations as in O. guajavae . Ductus spermathecae with 1-3.5 convolutions.  Puplesis & Diškus, 2002 share many characters with the type species: an almost identical venation (albeit without recognisable CuA in the drawing of E. guajavae) and very similar male genitalia. We therefore transfer both to this genus as Ozadelpha ovata (Puplesis & Robinson, 2000), comb. n. and O. guajavae (Puplesis & Diškus, 2002), comb. n. It is possible that some species associated with the Myrtaceae genus Myrceugenia from Patagonia (Nielsen 1985) belong here as well, although we observe some differences in the few specimens avail-able to us. The Vilnius group is planning to revise this group of species (J.R. Stonis, personal communication).
Etymology. Ozadelpha, a noun. The name is based on a combination of the Greek adelphe (αδελφη), meaning sister with the colloquial abbreviation "Oz" as often used for Australia. This to indicate the sister group relationship between Ozadelpha and the Australian clade with Roscidotoga Hoare, 2000, Pectinivalva Scoble, 1983, Menurella Hoare, 2013and Casanovula Hoare, 2013. The name is to be treated as feminine.
Remarks. The series reared from Conostegia that KN reared and sent via the Museo de Zoología, Universidad de Costa Rica, to EvN was at first considered to constitute a single species. We were only able to successfully amplify several genes from one large female (RMNH.INS.24680) that on closer inspection appeared to be different from the rest of the series, even though the differences externally are small, apart from the size. The position of the genus as sister of the Australian genera was strongly supported on the basis of this specimen; however, we do not name that by lack of a male specimen and life history data. The few genes that we did amplify from other specimens show that both species belong to the same clade, and thus the same genus. In fact, there are probably more species feeding on different species of Conostegia throughout its distribution area; just before finalising this manuscript, Kenji Nishida reared another species from another species of Conostegia.
In the morphology Ozadelpha shows similarities with both Stigmella, Enteucha and the Australian genera. The venation resembles Stigmella, but also Pectinivalva, apart for the thickened A in forewing. In our molecular phylogeny Ozadelpha always groups with the Australian genera, either as sister to all of them together, or as sister to Roscidotoga . Where both Ozadelpha and Pectinivalva in the old sense feed on Myrtales, it is possible that the ancestral hosts were also Myrtales and that the ancestor was a rainforest inhabitant (Hoare and van Nieukerken 2013 silver fascia at 1/3, a second silver fascia at 2/3, often broken or narrowed in middle, at dorsum widened in both directions, silver scales along dorsum may reach other fascia, usually not; more distal silver scales usually separated from fascia. Hindwing with a narrow ochreous hairpencil inserted near frenulum of 1/3 hindwing length, hindwing scaling brown-grey. Leg upperside and abdomen dark fuscous, tarsi paler. Abdomen with indistinct grey anal tufts. Female (Fig. 90). Antenna with 19 segments (n=2). Hairpencil absent. Ovipositor broadly rounded.
Female genitalia (Figs 74-76). Total length of bursa ca 660 µm. T8 with row of ca 8 setae on either side, partly on distinct sockets; no setose anal papillae. Anterior apophyses broadly rounded, posterior apophyses narrow, straight, longer than anterior ones. Vestibulum more strongly stained, with indistinct sclerotisation; ductus bursae not demarcated from corpus bursae, corpus asymmetric, curved; wall completely devoid of spines or pectinations. Ductus spermathecae slightly curved, only one incomplete convolution at vesicle.
Biology. Host plants (Fig. 81). Melastomataceae: Conostegia oerstediana O. Berg ex Triana and C. pitierri Cogn., evergreen trees. Both species are distributed from Nicaragua to Panama, and recorded between 700 and 2400 m elevations in Costa Rica (Almeda 2007). C. oerstediana is one of the widespread trees in the mid-elevation cloud forest of Costa Rica (K. Nishida, personal observation).
Leafmines (Figs 82,83,88). Narrow zigzag linear mine, pale brown in colour, on upperside leaf (n=42). Mature mine approximately 50 mm long (n=42). Some mines were found along leaf veins, i.e. mines were angular or square (n=7). The mines were mostly found on mature broad leaves of small treelets of less than a meter tall (n=ca 30). We recorded from a single mine up to 20 mines per leaf, with an average of 5 mines/leaf (n=15). Central portion of leaf mines filled with black frass, deposited in zigzag arcs (Fig. 86). Exit hole on underside of leaf at tip of mine, ellipsoid, 0.6-0.9 mm wide (n=5). Mines were found on leaves with a length of ca. 7-20 cm (mean 15.5) and width of ca. 4-13 cm (mean 9.75) (n=20).
Cocoon (Figs 92, 93). Oval shaped, flat, double chambered, outer cocoon 2.3-3.3 mm long and 1.3-2.0 mm wide, inner cocoon 1.6-2.0 mm long and 0.8-1.3 mm wide, pale brown to brown (n=4). Exit slit side of cocoon slightly thinner and paler than opposite end. Under rearing conditions inside plastic bags, the larvae pupated on host plant leaf, next to leaf veins or vein grooves either on upper or underside leaf (n=20) (Figs 88, 92). A species of a solitary koinobiont-endo parasitoid wasp was found inside the inner chamber of cocoon (n=2).   Pupa (Fig. 91). General appearance of mature pupa flat and dark brown, 1.5-1.7 mm long (n=2).
Voltinism and habits. Old and young leaf mines were seen all year round on treelets found along trails in forest understory at the Estación Biológica Monteverde. The larvae pupate away from the mine and host plant under natural conditions; however, it is unknown where the larvae spin their cocoons except for a single cocoon that was found on the underside of a mined leaf, but this was parasitized by a parasitoid wasp. Late stage to mature mines collected on 2.iii.2016 produced adults on 28-29. iii.2016. Mines collected on 21.v.2016 produced mature pupae (n=2) and mature pupae of a parasitoid wasp inside the moth cocoon (n=2) on 7.vi.2016. Mines collected on 25.v.2016 produced 6 cocoons between 2-6.vi.2016. Two eggs on a single leaf collected on 7.vi.2016 produced very early mines of ca. 1.2 mm and 8.6 mm long on 19.vi.2016. White, circular, pellet-like micro objects were found inside mines surrounding mining larvae (n=2) (Figs 84, 85). The pellets are calcium oxalate crystals (druses) according to R. Kriebel (personal communication). Thus the larvae appear to avoid feeding on druses.
Distribution. Costa Rica: Alajuela, Guanacaste and Puntarenas Provinces. DNA barcode. We failed to produce a DNA barcode from the DNA extracts, but we obtained sequences for 28S and COII from the holotype, RMNH.INS.24506. They will be made available through GENBANK with another paper in preparation.
Remarks. Ozadelpha conostegiae represents the first published record of a Nepticulidae feeding on Melastomataceae. However, we also collected and reared unnamed Acalyptris species from the genus Melastoma from Australia: Queensland and Indonesia: Borneo. The genus Conostegia comprises 77 species of shrubs and trees in Central America, northern South America and the Carribean (Kriebel 2014).
Etymology. The epithet conostegiae is a noun in genitive case, derived from the generic name of the host plant Conostegia.

Ozadelpha specimen EvN4680
Differential diagnosis. Externally very similar to O. conostegiae, but markedly larger (forewing length 2.8 mm against 1.8-2.0). Female genitalia recognised by longer anterior apophyses and very short bursa.
Biology. Host plants. Unknown. The moth was found resting on the upperside of a Conostegia oerstediana (Melastomataceae) leaf along one of the main veins, thus this may well be the host.
Voltinism and habits. The moth was collected on March 2nd, 2012. Distribution. Costa Rica: Puntarenas Province. Remarks. We leave this species presently unnamed, as we only have a single female and no data on life history. There appear to be more closely related species of Ozadelpha feeding on Conostegia in Costa Rica, and we therefore rather await more material in order to be able to discriminate the various species better.
DNA barcode. We barcoded the single specimen, currently the only DNA barcode we obtained from the genus Ozadelpha. This specimen was also sequenced for other genes and used in the molecular phylogeny  Diagnosis. Neotrifurcula can be recognised by the hindwing venation with trifurcate Rs+M, and a very long and separate CuA in forewing, collar with hairscales, in the genitalia male phallus with a long curved flagellum-like appendix; female with reticulate signa and complex vaginal sclerite. Glaucolepis has a similar venation, but usually a velvet patch of special scales on hindwing and three pairs of anal tufts. (Figs 97-99). Medial to large nepticulid moths, forewing length 2.7-4.8 mm, largest over 10 mm wingspan. Head with collar comprising piliform scales; antenna with 42-58 segments in male (n=4), 47 in female (n=1). Forewing with distinct or less distinct fascia, sometimes metallic, no subdorsal retinaculum in male. Hindwing in male with costal bristles, no androconial scales observed. Venation (Fig. 95): very complete, with closed cell, R+Rs+M with 6 terminal branches: R, Rs1+2, Rs3, Rs4, M1 and M2, CuA separate and long, approaching Rs+M; A thickened; Hindwing broad, with 5 veins, Rs+M trifurculate: Rs, M1, M2. Abdomen: anterior part of sternum 2 with two lobes on posterior margin (Fig. 96). Tergum 8 with distinct anal tufts, tergum 3-7 with lateral groups of many setae and scales.
Biology. Hostplant and immature stages unknown. Adults collected in Nothofagus forests from November to January.
Distribution. Chile and Argentina, southern parts at low and medial altitudes. Composition. Next to the type species, we include two unnamed species, for one of which we only have one female that was sequenced, another one only one worn male on loan from Copenhagen (ZMUC). We find the DNA barcode distance too large to include the female in N. gielisorum and the male is much smaller and shows some differences in the genitalia. According to Jonas Rimantas Stonis (personal communication) there is a group of several closely related species of Neotrifurcula in Patagonia, estimated to comprise at least five species.
Etymology. Neotrifurcula, a noun, a combination of the prefix neo-, new, here derived from Neotropics, and the Latin noun Trifurcula (= a three-pronged fork), another nepticulid genus with a 3-forked Rs+M in the hindwing. The name is to be treated as feminine.
Remarks. The anterior sclerite of sternum 2 (S2A) has anterior apodemes similar to Bohemannia Stainton, 1859. The venation also has several similarities to Bohemannia, although the latter seems more reduced by the fusing of CuA with Rs+M and the reduction of the closed cell. This supports the possible sistergroup relationship to Bohemannia or Bohemannia + Hesperolyra that we found in our molecular analyses . Overall the species of Neotrifurcula resemble Glaucolepis in several ways: the venation is almost identical, the transverse bar of transtilla is absent, but Neotrifurcula does not have the male "velvet patch", the phallotrema spines. The flagellum-like appendix of the phallus is a remarkable character of as yet unknown function. This character requires further detailed morphological study. It is likely an apomorphy for the genus.

Neotrifurcula specimen EvN4504
Differential diagnosis. Externally similar to N. gielisorum, but no distinct fascia. Female genitalia characterised by ductus spermathecae with 3.5 convolutions, distinct reticulate signa and omega-shaped sclerotisation in vestibulum.
Biology. Host plant. Unknown. Voltinism and habits. The moth was collected in February. Distribution. Chile: Valparaiso. Collected in open shrubby habitat, with shrubby species of Nothofagus (Fig. 106).
DNA barcode. We obtained a full barcode of the specimen, with BIN BOLD:ACU6693, showing a distance to barcodes of N. gielisorum of ca 12%.
Remarks. The huge barcode distance to N. gielisorum and the differences in the wing pattern show clearly the species status of this specimen. It could possibly be the female of the next species, but difference in size and number of antennal segments make this rather unlikely.
Biology. So far the host plant is only known for one species, H. saopaulensis: Myrtaceae, on which the larva makes normal gallery mines (Figs 137-138).
Distribution. Only known from the Neotropics: Belize, Brazil, Colombia and Ecuador.
Composition. We recombine here the species previously placed in the Fomoria molybditis group: Hesperolyra diskusi (Puplesis & Robinson, 2000), comb. n., H. molybditis (Zeller, 1877), comb. n., H. repanda (Puplesis & Diškus, 2002), comb. n. and describe one new species, Hesperolyra saopaulensis van Nieukerken sp. n. One unnamed species also belongs here: Hesperolyra species 29122 (Puplesis & Robinson, 2000).  Etymology. The name Hesperolyra is a combination of Hespero-from Hesperus (the evening star), referring to the occurrence in the western hemisphere, and lyra (lyre), referring to the lyre shaped transtilla, a common character for species where the male is known. The name is to be treated as feminine.
Remarks. The present genus was recognised first in our molecular analysis by the new species H. saopaulensis, that consistently grouped outside any known genus, often together with Neotrifurcula, but at a large distance. Since we did not have a male of H. saopaulensis, a generic description seemed problematic, until we noticed similarity to the recently obtained DNA barcode of Fomoria diskusi and the unusual venation of both species, quite different from any other Fomoria. Fomoria diskusi was placed with some other species in the Fomoria molybditis group (Puplesis et al. 2002b), here treated as a synonym of Hesperolyra. Puplesis and Robinson (2000) placed H. molybditis and diskusi in Fomoria on the basis of superficial similarity of male genitalia, even though the venation is markedly different, resembling more that of Acalyptris. The authors even stated that "the unusually reduced wing venation lends additional support." The lack of apomorphies for Fomoria s. str. make assignment of any species to that genus difficult without molecular support, and even with eight genetic markers the support is not high  Differential diagnosis. Hesperolyra diskusi is easily recognisable by its striking pattern: the pale costal streak that turns into an outward oblique fascia at 2/3. In the male genitalia the characteristic long spines separate it from congeners. Redescription. Male (Fig. 117). Frontal tuft yellow to orange; scape shining yellowish white, antenna with 37-39 segments. Collar of hairscales similar to frontal tuft. Thorax and forewing fuscous; yellowish white pattern comprising a costal streak running from wing base to slightly over 1/2, then changing into an outwards oblique fascia running to 2/3 of dorsum, and becoming narrower towards dorsum; scales along costa stronger yellow; terminal cilia yellowish white beyond an irregular cilia line. Thorax anteriorly pale, posteriorly fuscous, in rest joining to the forewing pattern. Forewing under fold with a conspicuous row of special scales, only visible in descaled wings (Figs 126, 127).
Female genitalia (Figs 129-131). T9 forming broad pair of anal papillae with each ca 15 setae. T8 with rather square posterior margin, some setae on either side. Anterior apophyses narrow, posterior apophyses slightly widened, of about same length. Total length of bursa 460 µm. Bursa without signa or other ornamentation. Ductus spermathecae distinct and longer than corpus bursae, with 7 convolutions and ending in wide and pointed vesicle.
Biology. Host plant unknown. Collected at light in April, during the dry season in secondary forest (Puplesis and Robinson 2000).
Distribution Differential diagnosis. Externally recognised by silver to leaden basal half of forewing and postmedial silvery fascia, with cilia line present and slightly edged scape. Female genitalia simple, without signa.
Description. Male. Unknown. Female (Figs 119, 120). Frontal tuft pale yellow; scape white, slightly edged with grey, antenna with 23 segments (n=1). Collar of hairscales similar to frontal tuft. Thorax and basal half of forewing shining leaden to silver; followed by a slightly postmedial fuscous fascia, then a silver fascia at 2/3 and a fuscous wingtip; terminal cilia silvery white beyond cilia line. Hindwing pale. Abdomen with blunt tip.
Female genitalia (Figs 132-136). T9 forming broad pair of anal papillae with each ca 16-17 setae. T8 with truncate posterior margin, some setae on either side. Anterior and posterior apophyses narrow, posterior apophyses slightly longer. Total length of bursa 450-500 µm. Bursa without signa or other ornamentation; vestibulum folded and more sclerotised. Ductus spermathecae distinct, but short, with 2-3 indistinct convolutions and ending in wide and pointed vesicle.
Biology. Host plants. Myrtaceae: unidentified tree (probably an Eugenia or Myrcia sp.). Leafmines (Figs 137, 138). The leafmine is a contorted gallery, with linear broken black frass throughout, sometimes forming a false blotch at the end. Larval emergence exit at leaf upperside.
Egg. The egg is deposited on leaf upperside, often on a lateral vein. Larva. Not described. Cocoon (Fig. 139) Distribution. Costa Rica: Guanacaste Province, Área de Conservación Guanacaste. Coordinates type locality: N10.83400, W85.61200. It is a sub-humid to humid tropical dry (deciduous) forest with five to six months of dry season (Herrera Soto and Gómez Pignataro 1993).
DNA barcode. We have DNA barcodes of all ten known specimens, the holotype differs 2.3% from the other specimens.
Remarks. The combination of the venation, absence of transverse bar of transtilla and presence of lateral support rods makes this a typical New World Acalyptris species, probably belonging to the A. scirpi group. The DNA barcode does not place it close to any Acalyptris species of which the barcode is known. Morphologically there are similarities to several species described from Belize: A. bifidus Robinson, 2000 andA. unicornis Puplesis &Robinson, 2000. It is remarkable that the holotype barcode has a 2.3% distance to the Malaise trapped specimens, found in almost the same locality. More material is needed to see whether this is a case of cryptic species, or a unusual high variation in the population. Etymology. Janzeni: a noun in genitive case, based on the family name Janzen, to honour Daniel H. Janzen, collector of part of the material, for his long time dedication to study the tropical Lepidoptera fauna of the Guanacaste Conservation area in great detail, both ecologically and taxonomically, and his enthusiastic support of DNA barcoding (eg. Janzen et al. 2009).