Splitting the leafmining shield-bearer moth genus Antispila Hübner (Lepidoptera, Heliozelidae): North American species with reduced venation placed in Aspilanta new genus, with a review of heliozelid morphology

Abstract The new genus Aspilantagen. n. is described to harbour Nearctic heliozelid moths with reduced venation, previously placed in Antispila Hübner, 1825, with type species Antispila oinophylla van Nieukerken & Wagner, 2012. The erection of this genus has become possible now that monophyly has been supported by a recent phylotranscriptomics analysis. Six species are combined in this genus: Aspilanta oinophylla (van Nieukerken & Wagner, 2012), comb. n., A. hydrangaeella (Chambers, 1874), comb. n., A. ampelopsifoliella (Chambers, 1874), comb. n., A. voraginella (Braun, 1927), comb. n., A. argentifera (Braun, 1927), comb. n., A. viticordifoliella (Clemens, 1860), comb. n. and two candidate species are recognised. DNA barcode COI sequences of Malaise trapped specimens suggest a rich fauna of Aspilanta in Central America. All are leafminers, with Vitaceae as main host family, and single species feeding respectively on Hydrangeaceae and Myricaceae. The species are briefly diagnosed, and data on biology, DNA barcodes and distribution are provided. To place the genus in context, a review of heliozelid morphology and phylogeny is presented and a key to Nearctic genera is given. The genus is confined to North and Central America, possibly also occurring in South America. Aspilanta oinophylla is also an invasive species on grapevine in Italy. The genus is sister to Coptodisca Walsingham, 1895. Another species is removed from Antispila: Heliozela eugeniella (Busck, 1900), comb. n., feeding on Eugenia (Myrtaceae), from Florida.


Introduction
The taxonomy of the Lepidoptera family Heliozelidae, the shield bearer moths, has recently received considerable attention due to the discovery of several previously unknown pest species, attacking grapevines and walnuts Bernardo et al. 2015; van Nieukerken and Geertsema 2015;Takács et al. 2020), as well as the discovery of a rich unknown fauna in Australia that includes pollinating species (Milla et al. 2017). Through international collaboration, it became possible to study the phylogeny of the family on the basis of a large selection of taxa with some nuclear and mitochondrial genes (Milla et al. 2017), and for a smaller subset with transcriptomes (Milla et al. 2019).
It soon became apparent that the relatively large genus Antispila Hübner, 1825, was a catch-all genus for species similarly patterned with a fascia before the middle of the forewing and a pair of opposing spots (the meaning of the Greek anti-spila / ἀντίσπίλα) in the distal half. Several African and Asian species in Antispila with a reduced venation were transferred to the previously monotypic genus Holocacista Walsingham &Durrant, 1909 (van Nieukerken andGeertsema 2015). The Nearctic species that invaded Italy, Antispila oinophylla van Nieukerken & Wagner, 2012 and its relatives, share a similar reduced venation, but without a phylogenetic generic framework it was still described in Antispila, although in the original description it was made clear that this was a temporary placement (van Nieukerken et al. 2012). The phylogeny in Milla et al. (2017) showed that all taxa with reduced venation form a clade: the genera Holocacista, Antispilina Hering, 1941, Coptodisca Walsingham, 1895 together with several species in Antispila, whereas the type species of Antispila, A. metallella ([Denis & Schiffermüller], 1775) and other species with the full venation form another clade. However, in that study, the species with reduced venation in Antispila were paraphyletic with respect to Coptodisca and formed two clades: the A. ampelopsifoliella group and Antispila group II. The transcriptomics study, in contrast, showed a strong support for monophyly of the sampled representatives of these groups, here A. oinophylla and A. argentifera Braun, 1927. As the continued use of the incorrect generic name Antispila can cause confusion, we describe here a new genus for this group of species and give the invasive grapevine miner A. oinophylla its correct taxonomic placement.
To place this description in a wider context, we review the taxonomic history of the genus Antispila sensu lato and review knowledge on the morphology and phylogeny of the family Heliozelidae, partly based on the unpublished thesis of the late Ebbe Nielsen (1980).
In North America, Clemens discovered Antispila leafmines on "Nyssa multiflora" (black gum, now known as Nyssa sylvatica, Nyssaceae), but at first did not know where the reared moths belonged and suggested in a letter to Stainton on 15 May 1859 the new name Diacopia. This name has unintentionally, but validly been published as Diacopia Clemens, 1872, when Stainton published their correspondence much later (Clemens and Stainton 1872). Undoubtedly Stainton informed Clemens about the correct position of this species, and shortly thereafter the first four Nearctic species were described in Antispila by Clemens (1860aClemens ( , 1860b; later more by Chambers (1874aChambers ( , 1874b, Busck (1900), Braun (1915Braun ( ,1927aBraun ( , 1927b and Lafontaine (1973).
A formal selection of Antispila stadtmuellerella as the type species was made by Fletcher (1929), but unfortunately, he and most later authors overlooked the fact that Hampson (1918) had already made a selection of the first species in Hübner's list: A. pagenstecherella (Hübner, 1825). However, this species was already recognised in the genus Eudarcia Clemens, 1860, then Tineidae, now in Meessiidae. Because of this unfortunate situation, where the well-known generic name Antispila in fact belonged to another family, Nielsen and Nye (1986) applied to the International Commission on Zoological Nomenclature to validate Fletcher's selection of the type species rather than Hampson's. This was granted two years later (ICZN 1988).
Until recently, the validity of the genus Antispila was rarely challenged; only the small Mediterranean species A. rivillei (Stainton, 1855) was moved to the genus Holocacista, when the different venation was noted by Walsingham (1909). Nielsen and Traugott-Olsen (1981) moved the American A. major Kearfott, 1907 to Stephensia (Elachistidae) and later one more misplaced Elachistidae was discovered: A. merinaella Paulian & Viette, 1955 from Madagascar, without assigning it to another genus (van Nieukerken and Geertsema 2015).
The presence of species with a reduced venation was already noted by Meyrick (1916) when describing the Neotropical A. trypherantis Meyrick, 1916, A. pentalitha Meyrick, 1916and later A. cyclosema Meyrick, 1921(Meyrick 1921. This was also noted by Kuroko (1961), when describing the first Japanese Antispila. In North America, Don Lafontaine (pers. comm.) had also independently discovered the difference in venation, and a difference in the shape of the "shield", the pupal case cut out from the leaf mine by the mature larva, when working on Heliozelidae, and planned the description of a new genus in the early 1970s; however, this never happened as he changed his study subject to larger moths.
When describing the new Antispila oinophylla¸ the non-monophyly of North American Antispila became apparent once again (van Nieukerken et al. 2012), and during the study of South African Heliozelidae, several African and Indian species were transferred from Antispila to Holocacista and Heliozela respectively ( van Nieukerken and Geertsema 2015). The two phylogenetic studies that followed (Milla et al. 2017(Milla et al. , 2019 corroborated this and strongly indicated the polyphyly of the old genus Antispila. Here we continue this process by placing six Antispila species in the new genus Aspilanta and recombining one species with Heliozela. After this action, five named Nearctic species remain in Antispila s. str., three feeding on Cornus (Cornaceae) and one each on Nyssa (Nyssaceae) and Vitis (Vitaceae). Globally the genus now has 27 named species in the Holarctic and Oriental regions, mainly feeding on Cornales and Vitaceae. The three European species, all feeding on Cornus, were recently revised by ; the 12 named and one unnamed Japanese species were treated by Lee and Hirowatari (2013) and in total five species were recently described from China (Liu and Wang 2017;Liao et al. 2019). One unnamed species was recorded from South Africa (van Nieukerken and Geertsema 2015). The majority of the nine species listed in Antispila from the Neotropics (Heppner 1984) probably belong in other genera, including Aspilanta.

Materials and methods
As most species treated here were already discussed in the paper describing Antispila oinophylla, we re-use several photographs from  and refer to that paper for the methodology used by EvN for collecting, preparation, imaging, and DNA barcoding, and updates in subsequent papers on Heliozelidae (van Nieukerken and Geertsema 2015; . Additional fieldwork in North America was carried out by EvN in 2015, 2016 and 2018. All examined and observed material, including parasitoids and all available barcoding data are listed in the occurrence dataset, published with GBIF (DOI: https://www.doi. org/10.15468/79mu7z). In the text we omit all material that was already listed by , even though some of these records later received registry numbers (in CNC and RMNH) or were dissected; these data can be found in the GBIF dataset. Material collected in 2011 was cited in the 2012 paper only from larvae and leafmines; adults reared in 2012 are cited here. We follow the new guidelines for fine-grained formatting of material examined. We divide the material into four categories: examined adults (numbers without male or female sign represent unsexed adults); examined larvae and leafmines; DNA barcoded material that we did not examine personally, but identified on the basis of the barcode (all DNA barcoded specimens can also be found in the BOLD dataset DS-ASPIL (DOI: https://www.doi.org/10.5883/DS-ASPIL); and observations from internet sites and other photographs that could be reliably identified on the basis of the photograph; unless otherwise indicated, these refer to leafmines with or without larvae.
Under the heading "Distribution" we list countries, states and provinces; an asterisk denotes a new record.
Leafmines were photographed by CSE in the field using a Nikon D50 digital camera and a Nikon AF Micro Nikkor 105 mm lens. The camera's built-in flash was used for photographs with reflected light; the sun was used for transmitted light. All other photographs by CSE were taken using a Canon EOS Rebel XSi SLR digital camera, MP-E 65 mm macro lens, and Macro Twin Lite MT-24EX flash unit. For transmitted light in leaf mine photographs, one of the flash heads was removed and aimed at the camera while the leaf was held tightly against it.
Leaves containing larvae were placed in sealed, 9-or 15-dram plastic vials, which were stored away from direct sunlight and checked daily (when possible) for emerging insects. After pupal cases were formed, extraneous leaf material was removed from the vial to minimise mould and a moistened, crumpled piece of toilet paper was added to prevent desiccation. Overwintering pupal cases were stored in a refrigerator at 1-3 °C from 6 November 2013 to 25 February 2014, 1 November 2014 to 1 March 2015, 1 November 2015 to 1 March 2016, 20 October 2017 to 1 April 2018, and 17 October 2019 to 21 February 2020. Adult moths were chilled and photographed as described by Eiseman and Lonsdale (2018) for agromyzid flies.
Morphological terminology follows  with the corrections proposed by van Nieukerken and Geertsema (2015).
Host plant names follow the Catalogue of Life (Hassler 2020), which agrees with the Flora of North America for the Vitaceae (Moore and Wen 2016) and Hydrangeaceae (Freeman 2016), but for Myricaceae rather follows Wilbur (1994). As the differences between the Virginia creeper species Parthenocissus quinquefolia and P. vitacea are rather subtle, it is possible that some records of P. quinquefolia may in fact refer to P. vitacea. According to Pringle (2010) the last species should be named P. inserta, but the cited sources do not follow this, despite the rather strong argumentation.
We list parasitoids that we reared from Aspilanta larvae. Those reared by EvN were routinely stored in ethanol 96-100%, and later DNA barcoding was performed (dataset DS-PARASP, doi: https://www.doi.org/10.5883/DS-PARASP). Chalcidoidea were morphologically identified by Sandrine Ulenberg, Braconidae were tentatively identified by EvN on the basis of DNA barcodes; all are stored in RMNH; Barcode Identification Numbers are listed after the taxon names. Parasitoids reared by CSE were photographed alive, then preserved in 95% ethanol and sent to specialists: all Eulophidae, except the Pediobius, were examined by Christer Hansson and deposited in BMNH; Braconidae were determined by José Fernández-Triana from photographs and are deposited at CNC. We have not found any literature records of reared parasitoids that were identified below family level. Species listed here as Braconidae: Microgastrinae belong either to the genus Dolichogenidea or Pholetesor; a detailed taxonomic revision of this group is necessary (José Fernández-Triana, pers. comm.). The record of this subfamily listed here under A. ampelopsifoliella, was in fact the basis for the first record of Heliozelidae as host family (Fernandez-Triana et al. 2020). All records of parasitoids are also listed in the GBIF dataset (DOI: https://www.doi.org/10.15468/79mu7z).
Abbreviations for depositories: Differential diagnosis. Very small moths, wingspan between 4.0 and 6.2 mm, with a forewing pattern of metallic-silvery markings, comprising an oblique fascia at ¼, a postmedial pair of spots (one costal and one dorsal), and usually a small apical spot (only absent in A. viticordifoliella); fringe line more or less distinct. Males never with androconial scales or hair-pencils. Antennae with only 16-20 segments. Antispila species in North America never have an apical spot. Aspilanta species are diagnosed by the reduced venation (Figs 9-13); in Antispila the discoidal cell is present and more veins are retained (illustrated in ; most Antispila species are larger and have more antennal segments. Separated from Heliozela species by more extensive colour pattern and apical spot, and Heliozela species have the venation with discoidal cell and a distinct epiphysis on foreleg. Coptodisca species are readily recognised by their colour pattern (Bernardo et al. 2015;Eiseman 2019). The genera Holocacista and Antispilina, not yet known from the New World, have a very similar venation, but always lack the apical spot. Moreover, Holocacista species have a small epiphysis on the foreleg, and the phallus bears an unusually long, often recurved appendix.
Head . Almost oval in outline. Eyes in latero-ventral position, ventral margin not reaching lower margin of head. No sutures present. Anterior tentorial arms very slender, prominently curved laterally before converging towards frons. Vestiture comprising lamellar scales, closely appressed on head, in dry specimens scales on vertex sometimes raised, probably an artefact as a result of drying. Mouthparts: labrum narrow, pilifers absent. Mandibles small, as long as broad, relatively well sclerotised (Fig. 16). Maxilla with galea well developed and almost twice as long as head; maxillary palp reduced to a single segment. Labial palp well developed, 3-segmented, drooping, slightly shorter than head capsule; distal segment almost 3× as long as second segment; depression for Organ von Rath not seen. Antenna (Figs 14,18) ca. half length of forewing with 14-18 flagellomeres (16-20 segments) [best counted in denuded specimens on slides], no sexual dimorphism. Scape and pedicel of equal length, slightly shorter and wider than flagellomeres. Flagellomeres ( Fig.18) cylindrical, ca. twice as long as wide, each with two annuli of scales, most dark, some apical flagellomeres may be white. Pecten present, but not easily visible; with ca. 4-5 hairs.
Thorax. Vestiture of appressed lamellar scales, either concolourous with ground colour of forewings or more metallic and similar to vestiture of head. Foreleg (Fig. 20) without epiphysis.
Wings. Male retinaculum a series of 7-12 hook-shaped bristles (Fig. 19), arising from a thickened serrate portion of Sc. Frenulum in male a strong curved bristle (Figs 12,13), in female two bristles present (Fig. 10); pseudofrenular bristles in male absent. Humeral field with scattered microtrichia, otherwise microtrichia restricted on wing membrane to area just posterior to retinaculum, arranged in longitudinal rows. Scale sockets regularly spaced, not in distinct rows.
Venation . Forewing with Sc to middle of costa. R unbranched, a separate vein, to costa, but a persistent trachea connecting R with Rs+M+CuA. Rs+M+CuA ending in 3-4 branches, interpreted as Rs1+2 to costa, Rs3+4 to termen, M and CuA  to dorsum. Hindwing with Sc+R to costa, Rs+ M with 2-3 branches, Rs to costa, 1 or 2 branches of M to termen and dorsum; CuA a separate vein to dorsum.
Wing pattern (Figs 1-8). On forewing typically comprising a silvery white metallic fascia at 1/3, widest at dorsum, and a similarly coloured pair of opposite spots at 2/3 on a dark background, brown to black, with a small silvery blue spot in apex, equidistant to dorsum, costa and fringe line, but the apical spot is absent in A. viticordifoliella. A fringe line often present, with fringe scales pale. Hindwing uniform grey. Androconial scales absent in all species examined.
Male genitalia . Vinculum (SIX) very long, anteriorly often reaching beyond anterior margin of segment VI, almost cylindrical; approximately 2/3 of total length of genitalia. Tegumen (TIX) narrow, often bilobed, or medially indented, or truncate; probably a composite structure with uncus. Gnathos absent. Valva approxi- mately triangular, with stalked pectinifer halfway to inner margin, pecten comprising 10-22 blunt sensilla (comb teeth); transtilla typically with medial anterior projection, sublateral processes long. Phallocrypt (manica) with some to many strongly sclerotised  conical spines, often arranged in an asymmetric fashion, or with many smaller spines. Phallus outer tube often with remarkable appendices of different sizes and shapes. Juxta present, often bilobed or arrow shaped, in hydrangaeella with extra spines laterally.
Larva . Yellow or whitish, usually with brown to almost black head capsule and prothoracic sclerites and dark anal plates; no plates present on other segments, but concentration of cuticular swellings can give impression of darker plates on abdomen in some species. Head prognathous, 2 stemmata at either side. Thorax with elongate dorsal and ventral sclerites without adornment; 10 th (last) abdominal segment with single dorsal and paired ventral sclerites, with several prominent setae; other thoracic and abdominal segments covered with small transverse swellings. Legs and prolegs absent but paired ambulatory calli on T2 and T3 (ventral and dorsal) and fused ventro-medial calli on A3-6. Number of instars unknown, but likely with four feeding instars and a fifth non-feeding instar that constructs the case in which it pupates, in analogy to Holocacista, Coptodisca, Antispila and Heliozela (Dziurzyński 1948(Dziurzyński , 1952Marchi 1956;Prota 1962;Maier 1988).
Biology. Host plants. Most species feed on Vitaceae, one each on Hydrangeaceae and Myricaceae.
Life history. Eggs are inserted in leaf tissue, often near a vein or leaf margin. All species construct leafmines, either starting as a narrow linear mine and later widening into a blotch, or sometimes starting almost immediately as a blotch mine. All frass is deposited in the mine, initially filling or irregularly scattered in the linear portion, later often scattered in the blotch or pushed by the larva to one side. During the penultimate (fourth) instar an oval section is cut out from both epidermal layers, forming a portable case or "shield". This shield (Figs 55,68,76,93,106), later forming the cocoon, is more or less flat, without the raised central ridge that is characteristic for Antispila. The larva descends with its shield on a strand of silk and may wander for some distance before finally attaching the shield at one end to a substrate (leaf, trunk, leaf litter, etc.), where it moults to the fifth non-feeding instar and later pupates. The pupa protrudes from the opposite end of the shield when the adult emerges. As far as we know, most species are bivoltine, overwintering as fifth instar larva in the cocoon, but A. voraginella and possibly A. ampelopsifoliella are univoltine. Adults are usually day flying and can be swept from the hosts, but rarely come to light. Several specimens were taken in Malaise traps (DNA barcoded material) and in several cases provide the only accurate phenology data for adults.
Distribution. North America; DNA barcodes suggest a rich fauna in Central America: Mexico, Honduras, Costa Rica, likely also elsewhere in the Neotropics (see under Composition).
Etymology. The name Aspilanta is an anagram of Antispila, where one "i" was replaced by an "a". The gender of the name is to be regarded as feminine. Composition. In the checklist below we provide the original genus in brackets, type locality, and the hostplant of the types. The species are listed according to the posi-  tion in the recent phylogenetic analyses (Milla et al. 2017(Milla et al. , 2019. We also include the candidate species Aspilanta "Vitis1_USA" (van Nieukerken et al. 2012) and A. "Vitis.arizonica_USA". Some publicly available DNA barcodes ( Fig. 44) closely match confirmed Aspilanta sequences, suggesting further candidate species and a rich fauna in Mexico, Honduras and Costa Rica, but until these taxa are examined morphologically, we omit them. The following BINs are concerned: Mexico: BOLD:ACZ5051, BOLD:ACP0240, BOLD:ACO9420, BOLD:ACU0821, BOLD:ACT4781; Costa Rica: BOLD:ADA1988, BOLD:ACL9188; Honduras: BOLD:ACF9350.
The Neotropic species Antispila trypherantis Meyrick, 1916 (Guyana), A. pentalitha Meyrick, 1916 (Guyana) and A. cyclosema Meyrick, 1921 (Brazil) may also belong in Aspilanta, based on their original descriptions that cite the presence of an apical dot, but without examination of types, we refrain from recombination here. Also the Patagonian group of species, associated with Nothofagus (Nothofagaceae), for which Nielsen in his unpublished thesis proposed the name "Neospila", could belong in Aspilanta on the basis of the very similar externals (Fig. 7), although following the latest phylogeny, where it occurs as Genus14, its inclusion would make Aspilanta paraphyletic (Milla et al. 2019) (Fig. 43).
In the Museum of Comparative Zoology (Cambridge, MA), there is a series of externally similar moths of unknown provenance, ex coll. Dietz, allegedly reared from poison ivy, (Toxicodendron radicans, Anacardiaceae) with the manuscript name "Antispila rhoifoliella" (handwritten label: on Rhus radicans. Coll 9.7. [18]99, many mines empty, larva pale, green frass line; head + 1 dark brown. [word crossed out] mine begins with a fine tract along edge of leaf, expands & frass collects along edge of mine.). These specimens have emergence dates of 1-20 June 1900 (written as "19C"). As we have never seen such mines on poison ivy we cannot exclude the possibility that Virginia creeper was mistaken for poison ivy, as both often grow together; in this case the series would likely represent Aspilanta ampelopsifoliella. We thus ignore this information until these specimens have been examined in more detail.  (Braun, 1927b): 191, comb n. (Antispila) USA, Utah, Washington Co., Vitis arizonica Engelm. Aspilanta argentifera (Braun, 1927a): 56, comb n. (Antispila) Canada, Ontario, Sparrow Lake., collected as adult without definite host association. Aspilanta viticordifoliella (Clemens, 1860): 209, comb n. (Antispila) USA, Pennsylvania, Easton, Vitis vulpina L.

Comparative morphology of Heliozelidae
In order to judge the validity of the morphological characters in Aspilanta, they should be compared with the same characters in other Heliozelidae. There is no single large treatment of the family, but Davis (1998) provides an introductory description and a lot of information is found in Nielsen's unpublished thesis (Nielsen 1980). However, there are very detailed monographs on all stages of species in the genera Heliozela (Prota 1962), Antispila (Dziurzyński 1948(Dziurzyński , 1952 and especially for immatures in Holocacista (Marchi 1956). For the last genus, the recent taxonomic paper provides further information (van Nieukerken and Geertsema 2015) and for Coptodisca a number of papers deal with various aspects (Snodgrass 1922;Peterson 1948;Maier 1988;Bernardo et al. 2015). Some information is also provided in the paper describing the genus Plesiozela (Karsholt and Kristensen 2003), but as that genus has now been shown to belong to Incurvariidae (Milla et al. 2019), it is only partly relevant. Unfortunately, there is currently little published information on the detailed morphology of the large Australian clades (Hoplophanes group, Pseliastes group, Milla et al. 2019), with only a brief adult description of Hoplophanes Meyrick, 1897, available in Nielsen's thesis.

Adult.
Head. All Heliozelidae share the smooth head, with appressed, lamellar scales. This is a strong apomorphy for the family. Antennae in Hoplophanes have ca. 28-30 flagellomeres, Heliozela 15-25, ca. 23 in Antispila, 18 in Antispilina, 12-18 in Holocacista, 12-16 in Coptodisca and 14-18 in Aspilanta. The flagellomeres are much longer than wide in Aspilanta and other genera in the Holocacista group, almost the same in Antispila, but in Heliozela hardly longer than wide; all flagellomeres have two rings of scales. Mandibles are small, but visible in all genera (according to Nielsen absent in the leafmining genera). The proboscis (galea) is more than twice the height of the head in Heliozela and Antispila, almost twice the height in Aspilanta, but hardly longer than the head in Holocacista, Antispilina and Coptodisca. The maxillary palpi are progressively reduced, from four segments in Hoplophanes and Heliozela, through three segments in Antispila and a reduction in the Holocacista group genera into a single segment (Nielsen considered them absent, but we clearly observed a short palp). Labial palpi usually have three segments, but only two in Antispilina ludwigi; the last segment is very long in Aspilanta (3 × second segment) compared to Coptodisca (slightly longer than second segment). Eyes are rather similar in most Heliozelidae, but comparatively large in Hoplophanes (interocular index 1.06, see (Davis 1975)), in other genera much smaller (less than 0.6), but in Aspilanta larger than in related genera.
Thorax. The most relevant character in the legs is the presence of an epiphysis on the foretibia, a plesiomorphic condition in Lepidoptera. It is present in Hoplophanes, Tyriozela, and Heliozela, probably also in the other Australian genera, but absent in Antispila and most of the genera in the Holocacista group, but Holocacista rivillei and H. capensis have a reduced epiphysis (van Nieukerken and Geertsema 2015).
Forewings in the majority of Heliozelidae are metallic shining. A forewing pattern of spots and fasciae is present in many Heliozelidae, in Antispila, Antispilina, Holocacista and Aspilanta comprising a fascia (or two spots) at one third and an opposite pair of spots at two thirds, somewhat reduced in Ischnocanaba. An apical dot is characteristic for most Aspilanta species, but something similar also occurs in Ischnocanaba and Nielsen's "Neospila". Larger terminal spots are known in some Antispila species. Most Heliozela species have the silver spots confined to the dorsal margin, but in some cases one of these reaches the costa, or an extra spot is present on the costa, and a few have a pattern similar to Antispila. Such species were previously incorrectly placed in Antispila (e.g., H. anna (Fletcher, 1920)), like many of the species in Holocacista and Aspilanta. Coptodisca species have a completely different pattern, resembling that of Leucoptera (Lyonetiidae).
Androconial scales on one or both wings are common in Antispila, but absent from most other genera.
Venation. Hoplophanes, Heliozela, and Antispila have the more complete venation, with a closed cell between Rs and CuA, three branches of Rs; in Antispilina, Holocacista, Aspilanta and Coptodisca the venation is reduced, Rs, M and Cu are coalescent from wing base, closing the cell, and ending in three or four branches, rarely two. The label CuA given for the A. oinophylla forewing by   fig. 6) is misplaced; the line next to it is a staining artefact. What is labelled M2+3 should be CuA (see Fig. 9). The hindwing of all Heliozelidae lacks the cross-vein M-CuA that is present in most other Adeloidea. The wing coupling comprises in the hindwing a single frenulum in males and 2-5 frenular bristles in females; in Heliozela sericiella it is 4-5 according to Prota (1962), but we observe only two in H. hammoniella, in Antispila three, in Holocacista, Aspilanta and Coptodisca only two. The forewing retinaculum comprises a distinct cuticular lobe in male Hoplophanes, Tyriozela and Heliozela, provided with ca. four or five hooked scales; the female has a row of setae inserted on the vein Rs+M. In Antispila the Sc forms a long fold with a row of scales, female as in Heliozela. Males in the Holocacista group have a series of hooked scales on the Sc fold, very similar to that structure in Nepticulidae ( van Nieukerken 1986).
Male genitalia. These are rather uniform throughout Heliozelidae, with a long and narrow vinculum; bilobed or truncate tegumen (sometimes termed uncus); triangular or elongate valvae, each with a single pectinifer with 8-30 blunt sensilla, often different numbers on each valva; a spear-shaped juxta; phallus very long and narrow, with anellar spines and often large projections at phallotrema. Whereas the male genitalia are diagnostic at species level, few characters can be used at higher phylogenetic level. The recurved appendix of the phallus is probably characteristic for Holocacista, whereas the very long vinculum (more than twice the valva length) separates most species of the Holocacista group from Antispila.
Female genitalia. The oviscapt is spear shaped with many small teeth in Hoplophanes, Tyriozela and Heliozela, whereas in Antispila and the Holocacista group it has just a few cusps at either side. In Antispila there is a distinct interapodemal process between the anterior apophyses, running almost to the caudal tip. In the Holocacista group the process is very short, the caudal part being absent.
Larva. Larvae of the Australian genera are unfortunately hardly known. Leafmining larvae of Heliozela sericiella (Prota 1962), Holocacista rivillei (Marchi 1956) and Antispila treitschkiella (Dziurzyński 1948) have been studied in detail, including all instars, and the fourth instar of A. metallella by Grandi (1933). The larvae are usually white to yellow or green, flattened and with prognathous head.
Head. Fourth instar larvae of Heliozela have two pairs of stemmata at either side, those of Antispila three, and Aspilanta two. In Holocacista and Coptodisca probably also two, but difficult to see in our slide mounted specimens. Davis (1978Davis ( , 1998 reports 2-5 for the family. Thorax. Legs are absent in most studied larvae, but some Heliozela species have much reduced legs during the 5 th , non-feeding instar (Prota 1962). Only for full grown H. aesella larvae fully developed thoracic legs are reported (McGiffen and Neunzig 1985;Davis 1987), which CSE has also observed in live larvae (including in early instars). At least some of the Australian taxa also have more developed legs, although many are also legless (Andy Young & Mike Halsey, pers. comm.). In most cases the prosternum and protergum are elongate sclerotised structures, in Coptodisca sometimes adorned with a ventral transverse row of swellings.
Abdomen. Prolegs are only reported from 4 th instar Heliozela larvae on segments 3-6, with 3-8 crochets in single or sometimes double mesoseries. Instead most larvae show either paired or unpaired calli, the position of which may be diagnostic, but is difficult to assess in slide mounted specimens. The integument is covered by slight, often transverse swellings, which may concentrate in the centre of the segments, and in some cases form a row of abdominal plates in species of Antispila. The presence of warts on the 8 th larval segment, used for stridulation or drumming (Low 2008;) is typical for many species of Antispila; they are unknown for the other genera.
Cocoon. Cocoons (shields) of leafmining species are prepared from both epidermal layers of the final part of the leafmine, held together by silk, and are very similar in most genera. Antispila has as extra character in that the larva contracts the case with silk, resulting in a lengthwise keel; also in all Antispila species there are silken projections from the anterior (and sometimes the posterior) end of the case; we do not know them from Heliozela, Holocacista or Coptodisca, but some Aspilanta species produce these projections at least sometimes ( Fig. 106) (Eiseman 2019). Cocoons are made by 4 th instar larvae and not by the final instar as some authors state incorrectly. The final moult takes place in the shield.
Pupa. The pupa of Heliozela sericiella was described by Prota (1962); otherwise there is little published information on pupae of Heliozelidae.
Phylogeny. We use the recent molecular phylogeny based on transcriptome data, including up to 1049 nuclear genes (Milla et al. 2019) as the basis for further discussions on the phylogeny. Part of this tree is here reproduced, with the name Aspilanta instead of Antispila group II, and with the listed apomorphies added (Fig. 43). The resulting tree is in fact not that different from the manually derived tree, based on morphology, in Ebbe Nielsen's thesis (see Milla et al. 2017). The most apparent differences are the removal of Plesiozela from Heliozelidae to Incurvariidae, the position of Antispilina (Nielsen placed it next to Antispila) and the non-monophyly of the Australian Hoplophanes and Pseliastis. Also, Nielsen did not study any of the species here placed in Aspilanta. Interestingly, the leafmining genus groups, viz. the Heliozela group, the genus Antispila (= Antispila group) and the Holocacista group together form a monophyletic group, already recognised by Nielsen (1980). Here we provide possible apomorphies for the genera and clades recognised in this part of the tree only, based on Nielsen's work and our own observations. Characters without an asterisk were already listed by Nielsen (1980). They are not explained further when treated above.
Possible apomorphies for the Heliozela, Antispila, and Holocacista groups are: (1) Size strongly reduced. The wingspan of all the species of these genera is small, always less than 9 mm, and usually smaller.
(2) Compound eyes in a posteroventral position. The anterior lateral margin of the compound eyes does not reach the anterior surface of the head capsule, but is present behind the laterally extended genae.
(3) Proboscis very long, two to three times as long as height of head capsule. However, as noted above, it is shorter again in the small moths of Holocacista, Antispilina, and Coptodisca. Nielsen also listed absence of mandibles as an apomorphy, but we think that reduced mandibles are in fact present.
A possible apomorphy for the Heliozela group (including Tyriozela) is: (4) Female with spear-shaped ovipositor tip. The flattened, deeply dented ovipositor observed in the Antispila and Holocacista groups is regarded as the relatively plesiomorphic state compared to the spear-shaped tip of the Heliozela group. The first mentioned type is present in most Incurvariidae and the most generalised genera of Prodoxidae. Two other characters given by Nielsen are in our view incorrect; the retinaculum is not as long as Sc as he states, and although some species have petiole-mining larvae, this is not the only larval feeding mode; many Australian and Oriental species are leafminers, and others are gall formers.
Possible apomorphies for Antispila and the Holocacista group are: (5) Head with transfrontal suture absent and scale-sockets on the entire frons and vertex. The area in all other Adeloidea is strengthened and devoid of scale-sockets. (6) Foretibia without fully developed epiphysis. (7) Pseudofrenular bristles absent in male. (8) Maxillary palpi three-segmented or shorter.
(9) Microtrichiation reduced on forewings. Only the proximal third is provided with microtrichia; in Hoplophanes the entire wing surface is covered with microtrichia, what is also regarded as a ground-plan condition within Adeloidea.
(13)* The presence of warts on the 8 th larval segment may be an additional apomorphy, but more likely only for a subset of the genus. Nielsen gave the presence of an interapodemal process in female as important apomorphy for the Antispila group, but we see a similar structure, albeit somewhat reduced, in Holocacista (van Nieukerken and Geertsema 2015: figs 53, 68) and Aspilanta (see Fig. 29). It is possible that this character is another apomorphy for Antispila + Holocacista group.
Possible apomorphies for Coptodisca are: (25). Forewing pattern specialised. DNA barcoding. We provide COI DNA barcode sequences as an aid for identification for all known species of Aspilanta. A Neighbor-Joining tree of these ( Fig.  44), including some nearest neighbours (NN) and some species of the sister-group Coptodisca. Details on the Barcode Identification Number (BIN), the average and maximum distance within each BIN, and the distance to NN are given for each species. All species are well distinguished by barcodes. Aspilanta hydrangaeella has two BINs, with a considerable distance, and A. viticordifoliella has three BINs, although the specimen with BIN BOLD:ACA6487 has not been studied; its identification is only based on the barcode.

Treatment of species
Key to known species of Aspilanta 1 Forewing with an apical silver spot in addition to a fascia at ¼ and two opposite spots at 2/3; antenna dark throughout or with the last 1-3 segments white . has distinct white antennal tips. Male genitalia with characteristic long curved process and a comb of 10-12 large teeth at phallotrema; tegumen truncate; valvae with 10-12 pecten sensilla. Female oviscapt with more cusps (4-5 at either side) than A. ampelopsifoliella (3). Leafmines differ from other Vitaceae miners by compact size, rather short linear portion close to a vein and frass in concentric lines, especially in thin leaves.
Leafmines. (Figs 51-58) Egg usually within 1-2 mm from a vein. The mine starts as a rather straight or slightly contorted linear mine towards the vein, usually forming a right angle and often following the vein for a short distance, then again turning away from the vein and expanding into a blotch. The linear portion of the mine is usually later incorporated into the blotch. The frass in the linear portion usually occupies the complete mine width, occasionally deposited in a thin line. In the blotch much of the blackish-brown frass is deposited close to the origin in semi-circular concentric frass lines, best seen in thin shade leaves; in sun-exposed leaves the frass pattern is often obscured. The whole mine occupies as a rule an area of less than 10 × 10 mm; only in thin leaves are mines appreciably larger. The larva cuts out an elliptic case ca. 3.2-4.0 mm long.
Larva. Yellowish green, with green gut contents; head and prothorax brown. Life history. Larvae in Canada and NE United States in July, August and September, more south until October; in Italy in June-July and August-October. Adults from June to August; apparently bivoltine.
Parasitoids Remarks. In the original description it was noted that the species had not been found on Parthenocissus in North America. Since then some specimens of this species were barcoded and reared from Parthenocissus by J.-F. Landry (material in CNC), in a situation where Vitis and Parthenocissus grew entangled. Even though this would weaken the argumentation about the identity of A. ampelopsifoliella by van Nieukerken et al. (2012) Antispila hydrangaeella ;Chambers 1877: 195;Chambers 1879: 126;Edwards 1889: 126;Davis 1983b: 4;van Nieukerken et al. 2012: 56;Eiseman 2019Eiseman : 190, 1333. Antispila hydrangiaeella; Dyar et al. 1903: 539;Forbes 1923: 226;Needham et al. 1928: 290;McDunnough 1939 Leafmines. (Figs 59-68) Egg inserted anywhere on the leaf, but frequently near a vein. Mine starts with a rather long linear portion, which may be partly or completely contorted, or often follows veins for some length; later the mine turns into an elongate blotch or a wide gallery. Frass starts usually as a rather narrow line, often broken and irregular, not always in the middle, later becoming more dispersed, in grains or sometimes smeared out, green to brown; in the blotch the frass remains in the middle, smeared out, dark green to black. Often many mines occur in the same leaf; mines can be rather extensive in the thin leaves of Hydrangea. The larva cuts out an elliptic case ca. 3.5-4.5 mm long.
Larva. Colourless or whitish except for green gut contents; head and prothorax dark brown, some darker dots on most segments visible (Fig. 68).
Life history. Poorly known; larvae found in June, August, September and October; adults emerged the following spring from larvae collected in October, but dates are too early due to forced emergence. Chambers provides no information whatsoever; Forbes (1923) reports moths in August. This, together with an early mine photographed in Kentucky on 30 June, suggests the species may be bivoltine.

Remarks.
In our barcode analysis we found two clusters with a distance of 5.25%, even the distance of two specimens of the Georgia population is rather large with 2.15%. By analysing the characters of adults of the two barcoded populations, we did not find any supporting argument for the earlier suggestion that two species might be involved (van Nieukerken et al. 2012). Mines of the Georgia and Ohio populations seem to have more contorted mines than the population from North Carolina, but there is also variation in this character within these populations. A further analysis of material is needed to assess this situation, and designation of a Neotype from Kentucky might be necessary to avoid confusion if the species appears to be part of a complex.

Aspilanta ampelopsifoliella
Differential diagnosis. Wingspan ca. 5.0-5.3 mm, forewing length 2.4-2.6 mm. Externally inseparable from A. oinophylla (see above) and A. "Vitis1_USA". Male genitalia characterised by the non-bilobed tegumen, pecten with 11-14 sensilla, and phallus with rather small curved spine at phallotrema. Female oviscapt with 3 cusps at either side. Leafmines recognised by long linear portion that is often sinuous. Host plants. Vitaceae: Parthenocissus quinquefolia, P. vitacea. Leafmines. (Figs 69-79) Egg in the majority of examined mines close to midrib, when at leaf margin then usually close to leaf base, occasionally near another vein. The mine starts with a long linear portion, often very sinuous, occasionally straighter, when it runs along the leaflet margin; gradually widening into elongate blotch. Frass brown to black, initially broken linear, later irregularly dispersed, not always in centre of linear portion; in blotch more or less in centre throughout. The larva cuts out an elliptic case ca. 3.5-4 mm long.
Larva. Colourless or whitish except for green gut contents, in contrast with yellow-green larva of A. viticordifoliella on the same host plant (compare Figs 69, 70 with 97-100); head and prothorax almost black.
Life history. Poorly known. Larvae found from June (Oklahoma), August and September, field caught adults only from early July. The adult from the June larva emerged the following spring, like those from August and September larvae, suggesting that this species is univoltine.
Barcode. BIN: BOLD:ABA3237, average distance 0.3%, max. distance 1.12% (n = 10), distance to nearest neighbour 8.79% (BOLD:ACO9420, an unidentified Heliozelidae from Mexico, most likely an Aspilanta species). Remarks. The identity of this species follows from the discussion in van Nieukerken et al. (2012). As there are no Syntypes to select as Lectotype, we suggest that a reared male specimen should be selected as Neotype, preferably collected as close to the type locality as possible.
Material Differential diagnosis. Wingspan ca. 4.7-5.4 mm, forewing length 2.3-2.6 mm. Externally separated from A. oinophylla and ampelopsifoliella by the darker brassy brown head scaling, the absence of a fringe line and the overall less shining metallic pattern. Relatively similar to A. argentifera, but unlikely to be sympatric with that species. Male genitalia with distinctly bilobed tegumen, fewer pecten sensilla (8-10) and phallotrema with smaller spines only. Host plant. Vitaceae: Vitis arizonica, unidentified Vitis. Leafmines. (Figs 80-83) Yellowish white blotches, without any linear part; greenish black frass irregularly in centre; mine usually formed from the confluence of several mines; as many as twenty or twenty-five pupal cases may be cut from a single leaf.
Life history. Larvae found between 4 June and 9 August (Arizona); in all cases the adults emerged the next spring (April to June), suggesting that the species is univoltine.
Leafmines. (Figs 88-96) Egg in the examined mines inserted at the midrib. The mine starts with a linear portion, following the midrib towards the apex, sometimes difficult to see when it is very close to the midrib (Fig. 94); later becoming a small blotch extending from midrib to leaf margin. Frass blackish, varying from dispersed to concentrated near origin of mine; in linear portion a wide line in middle. Mines on Comptonia are usually very compact, on Morella and Myrica more elongate; on Myrica gale we observed several mines crossing the midrib, resulting in a wilting leaf tip that may drop off (Fig. 90). The larva cuts out an elliptic case ca. 2.5-3.5 mm long.
Larva. Pale yellowish green, head and prothorax brown.  Life history. Bivoltine. Larvae found in July and September to November, adults in June, July and August; specimens reared from fall mines emerged (indoors) in April to May, from July mines in August. Mines of the first generation seem to be very scarce.
Remarks. Annette Braun (1927a) believed that the moth that she collected at Sparrow Lake belonged to long leafmines that she found on Betula in the same spot. Her description of the mine actually matches very well that of Phylloporia bistrigella (Haworth, 1828) (Incurvariidae) (Eiseman 2019; Ellis 2020), a Holarctic species, widespread in Canada, but in 1927 poorly known in North America as either Incurvaria labradorella Clemens, 1864or I. aureovirens Dietz, 1905 had not yet been associated with Betula and which were only later synonymised with P. bistrigella by Davis (1983a). On the basis of Braun's record of these leafmines, P. bistrigella should also be listed for Ontario, a record missing from . As we have never seen any heliozelid mine on Betula, and as the Holotype of A. argentifera externally matches the specimens reared from Myricaceae, we are convinced that argentifera is the correct name for this species (see also Milla et al. 2017: 140). According to Braun (1927a), the holotype was a male, but the photograph (Fig. 6), kindly made by Mike Palmer, clearly shows a female, with protruding oviscapt. We refrain for now from borrowing and dissecting the holotype, as the characters of female genitalia have not been sufficiently studied, and a future dissection could possibly be combined with newer techniques for obtaining DNA data. Already Forbes (1923) reported an Antispila species from sweet fern (= Comptonia), which strange enough according to him "has not been distinguished from A. isabella".
Specimens from Prince Edward Island were previously regarded as misidentifications (    Barnes and McDunnough 1917: 181;Forbes 1923: 226;Needham et al. 1928: 20, 153, 290;McDunnough 1939: 91;Davis 1983b: 4;Brower 1984: 29;McGiffen and Neunzig 1985: 55;van Nieukerken and Pohl 2018: 41;Eiseman 2019: 189, 190, 729, 733. Antispila cf. viticordifoliella van Nieukerken et al. 2012: 58;Eiseman 2019: 729. Differential diagnosis. Wingspan ca. 5.5 mm, forewing length 2.2-2.5 mm. Externally different from all other Aspilanta species by the absence of the apical spot in the forewings. Differs from most Antispila species, which have a similar wing pattern, by the antennae with distinct white tips, the complete absence of androconial scales in the males and the smaller size. Genitalia not yet examined. Leafmines differ from other Aspilanta on Vitaceae by the absence of a linear portion and by the larva that is yellow-green (vs. whitish/colourless in A. ampelopsifoliella and oinophylla); from Antispila species by the black dispersed frass rather than brown and the markedly smaller cut-out. The record of an Antispila sp. from similar mines (see Remarks) complicates this, though.
Host plants. Vitaceae: Parthenocissus quinquefolia, P. vitacea, Vitis vulpina. Leafmines. (Figs 97-106) Egg in the majority of examined mines in leaf margin, occasionally anywhere in the leaf, usually not associated with veins. The mine starts almost immediately as a blotch, occasionally with a minute linear portion along leaf margin; blotch rather compact, often along leaf margin or slightly extending towards  middle, in narrow leaflets until midrib. Frass black, a thick clump near origin, more dispersed in centre of mine. The larva cuts out an elliptic case ca. 3.5-4 mm long.
Larva. Yellow-green except for green gut contents, in contrast with colourless or whitish larva of A. ampelopsifoliella on the same host plant (compare Figs 69, 70 with 97-100); head and prothorax dark brown.
Life history. Larvae found from August to September through most of the range, but in Florida in April. Field caught adults found in July; larvae collected in fall resulted in emerging adults in April-May, the adult from the April larva emerged in May.
Distribution. Canada: Ontario, Québec*; USA: Connecticut, Florida, Kentucky, ? Maine, Massachusetts*, New York, Pennsylvania, ? Texas, Vermont. The previous record from Vermont by Grehan et al. (1995) was incorrect, as the cited material in the collection of D.L. Wagner belongs to an Antispila species. Records from Maine (Brower 1984) and Texas (Frey and Boll 1878) are uncertain; the last authors give a description but did not mention the white-tipped antenna.
Remarks. Antispila viticordifoliella was described on the basis of larvae and leafmines on Vitis vulpina (as V. cordifolia), with insufficient detail to recognise the species. The first adults were described by Chambers (1874a), again as new species. He clearly described a moth, reared from the same host, without an apical spot and antennae "with about six terminal joints silvery white". His description has later served as basis for the identity of this species, and one of the specimens in MCZ from his collection, incorrectly termed "types" might probably be the best to serve as Neotype, after a male has been dissected and DNA been extracted. Unfortunately the specimens are rather worn, and antennae are missing. The name has also several times been incorrectly used for one of the true Antispila species on Vitis, A. isabella Clemens, 1860 or a related species (see . Earlier we excluded the specimens reared from Parthenocissus from this species ), but we are now more convinced that they are conspecific. Since then it has become clear that Vitaceae-feeding Heliozelidae frequently use more than one plant genus as host (eg van Nieukerken and Geertsema 2015), as we also observe in A. oinophylla. Still we urge that a species level revision using more material should be carried out.
From very similar mines in Parthenocissus collected in North Carolina, CSE and Tracy Feldman have reared an unidentified species of Antispila; we have thus omitted possible records of Aspilanta viticordifoliella from Iowa, Kansas and Ohio that are based only on photographed mines. In photographs of the North Carolina mines taken in the field, the larvae appear to be paler than those of A. viticordifoliella, but this requires confirmation; the Antispila mines are otherwise only recognisable when the typical Leafmine. Not examined. Life history. Larvae in June and July, single adult reared in July. Possibly bivoltine. Distribution. USA: Connecticut, Florida. Barcode. BIN: BOLD:ABA2832, average/max. distance 0.35% (n = 2), distance to NN 9.72% (BIN: BOLD:AEA4069, unknown, not a public record).

Discussion
The phylogenetic basis for the erection of the genus Aspilanta is the transcriptomic study, which provides a high support for its monophyly (Milla et al. 2019). Unfortunately, we were unable to find more than one possible morphological synapomorphy, the presence of an apical spot in the forewing. Possibly, a more detailed study of immature stages would be able to find more apomorphies. For solving the polyphyly of the genus Antispila in the old sense we considered also two alternative options: 1) The inclusion of the whole Holocacista group in an enlarged genus Antispila. Although this would have solved the problem of polyphyly, it would create a large genus with huge differences in many characters, both encompassing species with reduced venation and the more complete venation, and make recognition of species belonging to this genus almost impossible. 2) Combining all genera of the Holocacista group into one genus. Such a genus would be monophyletic and be based on a whole suite of characters, such as the reduced venation. Such a genus should also encompass the undescribed group I (Milla et al. 2017) and the genus14, also known by the manuscript name "Neospila" (Milla et al. 2019). The genus should then be named Coptodisca, as this is the oldest name.
As this name has been in use for more than a century for a rather specialised group of North American leafminers, with a characteristic wing pattern, it would be unfortunate to widen this and include grapevine pests in other parts of the world. Despite the suite of characters, recognition of adults would remain problematic.
The choice of generic borders is in principle not ruled by any other criterion than monophyly, and thus partly subjective. In all three options a number of changes in names are to be expected, but in the chosen option this is limited to just six species, only one of which is an economically important insect. In alternative 1) the number of new combinations would be highest, 26 in total (all Coptodisca, Holocacista and some others, including five or more economically important species), in 2) still 15 are needed (Holocacista, the current Aspilanta species, Antispilina, Ischnocanaba, including at least three economically important species). So in conclusion, the choice for a new genus Aspilanta was obvious in terms of stability and recognisability of the groups. Also biogeographically this choice has advantages, as we have now two old world genera (Holocacista, Antispilina) and two new world genera (Aspilanta, Coptodisca), whereas Antispila remains almost global (unknown from Australia, and so far from the Neotropics). Further study of the Neotropical fauna is important to see whether the monophyly of Aspilanta will stand, especially in the light of the position of the Patagonian group of species cited as "Neospila". If they belong to Aspilanta, which seems likely morphologically, the paraphyly of the combination Genus14 + Antispila group II as noted by Milla et al. (2019) could be problematic. However, it should be noted that the position of Genus14 was poorly supported in that study, and in fact the alternative of a monophyletic Genus14+ Antispila group II was found in one of the analyses (Milla et al. 2019: Fig S5). An analysis comprising several Neotropical species would be necessary for elucidating this phylogeny.
We choose to describe this genus without a full revision of its species. For several of the species we lack sufficient material, and additional fieldwork is much needed, not only for this genus, but for all North American Heliozelidae. Additionally, in Antispila, Coptodisca, and Heliozela we already know of several unnamed species, partly only from Malaise trapped specimens for barcoding campaigns, but also online photographs suggest there are more species. Inclusion of the rich Mexican fauna in such revisions would also be desirable. Further it would be important to settle the status of some old names by dissecting old types while extracting DNA, and for some species to designate neotypes.
This species is rare; since the unique holotype was reared, we are only aware of a few specimens reared by David L. Wagner (pers. comm.); all subsequent literature records to Dyar (1901) are simple checklist entries. According to Busck (1900) the host plant was Eugenia spec., which was according to Dyar (1901), who collected the mine, probably E. procera (identified by F. Kinzel), a misapplied name for the Red stopper, Eugenia rhombea Krug & Urb. ex Urb. However, we doubt that this identification is correct, as E. rhombea is a very rare tree only of the Florida Keys, and not currently occurring in Palm Beach, where only E. axillaris (Sw.) Willd. and E. uniflora L. have been recorded (Wunderlin and Hansen 2003;Wunderlin et al. 2020). However, we cannot exclude that it did occur there in 1900, as most of the original habitat is now probably lost to development. We illustrate the holotype here (Fig. 107): a moth with a small white dorsal spot at 1/3 and a fascia at 2/3, which is narrowest at dorsum and widens towards costa.
The colour pattern of Antispila and Aspilanta is almost always with a fascia at one third and two opposite spots at two-thirds. The pattern of H. eugeniella is different, with a postmedial fascia, and resembles more some species of the Old World genus Holocacista (van Nieukerken and Geertsema 2015). Most species of Heliozela have a basic pattern of two dorsal spots, but there are some species where one or both of these extend towards the costa, e.g., H. anna (Fletcher, 1920) from India with two fasciae and H. argyrozona (Meyrick, 1918) from South Africa, with a pattern very much resembling that of H. eugeniella (van Nieukerken and Geertsema 2015). Both were reared from Myrtaceae, the latter from Syzygium cordatum Hochts. ex Krauss, and H. anna from Eugenia jambolana Lam. (Fletcher 1920), the correct name of which is now Syzygium cumini (L.) Skeels (Malabar plum, Jambolan). Also in Australia and Southeast Asia Myrtaceae is a major host family for Heliozela (Milla et al. 2017(Milla et al. , 2019. These facts suggest that eugeniella is better placed in Heliozela than in Antispila, although still further morphological and molecular confirmation is needed.