Hawaiian Philodoria (Lepidoptera, Gracillariidae, Ornixolinae) leaf mining moths on Myrsine (Primulaceae): two new species and biological data

Abstract This paper provides new taxonomic and biological data on a complex of gracillariid moths in the endemic genus Philodoria Walsingham, 1907 that are associated with Myrsine (Primulaceae) in the Hawaiian Islands, United States. Two new species, Philodoria kauaulaensis Kobayashi, Johns & Kawahara, sp. n. (host: Myrsine lanaiensis, M. lessertiana, and M. sandwicensis) and P. kolea Kobayashi, Johns & Kawahara, sp. n. (host: M. lessertiana) are described. Biological data are provided for two previously described species that also feed on Myrsine: P. auromagnifica Walsingham, 1907 and P. succedanea Walsingham, 1907. For the first time we detail and illustrate genital structures, immature stages, biology, and host plants of P. auromagnifica and P. succedanea. Philodoria kolea, P. auromagnifica, and P. succedanea occur in sympatry on the island of Hawaii (Big Island), but each species differs in behavioral characters: P. kolea utilizes leaves of seedlings and forms a serpentine mine, whereas the latter two utilize leaves of larger plants, and form linear or serpentine to blotch mines. More broadly, leaf mine forms and diagnostic characteristics of the Myrsine-feeding species complex of Philodoria (as currently known) are reviewed and illustrated.

Hawaiian Philodoria (Lepidoptera, Gracillariidae, Ornixolinae) leaf mining moths on Myrsine (Primulaceae): two new species and biological data Introduction Hawaii constitutes one of the most geographically isolated archipelagos and harbors thousands of unusual, highly threatened endemic species. Phytophagous insects that rely on endemic Hawaiian plants are of special risk as they depend on the survival of their native host plants. The Hawaiian Islands measure just 0.02% of the area of the United States, but account for nearly 70% of the United States' historically documented plant and animal extinctions (Wagner et al. 1999). In all, over 360 Hawaiian animal and plant taxa are currently listed as either threatened or endangered under the federal and state Endangered Species Acts. More than 38% of native Hawaiian plants are threatened and 94% of Hawaiian insects are endemic (Evenhuis and Eldredge 1999). Leaf miners have achieved extraordinary localized diversity and are a major component of island ecosystems throughout the Pacific.
Philodoria Walsingham, 1907 is a genus of endemic Hawaiian leaf-mining micromoths, containing approximately 30 species, for which the classification remains largely in disarray. The genus can be distinguished from other genera in the Gracillariidae subfamily Ornixolinae by a hindwing with small frenular bristles along the costa in both sexes (Zimmerman 1978, figs 432-435); by a dorsal flap extending from the posterior margin of tergum VIII in the male; and by the female lamella antevaginalis that is sclerotized and semicircular in shape. Many Philodoria host plants are threatened along with their native habitat. Indeed, herbarium samples provide one of the few documented cases globally of a probable moth extinction, albeit an undescribed species (Johns et al. 2014). The genus was first described with seven species by Walsingham (1907), and the type species was designated as P. succedanea Walsingham, 1907. Zimmerman (1978 published a monograph of Hawaiian insects following Walsingham's work and many papers by Swezey (1910Swezey ( -1946. Zimmerman divided Philodoria into two subgenera, P. (Eophilodoria) and P. (Philodoria), based on the size of the maxillary palpus. His classification was recently rejected by Johns et al. (2016), who constructed a preliminary molecular phylogeny of Philodoria based on three genes for 11 Philodoria species. In their analyses, the two Philodoria subgenera were not monophyletic and morphological characters used to classify them were inferred as homoplasious; the subgenus Eophilodoria Zimmerman, 1978 was established as a subjective junior synonym of the genus Philodoria Walsingham, 1907. In addition, Johns et al. (2016) provided new host plant and distribution data for these 11 species. While Philodoria was historically treated as similar to Elachista (Elachistidae, Gelechioidea), it unequivocally belongs in Gracillariidae (Kawahara et al. 2017) and the genus is unrelated to Gelechioidea (Breinholt et al. 2018). Based on taxon sampling of exemplar gracillariid genera, Philodoria appears to be phylogenetically closely related to the ornixoline genus Chileoptilia Vargas & Landry, 2005 from Chile (Kawahara et al. 2017).
Larval host plants of Philodoria are diverse, with up to six plant orders (Asterales, Apiales, Ericales, Malvales, Myrtales and Rosales) reported as hosts, among which Asterales (Asteraceae: Dubautia) and Rosales (Urticaceae: Pipturus) appear as dominant hosts (Swezey 1954;Zimmerman 1978). Another host plant that is used by multiple Philodoria species is Myrsine (Ericales: Primulaceae). According to Zimmerman (1978) and label data from Philodoria specimens in the collection of the Bernice Pauahi Bishop Museum (BPBM), there appear to be numerous undescribed Philodoria species on Myrsine. In total, 19 Myrsine species are known to be endemic to the Hawaiian Islands (Wagner et al. 1999), and two species of Philodoria that feed on Myrsine have been described: P. succedanea Walsingham, 1907 (type species of the genus) and P. auromagnifica Walsingham, 1907, both with similar scale colors and genital characters (Walsingham 1907, Zimmerman 1978. In late April 2016, several of the authors collected numerous blotch mines on leaves of Myrsine species at two sites on the island of Hawaii (Big Island). Initially, we believed that these mines were created by a single Philodoria species, but after studying them, we realized that they comprised diverse larval habits (e.g., forms with spiral or linear mines, larvae in fallen or in situ leaves, and some adults which emerged with relatively black forewings). Recent studies (Kawahara et al. 2009, Davis and Wagner 2011, Davis and De Prins 2011, Brito et al. 2013, Moreira et al. 2017) have shown that important diagnostic characters of gracillariids are present in larvae and pupae. However, insufficient early stages have been preserved until now for diagnostics and identification. In this paper, we describe two new species, Philodoria kauaulaensis (hosts: Myrsine lanaiensis, M. lessertiana, and M. sandwicensis) and P. kolea (host: M. lessertiana), and also the genitalic structures, immature stages and new host plant information for the two previously described Myrsine-feeding species, P. succedanea and P. auromagnifica. Four species were reared, and their mine forms and characters are here reviewed and illustrated.

Taxon sampling
All adult moths were reared from leaf mining larvae and their pupal cocoons. Leaf mines and cocoons were collected between 2013-2016 in the locations listed in Table 1. Among the material examined, the final dates refer to the adult emergence and 'em.' signifies that an adult emerged and was mounted as a dry pinned specimen; 'stored' signifies a dead adult that was stored in 99 % ethanol or RNAlater solution (Thermo Fisher Scientific). Type material designated by Lord Walsingham and specimens collected by Dr K. & Mrs. E. Sattler in the Natural History Museum (NHMUK), and those collected by Mr. O. H. Swezey at the BPBM and the National Museum of Natural History, Smithsonian Institution (USNM) were also examined. Immatures in leaves were reared in plastic cups (420 ml: 129 mm in diameter at top and 60 mm in depth) containing wet cotton at 20 ± 5 °C under a photoperiod condition in the laboratory of 13-16L (hours light) 8-12D (hours darkness).

Morphology and nomenclature
Descriptions focused on the adult stage and leaf mines because of limitations of other material, and because these stages provide a wealth of morphological traits useful for diagnosis. Photographs of leaf mines were taken primarily in the field using Canon EOS 60D and 5D MKIII digital cameras. Some leafmines were scanned using an EPSON Perfection V600 Photo scanner. Observations and measurements were made under a Leica M2 16 dissection microscope at 71-115× and a Leica S6E microscope at 6.3-40× with the aid of a micrometer scale. Images of adults were captured using a Olympus E-330 camera and Moticam 580 5.0 MP. Images were taken at various depths and subsequently stacked using the Helicon Focus 6.22. All images were then edited with Adobe Photoshop Elements 9 into final figures. For genitalic dissections, the whole abdomen was removed and boiled for 3-4 min in 10% aqueous KOH, and residual scales and soft parts were removed in 70% ethanol. Genitalia were then stained in Chlorazol Black E (1% solution in 70 % ethanol) or acetocarmine for 0.5-1h, dehydrated in a series of 70−100 % ethanol and mounted in Canada balsam on a glass slide.
Type material and additional specimens used in the present study are preserved in the collections of the BPBM, the McGuire Center for Lepidoptera and Biodiversity, Florida Museum of Natural History (FLMNH) and Naturalis Biodiversity Center (RMNH). Terms used for describing wing color pattern are summarized in Fig. 1, and forewing characters follow the terminology of Walsingham (1907) and Zimmerman (1978). Terms for genitalia essentially, follow Zimmerman (1978) and "gnathos" is employed to indicate the sclerotized V-shaped transverse band joining the ventral base of tegumen. Scientific names of plants follow the Plant List (www.theplantlist.org).

DNA sequencing and analysis
A total of 16 specimens were DNA barcoded. DNA extraction, PCR amplification and sequencing of the 658 base pair Cytochrome Oxidase 1 (COI) "barcode" region for two specimens were carried out at the Canadian Centre for DNA Barcoding (CCDB, Biodiversity Institute of Ontario, University of Guelph) following a published protocol (deWaard et al. 2008). Five specimens were extracted at the Florida Museum of Natural History, McGuire Centre for Lepidoptera and Biodiversity at the University of Florida, Gainesville, FL, USA, using the OmniPrep extraction kit and sequenced at University of Florida's Interdisciplinary Center for Biotechnology Research (ICBR), one specimen was extracted at the Department of Life and Environmental Sciences, Kyoto Prefectural University, Shimogamo, Kyoto, Japan (KPU) using the DNeasy Blood & Tissue Kit (Qiagen, Inc., Valencia, California), and single-stranded PCR and sequencing for this specimen was carried out at the Operon Sequencing Center following the manufacturer's protocol (Eurofins, Tokyo, Japan). Eight specimens that were sequenced at Naturalis Biodiversity Center were extracted using a Macherey-Nagel magnetic bead DNA extraction kit on a KingFisher automated DNA extraction robot (Table 2).
We conducted an ML analysis of the COI gene using RAxML 8.2.10 (Stamatakis 2014), searching for the best tree using the GTRCAT model and GAMMA-based likelihood optimization for the final tree, and otherwise default settings. Subsequently, 1,000 parametric bootstrap analyses with automated stopping following the extended  (Rambaut 2009). Intra-and interspecific genetic distances were estimated using the Kimura 2-parameter model implemented within the analytical tools available in BOLDv4 (Table 3). We also used BOLD to obtain Barcode Index Numbers (BINs) (Ratnasingham and Hebert 2013). Kimura 2-parameter (K2P) distances (%) for barcode DNA sequences of the seven analyzed species in the genus Philodoria; minimal pairwise distances between species are given for each species pair; values in square brackets represent maximal intraspecific distances.
While single-marker COI analyses can be prone to insufficient resolution and error (Rubinoff and Holland 2005), we were unable to obtain additional genetic data for these species during the time of this study. We therefore chose to use a gene-tree based approach (Hebert et al. 2003;Hajibabaei et al. 2007) as another source of evidence to complement morphology to assess species limits. Sequences, voucher data, images, and trace files are deposited in the Barcode of Life Data Systems (BOLD) (Ratnasingham and Hebert 2007;www.barcodinglife.org). Furthermore, all sequences are deposited in GenBank, and are available as a single dataset DS-PHDRIA (http://dx.doi.org/10.5883/DS-PHDRIA)

Voucher specimen numbers
Institutional voucher numbers are given here for primary type material and museum collections. In the cases of NHMUK/BMNH numbers, for clarity and consitency they are cited without a space nor hash symbol (#) that might be read between the alpha and numeric parts of the code, since spaces and hashes create ambiguity for search, and series of institutional numbers have appeared in the past with or without such symbols Abbreviations for collections:

Key to adults
1 Forewing leaden grey, externally with fuscous brown (Fig. 3)    This species was described from 19 specimens: 'type ♀ (26695); ♂ (28505)' and 17 'paratypes' from Kauai and Haleakala, Maui. This seems to indicate that Lord Walsingham considered them as holotype, allotype, and paratypes, as indicated on their specimen labels. But as a holotype was not specified in the description, the so-labelled types and paratypes are all to be considered syntypes under the present Code, Article 73.2 (ICZN 1999), and any one is thus eligible for designation as lectotype. The syntype 'type ♀ (26695)', which Walsingham listed first and figured, is here designated as lectotype ( Fig. 2A). The remaining syntypes are therefore paralectotypes.    Diagnosis. This species is very similar to P. auromagnifica feeding on the same hostplant, Myrsine, but is recognizable by the rather bright orange patches and black triangular shaped basal patch in the forewing (Table 4; Figs 2A-D, 5A-D); in the male genitalia by the rather broad valva, slender and long saccus curving toward dorsal side ( Fig. 6A-C); in the female genitalia by signa with slender and long spines (Fig. 7E, F).
Forewing shiny, metallic bronze with bright orange-ochreous patches: a black triangular basal patch along the costal fold (Figs 5A, C, 8A); an oblique transverse fascia before the middle of wing, bordered with black scales; a large transverse patch after the middle to costal 3/4, distinctly narrowing in the dorsum, extending to dorsal 2/3, containing white costal spot; one white color band on the middle of the first   (Zimmerman 1978 andJohns et al. 2016) and see s also pecies descripution. b Plant species and island name in bold indicate new records in the present study. Islands underlined denote type-locality islands.
bronze color band, others on both extremities of second and third bands; a fuscous patch extending toward the termen and apex with a black apical spot, sometimes with orange-ochreous color encroaching on the apical part; cilia tawny, with two metallic silver basal lines, one at the apical cilia, another from termen to tornus. Hindwing dark tawny; cilia tawny. Abdomen tawny above, silvery beneath. Legs tawny, with silvery spurs and slightly paler tarsi. Male genitalia (Fig. 6A-D) (n = 7). Capsule 960 µm. Uncus absent. Tegumen 570-580 µm long, 1.2-1.3× length of valva with series of long hairs at lateral side of base (Fig. 6C). Tuba analis membranous with weakly sclerotized subscaphium; gnathos V-shaped transverse band, terminal margin weakly joining subscaphium and anterior process connecting ventral base of tegumen. Valva broad, 430 µm in length covered with fine setae distally, and having a short dorsal process (Fig. 6A).Vinculum U-shaped; saccus 250 µm long, slender , curved toward dorsal side (Fig. 6B, C). Phallus 720 µm long, tubular and long about 1.2-1.3× length of valva, sinuous in lateral view with two series of minute spiniform cornuti in vesica; coecum slightly curved toward inner side (Fig. 6D).
Female genitalia (Fig. 7E, F) (n = 7). Ostium bursae rather small, opening at the middle of 7 th abdominal segment; antrum cup-shaped with slender a pair of lateral lobes; ductus bursae slender, tubular, extremity connected to antrum very slender and membranous, curved inside of body, and middle part weakly sclerotized and plateshape; end of the ductus bursae broad; inception of ductus seminalis on the posterior part of ductus bursae. Corpus bursae pyriform, anterior end weakly sclerotized; some  Table 1 and alphabetical symbols (A-K) correspond to figure numbers in Figure 5. Biology. (Figs 8A, 9, 14A). The larvae mine the adaxial side of leaves of Myrsine species, forming a long linear mine (Fig. 9B, G, H). The mine is at first tornus-shaped (Fig.  9C, D, I, J) and the larva broaches the mid vein towards the petiole of the leaf, forming a straight mine; the vein mine and surrounding pattern are red in coloration (Fig. 9F, H) and later instars leave the mid vein usually near the base of the leaf, gradually expanding as they feed and grow forming a full-depth mine (Fig. 9E, F). There were usually one to two mines per leaf (Fig. 9B, G, M). The pupal cocoon is situated outside of the mine, usually on the leaf surface, and also on the woody tissue of the host plant with leaf mines and larvae. At Hawai'i Volcanoes National Park, larvae were collected from leaves that had fallen   to the ground and reared to adulthood (Fig. 9A). The adult has been observed during the day (Maui and Hawaii Island), resting on the upper leaf surface of the host plant (Fig. 8A).
DNA barcoding. BIN BOLD:ADF5435. The two specimens sequenced for COI, one from Maui and one from Hawaii, have identical DNA barcode sequences. The pdistance to the nearest neighbor, P. kauaulaensis, is 6.63%.
Remarks. We identified two adult moths (Coll ID CJ-144 / GenBank accession no. ID KT982414 and CJ-145) as P. succedanea, based on the presence of a basal black patch on forewing, from which whole bodies were sacrificed for molecular analysis (Johns et al. 2016;Figs 6O, 12). Zimmerman (1978) did not recognize Walsingham's (1907) Kauai record of this species because Walsingham had only one specimen at hand, which was in poor condition (specimen data: 1 ♂, Mts [which Mts not further specified], 3-4000 ft., Kauai, vi. 1894 Perkins.27297| PARATYPE 17/17 (?)|'NOT succedanea Det. by E. C. Zimmerman|NHMUK010862812). We could not find the specimen from Kauai. However, we found Myrsine knudsenii (Endangered, IUCN) leaves with mines with active larvae from Kokee, Kauai Is. 442), which were similar in appearance to mines of P. succedanea on M. lessertiana. Judging from these data, we consider the larval mines on M. knudsenii were made by P. succedanea. We also collected active Philodoria leaf mines from M. linearifolia (Endangered, IUCN) at the same location as M. knudsenii, but were unable to rear adult moths. It is thus possible that P. succedanea also mines M. linearifolia, but this needs to be further examined.  holotype is incomplete but we consider it distinctive enough to be worth describing. What remains of the holotype was mounted by placing three wings without mountant under a coverslip: two forewings (3/4 of right wing and half of left wing), and the apical portion of one hindwing (Fig. 5I). The head, antenna, thorax, and legs were sacrificed for molecular analysis.
Description. Adult (Fig. 5I, J). Forewing length 2.4 mm, basal part of holotype forewing missing. Descriptions of the basal forewing and part of the body were based on photographs of adult moths (CJ-064, 072). Head and frons fuscous; maxillary palpus unknown; labial palpus ochreous. Antennae dark fuscous. Thorax unknown. Forewing shiny, metallic bronze with three large bright orange transverse fascia: an oblique one from costal fold to dorsal 1/4; a second at the middle of wing, narrowing greatly in the dorsum, containing a white costal spot; a third at 3/4 in the middle, interrupted by a blue patch; all fascia bordered with black scales: one white color band at middle of the first bronze color band, others on both extremities of third and fourth bands; a fuscous patch extending toward termen and apex with a black apical spot; cilia shiny, dark bronze grey. Hindwing dark tawny fuscous. Abdomen fuscous above, whitish beneath.
Male genitalia. Unknown. Female genitalia. (Fig. 7H) (n = 1). Similar to P. succedanea and P. auromagnifica, but different in having rather smaller and rounded spines on the signa. Biology. (Figs 10, 14B). The mine is initially spiral-shaped (Fig. 10B, C) and gradually expands as the larva feeds and the mine later gets the form of a blotch (Fig. 14B). The pupal cocoon is situated outside the mine, usually on leaf surface (Fig. 10D).
DNA barcoding. BIN BOLD:ADI5327. The two specimens sequenced for COI are from Maui and have a 0.17 p-distance between them, the p-distance to the nearest neighbor, P. auromagnifica, is 5.58%.
Etymology. The specific epithet is derived from the type locality, Kaua`ula (pronounced 'cow-wa-u-la') Valley, an important site for Hawaiian endemic plants and culturally and spiritually for Native Hawaiians.
Remarks. Johns et al. (2016) collected larvae from Myrsine lessertiana and M. sandwicensis in West Maui and identified the reared adult moths as P. auromagnifica (Coll. ID CJ-064 / GenBank accession no. KT982404 and CJ-072 / KT982407). Comparison of adult morphology and larval behavior with other species shows that these moths belong to P. kauaulaensis (Figs 10, 14). Unfortunately these specimens were sacrificed for molecular analysis, so that they cannot be added to the type series. Walsingham, 1907  Figs 2E, F, 5E-H, 6E-H, 7G, 8B, 11, 14C Philodoria auromagnifica Walsingham, 1907: 718, pl. 25, fig. 20;Swezey 1913b: 223. Philodoria (Philodoria) auromagnifica Walsingham, 1907: Zimmerman 1978 1910-427.|NHMUK010305330| in NHMUK. This species was described based on a single specimen from Oahu. The 'type' specimen, designated by Walsingham is here thus the holotype following article 73. 1.2 (ICZN 1999). Diagnosis. This species is very similar to P. succedanea, but recognizable by the dark brownish orange patches and brownish orange basal patch in the forewing; a fuscous patch with dark orangish scales in the apical portion (Table 4; Figs 2E, F, 5E-H); in the male genitalia by the rather long valva narrowing in the middle, vinculum large, inflexed on the ventral side, broad and straight saccus (Fig. 6E-G); in the female genitalia by signa with rather blunt spines (Fig. 7G). See also diagnosis of P. succedanea. Redescription. Adult (Fig. 2E, F). Wingspan 8 mm in holotype, 7-9 mm in other specimens; forewing length 3.5 mm in holotype, 3.2-3.9 mm in others. Head and frons dark steely fuscous; maxillary palpus reduced; labial palpus ochreous to brown. Antenna dark fuscous. Thorax: dark brownish orange, becoming fuscous posteriorly. Forewing shiny, metallic bronze with dark brownish orange patches: a large one at base bordered with black ground color (Figs 2E, F, 5E, F, H), sometimes missing orange color (Fig. 8B); an oblique transverse fascia before the middle of wing, bordered with black ground color, sometimes missing orange color (Fig. 11A); a large transverse patch after the middle to costal 3/4, narrowing greatly in the dorsum, extending to dorsal 2/3, containing a white costal spot; one white color band on the middle of the first bronze color band, others on both extremities of second and third bands; a fuscous patch mixed with dark brownish orange scales extending toward the termen and apex with a black apical spot; cilia shiny, dark bronze grey. Hindwing dark tawny fuscous. Abdomen and legs fuscous above, white beneath.
Female genitalia (Fig. 7G) (n = 7). Similar to P. succedanea, but different in having rather slender tapering antrum and rather blunt spines on the signa.
Distribution. Kauai: new record, Oahu (Walsingham 1907), Molokai (Swezey and Bryan 1929), and Hawaii (Big Island) (Zimmerman 1978). Biology. (Figs 8B, 11, 14C). The larvae mine the adaxial side of leaves of Myrsine species, forming a long serpentine mine (Fig. 11B) and gradually expanding as they feed (Figs 11C, 14C2, C3). Old mines are ocherous to brown in coloration (Fig. 14C1). There were usually one to two mines per leaf (Fig.11B). The pupal cocoon is prepared outside the mine, on either surface of the leaf, and one was found on the bark of the host. DNA barcoding. BIN BOLD:ADD6965. The two specimens sequenced for COI are from Hawaii and diverge by 0.31%, whereas the p-distance to the nearest neighbor, P. kauaulaensis, is 5.58%.
Description. Adult (Figs 3, 5K, L, 12N). Wingspan 6.7 mm in holotype, 6.6, 8.5 mm in paratypes; forewing length 3.0, 3.1 mm in holotype, 2.7-4.0 mm in paratypes. Head leaden grey; frons whitish grey; maxillary palpus reduced; labial palpus greyish ochreous, terminal joint with fuscous band at middle and at apex. Antenna greyish fuscous. Thorax leaden grey. Forewing base leaden grey, externally suffused with brownish fuscous patches: a triangular basal patch along the costal fold; an oblique transverse fascia before the middle of wing, bordered with black scales; a large transverse patch after the middle to costal 3/4, narrowing in the dorsum, extending to dorsal 2/3, containing small white costal spot; leaden grey small median line at base with dorsal narrow patch from base to near middle; one white color band at the middle of the first bronze color band, others on both extremities of second and third bands; a leaden grey patch extending toward the termen and apex with small shiny black apical spots; cilia leaden grey with a black fringe basal line; tonal cilia with a shiny white fringe basal line. Hindwing and cilia leaden grey. Abdomen greyish fuscous above, banded with white beneath. Legs pale greyish fuscous, spurs white.
Distribution. Hawaii (Big Island). Host plants. Primulaceae: Myrsine lessertiana A. DC. Biology. (Figs 12, 13, 14D). Larvae mine the adaxial side of leaves of M. lessertiana, forming a slender serpentine mine (Fig. 12A, B), and gradually expanding as they feed and grow forming a full-depth mine (Fig. 12G, L). Larvae consumed small amounts of leaf tissue (under 2 cm in leaf length) when feeding on seedlings (Fig. 13G, H). The young larva is about 1.5 mm long (Fig. 12I) and later instar larvae are 4-8 mm long (Fig. 12J, K). Larvae were collected from fresh leaves of seedlings. There was usually only one mine per leaf (Fig. 12A, B, G, H). The pupal cocoon is prepared outside the mine, on either surface of the leaf; the cocoon is greyish white to ochreous and near ellipsoidal in shape (Fig. 12M); 4.0-5.0 mm in length, 1.0-3.0 mm in width.
DNA barcoding. BIN BOLD:ADF137. The five specimens sequenced for COI are from two localities on Hawaii and have maximum intraspecific p-distance of 0.17%. The p-distance to the nearest neighbor, Philodoria kauaulaensis, is 6.98%.
Etymology. The specific epithet, kolea, is a noun in apposition taken from the Hawaiian name of the host plant, Myrsine.

Discussion
Hawaiian Philodoria leaf mining moths were extensively studied in the early 1910s-1940s by Otto Herman Swezey. However, little taxonomic work has been conducted since, and our investigation is revealing that several undescribed cryptic species remain to be discovered, as found in other Hawaiian micromoths (e.g., Bedellia, Bedelliidae: [Zimmerman 1978]; Hyposmocoma: Cosmopterigidae [Kawahara and Rubinoff 2012;Rubinoff 2008]). Philodoria is critically in need of taxonomic work considering the endemic distribution of its species on the Hawaiian islands, and the close association of the genus with native endemic and endangered host plants. Some host plants and their associated Philodoria have already become locally extinct (Johns et al. 2014).
Swezey collected Myrsine-feeding P. succedanea and P. auromagnifica from numerous localities on Oahu in the early 1900s. Myrsine lessertiana plants remain relatively abundant on Oahu, but Myrsine-mining Philodoria have become exceedingly difficult to find there, especially in the southeast where intense urban development has taken place over the last century. During our Oahu surveys, we were unable to find leaf mines on M. degeneri, M. fosbergii, M. juddii (Critically Endangered, IUCN), M. lanaiensis, M. pukooensis, M. punctata, or M. sandwicensis, despite extensive searches for leaf mines on these host plants. It is not clear whether these absences are more due to environmental changes causing population reductions than to original host plant restriction among Myrsine species.
On Maui, P. kauaulaensis and P. succedanea were found in April-May 2013 at two sites separated only by 3.3 km, below the summit of Eke and on Haelaau Ridge, within the Pu'u Kukui Watershed Preserve ( Fig. 2A; Johns et al 2016, fig. 1, Coll. ID CJ-064, CJ-072). In the present study, we observed P. auromagnifica, P. kolea and P. succedanea occurring in sympatry on April 2016 at the Hawai'i Volcanoes National Park, the island of Hawaii (Big Island) (Figs 2A, 6A-J, 8B, C, 9A-M).
We collected larvae of P. auromagnifica (Fig. 8B, C) on plants that were also used by P. succedanea (Fig. 6A). The latter species was still mining leaves from the same plants that had fallen to the ground. Larvae of P. kolea occurred on leaves that were intact on short (about 10-20 cm high) Myrsine plants at the same site (Figs 9A, B, 10D). The genetic similarity between these species could imply that perhaps competition and niche partitioning may have been the cause of speciation. Fine-scale niche partitioning has been documented in other gracillariids and their host plants, such as Phyllocnistis on Persea (Davis and Wagner 2011) and Phyllocnistis on Salix (Kobayashi et al. 2011). Our ongoing research efforts will examine the evolutionary history and colonization patterns of Philodoria on the Hawaiian archipelago.
In addition to providing morphological and molecular evidence to delimit species limits among the Hawaiian Myrsine-feeding Philodoria, we include a pictorial key to their leaf mines (Fig. 14). We include this information as leaf mining moths can be difficult to observe as larvae or adults to a non-specialist. Larvae of P. succedanea form red, long linear mines along the leaf vein (Fig. 14A), P. kauaulaensis produces at first spiral and later blotch mines (Fig. 14B), P. auromagnifica makes brown serpentine mines (Fig. 14C), and P. kolea creates complete serpentine mines fully occupying the adaxial side of leaf surface of Myrsine seedlings (Fig. 14D). We hope that local Hawaiian park rangers, naturalists, and educators can use this key as a means to identify these species, so that the collection of these much-needed data can persist.
It is likely that detailed molecular work among islands will reveal further cryptic species but native hostplants and habitats are under great threat.