Review of the North American fauna of Drymeia Meigen (Diptera, Muscidae) and evaluation of DNA barcodes for species-level identification in the genus

Abstract The North American fauna of Drymeia Meigen was studied. Four new species are described (Drymeia huckettisp. nov., Drymeia pontisp. nov., Drymeia vockerothisp. nov., Drymeia woodorumsp. nov.), and three new synonymies are proposed: Drymeia amnicola (Huckett, 1966) (= Drymeia rivalis (Huckett, 1966), syn. nov.); Drymeia glacialis (Rondani, 1866) (= Drymeia alpicola (Rondani, 1871), syn. nov.); and Drymeia spinitarsis (Aldrich, 1918) (= Drymeia longiseta Sorokina & Pont, 2015, syn. nov.). An annotated checklist, DNA barcodes (when available), and keys for each sex of the 24 named species of North American Drymeia are provided. The utility of DNA barcodes for the identification of Drymeia species across a wide geographical range was explored using sequences from five countries. A match between morphology and DNA barcodes was found for 71% (22/31) of species studied (including three unnamed taxa). The remaining nine species clustered into two groups of taxa with very little interspecific variation within clusters (groups of two and seven species). Caution is advised against using DNA barcoding as the only determination tool for Drymeia material without prior knowledge of its limitations for certain species groups.


Introduction
The genus Drymeia Meigen, 1826 includes more than 130 species of mid-sized muscids that can usually be distinguished from other members of the Muscidae by the combination of a dilated and densely setose gena, proclinate orbital bristles in the female, and three or more posterodorsal bristles on the hind tibia. An apical ventral spur of variable size can be found on the hind tibia of males in many species (see Fig. 2) and some taxa such as Drymeia pribilofensis (Malloch, 1919) and D. spinitarsis (Aldrich, 1918) exhibit elongated mouthparts (see Fig. 8A, B).
While these flies can be collected using various passive collecting devices, net sweeping will often yield a good range of species since many can be found basking on sunny rocks or visiting flowers. Many species of Drymeia are known to consume pollen in addition to nectar, and Michelsen (2011) demonstrated that the morphology of the labellum has been modified to that effect in several species, including some found in the Nearctic region such as Drymeia groenlandica (Lundbeck, 1901) and D. segnis. The contribution of various Drymeia species to pollination services has not been widely investigated but in a study of the role of insect visitors in the pollination of Dryas in Northern Greenland, Tiusanen et al. (2016) demonstrated that in spite of comparable abundances and frequencies of flower visitation, Spilogona sanctipauli  had a much stronger association with seed set than D. segnis.
Not much is known of the immature stages of Drymeia. Some larvae can be found in humus soil or cow dung and Skidmore (1985) reported that the larvae of the few species examined are dimorphic obligate carnivores with an extremely elongate and highly sclerotised cephaloskeleton.
The vast majority of North American species were originally described under Pogonomyia Rondani and, to a lesser extent, Bebryx Gistl, Trichopticoides Ringdahl, Eupogonomyia Malloch, and Pogonomyioides Malloch before these taxa were synonymised with Drymeia by Pont (1986), and Huckett and Vockeroth (1987). Malloch (1915Malloch ( , 1919 and Huckett (1965aHuckett ( , 1966 were the most important contributors to the Nearctic fauna of Drymeia, describing several species, 14 of which are still valid today (seven from each author).
All taxonomic contributions to the North American fauna of Drymeia to date have focused on local fauna, making it difficult to identify specimens outside of these areas. Consequently, the first objective of this work was to develop comprehensive and well-illustrated keys to both sexes of all North American species of Drymeia. We also aimed to produce an annotated checklist of these species (including new synonymies and updated distribution ranges) and describe four new species. Finally, we explored the utility of DNA barcodes and Barcoding Index Number (BIN) assignments (Ratnasingham and Hebert 2013) for the identification of Drymeia species.

Materials and methods
Drymeia specimens examined in this study are housed in the Academy of Natural Sciences of Philadelphia, Philadelphia, USA (ANSP), the Bishop's University Insect Collection, Sherbrooke, Canada (BUIC), the California Academy of Sciences, San Francisco, USA (CAS), the Canadian National Collection of Insects, Arachnids and Nematodes, Ottawa, Canada (CNC), the Centre for Biodiversity Genomics, Guelph, Canada (BIOUG), the Essig Museum of Entomology

Morphology
Morphological terms follow McAlpine (1981) for external structures and Savage and Wheeler (2004) for male genitalia. Body length was measured in millimetres (mm) from the anterior margin of the head without antenna to the apex of the abdomen. Frons length was measured between lower margin of frons and upper margin of ocelli triangle. Genitalia were cleared using lactic acid and placed in a microvial pinned under the specimen following dissection. Whole specimens and external structures were photographed with a Leica DFC450 or Leica DFC5400 camera using Leica Application suite X (LAX) and images were stacked using LAS X and Zerene Stacker 1.04. Male genitalia were photographed using a Luminera 1 camera and images were stacked using Helicon Focus 6.3.0 (Helicon Soft Ltd., Kharkov, Ukraine). Final image editing and plate assembly was completed in Adobe Photoshop CS.
The following abbreviations are used in the text: Label data of primary type material examined are presented verbatim, with "/" indicating a change of line and ";" a change in label.

DNA barcoding
The Barcode Index Number (BIN) system uses the Refined Single Linkage (RESL) algorithm to group COI sequences in the Barcode of Life Data System (BOLD) (http:// boldsystems.org) into genetic clusters (BINs) representing proxies for species (Ratnasingham and Hebert 2013). While the present work deals mainly with the North American fauna, we assessed the match between BIN assignment and the morphology of current specie concepts globally and therefore included all available sequences regardless of geographical origins. This broad geographical approach had the potential to uncover new species (and/or synonymies) and yield new data on species distribution ranges. If the morphology of specimens from BINs that could not be assigned to a named species was clearly unique and the material in good enough condition, the species was described.
All DNA barcodes (Folmer region of COI gene) were sequenced at the University of Guelph Biodiversity Institute of Ontario following protocols published in Hajibabaei et al. (2005) and the LepF1/LepR1 (Hebert et al. 2004) and LCO1490_t1/HCO2198_ t1 (Foottit et al. 2009) primers were used for amplification. All sequences were at least 550 base pair long with zero ambiguous base and these were aligned in BOLD) using the BOLD Aligner tool (Amino Acid based HHM). A neighbour-joining tree was built in BOLD to provide a graphical representation of pairwise distance between specimens in the data set, detect potential anomalies, and examine BIN assignments. Uncorrected pairwise distances (p-distance) were calculated in MEGA X (Kumar et al. 2018).
In the public dataset "Drymeia specimens examined for Savage and Sorokina 2021" (dx.doi.org/10.5883/DS-DRYNEW) available on BOLD. Specimen details, including BIN assignment, sex, and GenBank accession numbers for this data set can also be found in Suppl. material 1: Table S1. All specimens were examined by the authors except for the only representative of Drymeia hamata (Fallén, 1823) (BOLD process ID: GBDP5828-09); we are, however, highly confident in the identity of that specimen since the unique elongated mouthparts characteristic of the species are clearly visible in the image of the male for this record. Frons at narrowest point at least 1.5 × as wide as ocellar triangle and with frontal vitta widely exposed (Fig. 1A); T3 with ventral apical process short but clearly visible and with apical pv absent ( Fig. 2A Frons broad with black frontal vitta exposed and at least 2 × as wide as fore ocellus (Fig. 1C); parafacial broad in lateral view (Fig. 4B Frons broad with black frontal vitta exposed and at least as wide as fore ocellus at narrowest point; fore and mid coxa with very long fine hair, crinkled near apex; mid tarsomere 1 with spines in av and pv rows long, some longer than width of tarsomere ( F2 with av and/or pv rows variable, but always with some (or many) bristles longer than basal diameter of femur (as in Fig. 6K, L); T3 ventral apical process short but usually distinct (Fig. 2I)  Lower margin of face projecting slightly beyond lower level of frons (Fig. 16A); arista with hairs restricted mostly to basal 1/2 and with no or few dorsal setae on apical 1/2; abdomen lightly dusted and mostly shiny, tergites without distinct dark central vittae (Fig. 7B)  Mid tarsomere 1 with spines in av and pv rows long, some longer than width of tarsomere (Fig. 5C); head in lateral view higher than wide (Fig. 8A)  T3 with apical pv present, at least 1/2 as long as apical av; arista with longest hair usually slightly longer than base of arista; fore tarsomere 5 at most slightly flattened (Fig. 10C)  Fronto-orbital plates and ocellar triangle mostly undusted and shiny (Fig. 11C); parafacial with large undusted shiny patch near base of antenna expanded ventrally (Fig. 11C); fore tarsomere 5 moderately to distinctly flattened (Fig. 10D Parafacial in lateral view no broader than width of first flagellomere (as in
Remarks. In his original descriptions, Huckett (1966) relied on subtle differences in variable characters to separate D. rivalis from D. amnicola. Following the examination of type material from these two taxa we concluded that all features listed in the original descriptions are variable (including the pollinosity of the female parafacial, a feature we found to vary from entirely dusted to mostly glossy in the type series of D. amnicola), and that D. rivalis is a synonym of D. amnicola.
Furthermore, Huckett (1966Huckett ( , 1975 uses the strength of av bristles on the apical 1/2 of F2 to separate males D. santamonicae (strong and long) from D. amnicola (weak) and junior synonym D. rivalis (weak but slightly stronger). While the male holotypes of these taxa do exhibit a difference in the strength of the F2 av, we have found this feature to be variable. The configuration of the F2 pv row and the pollinosity of the thorax and abdomen appear to be more stable characters to separate the males of D. santamonicae from those of D. amnicola (and the similar D. aldrichi) (see couplet 21 of male key). Remarks. This species was recently redescribed by Nihei and Carvalho (2004) but we are doubtful of the Nearctic records from Montana and Washington (USA). Not only do they create a highly disjunct distribution pattern for the species, they are based exclusively on female specimens which are similar to those of a number of other taxa including Drymeia minor, a species with a widespread distribution in the USA. Having examined numerous females associated with males of D. aterrima from El Salto, Mexico, we observed some differences with the description of Nihei and Carvalho (2004), mainly longer aristal hair (the longest 1.5-2.0 × as long as base) and the presence of 1 or 2 distinct av on the apical 1/3 of the mid femur (these reduced in a few specimens). ( Table S1 for GenBank accession numbers.

Drymeia cantabrigensis
Remarks. While always lacking an anteroventral bristle on the midtibia, the females of this species are sometime indistinguishable from those of D. setibasis where this bristle is either present or absent. DNA barcodes for D. cantabrigensis were very similar and in some case identical to those of other species forming a cluster (or BIN merge) of seven named species (including D. setibasis) in BOLD:AAD7664 (Fig. 25).
The males can be easily identified based on the distinctive chaetotaxy of the mid femur but since DNA barcodes do not discriminate between D. cantabrigensis and several other species including D. setibasis, the identification of females from these two species can be problematic in the Nearctic region.
Remark. DNA barcodes for material from Canada and Russia (Fig. 25) were very similar with a maximum intraspecific p-distance of 0.5%. (Huckett, 1965) Table S1 for GenBank accession numbers.

Drymeia flavinervis
Remark. DNA barcodes were available for material from Quebec and Ontario (Canada) with p-distances ranging from 0.0% to 0.16% (Fig. 25). ( Table S1 for GenBank accession numbers.

Drymeia glacialis
Remarks. We were recently informed that the holotype of Aspilia glacialis Rondani, previously considered lost (Hennig 1962b: 677;Pont 1986: 73), had been located in the Museo di Storia Naturale, Sezione di Zoologia ''La Specola'', Università di Firenze (MZUF). While we did not examine this material ourselves, the specimen has been unambiguously recognised by A.C. Pont as D. alpicola (pers. comm.) and we consider his expertise sufficient to recognise that the earlier name of D. glacialis must be given precedence over D. alpicola. Additional details about this new synonymy will be published in an upcoming work (A.C. Pont, pers. comm.).
In the Nearctic region, females of this Holarctic species can be distinguished from those of D. quadrisetosa only by the slightly darker wing base. However, this colour character appears variable in the Palaearctic region where Russian material shows a darker wing base (congruent with Nearctic females) while females from a series we examined from Austria as well as the holotype of D. glacialis (A.C. Pont pers. comm.) display a pale wing base similar to that of D. quadrisetosa, a species known only from the Nearctic region.
DNA barcodes for D. glacialis (all from Russian specimens) were very similar to those of D. quadrisetosa (specimens from Russia and Canada), forming a cluster with p-distances ranging from 0.0% to 1.72% for BOLD:AAC1021 (Fig. 25). Males of these two species can be easily distinguished based on distinctive leg chaetotaxy (see key to males) but since DNA barcodes do not discriminate between the two species, the identification of females can be problematic in the Nearctic region, as differences in wing base colour between the two taxa can sometimes be very subtle, especially for material kept in ethanol for long periods.    Table S1 for GenBank accession numbers.
Remark. DNA barcodes for material from Canada, Greenland, and Russia (Fig. 25) were similar with a maximum intraspecific p-distance of 0.76%. Etymology. The species name is a patronym in honour of Hugh C. Huckett, a major contributor to the study of Nearctic Muscidae.

Drymeia hucketti
Diagnosis. Small glossy black species with a long, strong prealar, lower margin of face projecting slightly beyond lower level of frons (Fig. 16A, 18A), and 2+3 dc. Male F2 with av row strong and regular, covering apical 2/3 to 3/4, and with pv row long and strong, at least 2 × as long as width of femur on apical 1/2, and T3 with a short but distinct ventral apical process (Fig. 16D). This species is similar to Drymeia minor  but can be distinguished from it in the female by the presence of a large undusted glossy patch on the parafacial near the base of the antenna (Fig. 18B) and, in both sexes, by the projecting face, broader ventral margin of parafacial, pubescence pattern of arista and mostly shiny abdomen. Females are also similar to those of D. amnicola (see couplet 14 of female identification key) but these taxa have very different distribution ranges.
Description. Male. Body length: 4.6-6.6 mm; wing length 4.1-5.1 mm. Head: Ground colour black; eye bare; fronto-orbital plate and parafacial silvery pruinose; face grey, gena and lower occiput grey pruinose; fronto-orbital plates touching in the middle; frons at narrowest point 2-3 × as wide as width of anterior ocellus; parafacial in lateral view with ventral margin broader than width of first flagellomere; lower margin of face projecting slightly beyond lower level of frons (Fig. 16A); gena at narrowest point 1.3 × length of first flagellomere, densely setulose and with a group of upcurved setae on anterior part of genal dilation; 11-13 frontal setae (including interstitials) reaching to anterior ocellus; antenna black; first flagellomere 1.2 × as long as wide; arista with hair much denser on basal 1/2 (longest hair as long as basal diameter of arista) and usually with sparse and very dorsal short hair on apical 1/2; palpus black; proboscis elongate with prementum much longer than palpus, mostly undusted and shiny; labella moderately developed.
Legs: Black; T1 with 1-3 pv; F2 with av row strong and regular, slightly longer than width of femur, covering apical 2/3 to 3/4 and with bristles curved on basal 1/2 (Fig. 16C), 2 or 3 preapical pd-p, with complete row of long strong pv, at least 2 × as long as width of femur on apical 1/2; T2 usually without av (one paratype with 1 short av on one side), 3-6 pd, 2-4 pv (some specimens also with 2 or 3 p); F3 with av row stronger on apical 1/3, without pv except one hair near base; T3 with 4 or 5 av, 6 or 7 ad, 4 or 5 pd, 4 or 5 short hair-like pv in middle part, ventral apical process short but distinct (Fig. 16D), apical pv distinct but no longer than 1/2 the length of apical av.
Wing: Brown, darker near base; basicosta and tegula black; costal spinules weak and costal spine reduced; calypters with membrane and edges yellow.
Terminalia: Fig. 17A, B. Female. Body length: 5.7-6.5 mm; wing length: 4.6-5.0 mm (Fig. 18A). Differs from the male as follows: Head: Frontal triangle undefined; ocellar triangle mostly glossy; frontal vitta black, deep brownish dusted; parafacial with large undusted shiny patch near base of antenna reaching up to or almost up to eye (Fig. 18B); frons at midpoint approximately 0.35 × as wide as head and approximately 1.2 × as long as wide; fronto-orbital plate narrow, approximately as wide as distance between inner margins of posterior ocelli; 3-6 medioclinate frontal setae and several weaker interstitials, two short reclinate and lateroclinate orbital setae followed by one (occasionally two) stronger proclinate orbital seta; arista as in male (Fig. 18C).
Wing: Veins yellow at least near base, membrane deep yellow near base, the remainder pale yellow to pale brown.
Remarks. The discovery of this species resulted from an exploration of all public Drymeia COI sequences found in BOLD (> 2800) which brought our attention to BOLD:ACA9214, a BIN including several well-preserved undetermined specimens of both sexes which turned out to be morphologically distinctive from any other species previously known to us. DNA barcodes for material from Canada and the United States ( Fig. 25) were available with intraspecific p-distances ranging from 0.0% of 0.35%. Remarks. The female holotype, with its unique combination of a broad frons, mid femur with weak av setae on apical 1/2, mid tibia with a single strong av, hind tibia with a strong apical pv, and wing membrane pale brown, is distinctive from any other specimen we have examined.  Table S1 for GenBank accession numbers.

Drymeia latifrons (Malloch, 1918)
Remarks. Based on the distribution of examined specimens (including the holotype of D. minor) as well as on the distribution of all public sequences for BOLD:ADZ5293 (D. minor) and BOLD:ACA9214 (D. hucketti sp. nov.), we suspect that previously published Canadian records of D. minor may actually belong to D. hucketti sp. nov. but additional data will be necessary for confirmation. Only specimens from Colorado (USA) were available for DNA barcoding (Fig. 25)   Baker Lake, Cambridge bay, Char river nr Ranking Inlet, Chesterfield, Landing Lake (7.5 km NW of Rankin inlet). Palaearctic: Russia: Chukotka AO: Wrangel Island (BUIC, CNC, SZMN).
Remarks. Huckett (1975: 103) mentions an undescribed Eupogonomyia species from California (female only) that would be distinguished from D. neoborealis based on the complete absence of a prealar bristle (the prealar is short but often visible in D. neoborealis) as well as projecting oral margins. We have found these features to be variable in the material we have examined (including specimens with DNA barcodes) and conclude that the unnamed species mentioned by Huckett (1975) falls within the range of known variations for D. neoborealis. While the COI sequence from the only Russian specimen in our data set showed a minimum intraspecific p-distance of 1.37% with the Canadian specimens, all DNA barcodes for D. neoborealis clustered together in BIN BOLD:ACA8935 (Fig. 25) with a maximum intraspecific p-distance of 1.52%. Etymology. The species name is a patronym in honour of Adrian C. Pont (UK), an exceptional dipterist and mentor to both co-authors.
Diagnosis. Small dark species with strong prealar, 2+3 dc and strong costal spine. This species is similar to Drymeia aterrima (Wulp, 1896), especially in the dark male calypter, but can be distinguished from it in the male by a broad frontal vitta (Fig. 13B) and a distinct ventral apical process on T3 (Fig. 13D), and in the female by the presence of a flattened fore tarsomere 5 (Fig. 15C).
Description. Male. Body length: 4.1-5.7 mm; wing length: 3.8-5.2 mm. Head: Ground colour black; eye bare; fronto-orbital plate and parafacial dark brown pruinose; face black, gena and lower occiput dark brown pruinose; frons at narrowest point approximately 1.5 × width of ocellar triangle with black frontal vitta exposed (Fig. 13B); parafacial in lateral view equal to or slightly wider than width of first flagellomere along most of its length (Fig. 13A); lower margin of the face projecting slightly beyond lower level of frons in lateral view; gena at narrowest point as high as length of first flagellomere, densely setose and without a group of upcurved setae on anterior part of genal dilation; 9-12 frontal setae (including interstitials) reaching to anterior ocellus; antenna black; first flagellomere 1.2 × as long as wide; arista swollen near base and pubescent, with longest hair as long as basal diameter of arista; palpus black; proboscis long and narrow with prementum approximately 2 × as long as palpus, undusted and glossy; labella small.
Legs: Black; T1 with 2 pv on apical 1/2; F2 straight, with matching rows of long strong av and pv on apical 2/3, these much longer than width of femur (Fig. 13A, C), a complete row of short ad, and 2 or 3 preapical pd-p; T2 with 1 long and 2 shorter ad on apical 1/2, 4 pd and 1 pv; F3 with a complete row of ad, a row of av, longer on apical 1/2, a row of p on basal 2/3 and a row of long strong pv on apical 1/3; T3 with 2 av, 4 ad, 3 pd and with 4 short delicate pv on apical 1/2, ventral apical process short but distinct (Fig.  13D), apical pv absent or reduced, no longer than 1/2 the length of apical av when visible.
Wing: Brown, darker at base with dark brown veins; basicosta and tegula black; costal spinules strong, with costal spine 2 × as long as costal spinules; calypters with membrane and edges dark brown.
Terminalia: Fig. 14A, B. Female. Body length: 4.5-6.5 mm; wing length: 4.0-5.5 mm. Differs from the male as follows: Head: Frontal triangle undefined; frontal vitta black; parafacial mostly dusted, with a small narrow shiny patch near base of antenna (Fig. 15A, B); frons at midpoint approximately 0.4 × as wide as head and approximately 0.9 × as long as wide; frontoorbital plate narrow, approximately as wide as distance between inner margins of posterior ocelli; three or four medioclinate frontal setae and several weaker interstitials, three orbital setae, the upper two reclinate and lateroclinate, the lower one proclinate.
Wing: Light brown, darker near base; with membrane and edges dark yellow. Abdomen: as in male. Distribution. Neotropical: Mexico (Mexico). DNA Barcode. None available.

Drymeia pribilofensis (Malloch, 1919)
Figs 4A, 5A, 8B mid femur but the females can be difficult to distinguish from those of D. setibasis and D. cantabrigensis (see key to females). All barcoded specimens of D. pribilofensis from our dataset clustered closest to one another on the neighbour-joining tree (Fig. 25) with a maximum intraspecific p-distance of 0.15% but the distance to D. cristata, one of the seven species found in the BOLD:AAD7664 BIN merge was very low (min p-distance = 0.61%). We would therefore not recommend using COI to discriminate specimens of D. pribilofensis from those of other species in this BIN.
Remarks. See comments about DNA barcodes under D. glacialis.   Table S1 for GenBank accession numbers.
Remark. DNA barcodes for D. segnis material from Canada and Greenland were very similar to several of the seven species found in the BOLD:AAD7664 BIN merge (Fig. 25) and in some cases, identical to those of D. setibasis. However, both sexes of D. segnis can easily be distinguished from all other species in this BIN by the presence of a haired anepimeron. We would therefore not recommend using COI to discriminate specimens of D. segnis from those of other species in this BIN.
Remark. All three specimens of D. similis in our DNA barcoding data set were from Saskatchewan (Canada) with p-distances ranging from 0.0% to 0.16% (Fig. 25).  Table S1 for GenBank accession numbers.
Remarks. The comparison of type material from D. longiseta Sorokina and Pont from Russia with a series of high elevation specimens from Colorado, USA, housed in the CNC and matching the original description of D. spinitarsis has led us to recognise D. longiseta as a junior synonym of D. spinitarsis. This change effectively expands the distribution of D. spinitarsis to the Palaearctic region and while DNA barcodes are currently only available for Russian specimens, D. spinitarsis is such a large, distinctive species (strong spines on mid tarsomere 1, distinctive chaetotaxy of fore and mid coxae in the male, high elevation distribution) that we are quite confident in this new synonymy.   Etymology. The species name is a patronym in honour of famous dipterologist John R. Vokeroth (Canada), who collected the holotype and several paratypes and designated these specimens as a "new species near D. neoborealis".

Drymeia vockerothi
Diagnosis. Large glossy black species with short proboscis, very narrow parafacial in lateral view, arista almost bare, prealar weak or absent, 2+3 or 4 dc and well developed presutural acr. Males with narrow frons (Fig. 19B), F2 with at most a few long av near base and with ventral apical process of T3 short but distinct (Fig. 19C). This species is similar to Drymeia neoborealis (Snyder, 1949) but can be distinguished from it in both sexes by the presence of strong presutural acr and a narrow parafacial in lateral view. Male. Body length: 7.2-9.5 mm; wing length: 6.0-6.7 mm (Fig. 19A). Head: Ground colour black; eye bare; fronto-orbital plate and parafacial yellowish silver pruinose; face grey, gena and lower occiput grey pruinose; frons at narrowest point 1.5-2.5 × width of anterior ocellus with fronto-orbital plates touching or nearly touching (Fig. 19B); parafacial very narrow in lateral view, < 1/2 × width of first flagellomere along full length; lower margin of the face equal to or slightly behind lower level of profrons in lateral view; gena at narrowest point 0.5 × length of first flagellomere, densely setulose and with a group of upcurved setae on anterior part of genal dilation; 14-16 frontal setae (including interstitials) reaching almost to anterior ocellus; antenna black; first flagellomere 1.3 × as long as wide; arista swollen near base and almost bare, with longest hair much shorter than basal diameter of arista; palpus black; proboscis broad with prementum shorter than palpus, mostly undusted and glossy; labella large and fleshy.
Wing: Light brownish, yellow at base; basicosta and tegula black; costal spinules weak and costal spine reduced; calypters with membrane and edges yellow.
Abdomen: Conical, ground colour black; densely grey dusted with narrow black median vittae on tergites II-V; sternite I bare; sternite V as in Fig. 20C.
Terminalia: Fig. 20A, B. Female. Body length: 7.5-9.0 mm; wing length: 6.2-6.8 mm (Fig. 21A). Differs from the male as follows: Head: Ground colour black with grey dusting; frontal vitta brown with grey dusting; frontal triangle indistinct, ocellar triangle undusted and shiny; parafacial entirely dusted, without shiny patch near base of antenna; frons at midpoint 0.36 × as wide as head and approximately 0.85 × as long as wide; fronto-orbital plate wide, at midpoint 0.4 × as wide as frontal vitta; seven medioclinate frontal setae (including several interstitials), three orbital setae, the upper two reclinate and lateroclinate, the lower one proclinate.
Thorax: Ground colour as in male but with light grey dusting (Fig. 21B); acr as in male but often in irregular row.
Wing: As in male. Etymology. The species name is a patronym in honour of the late Canadian dipterologist D. Monty Wood and his wife Grace Wood (Canada), who collected the holotype.
Diagnosis. Small grey species with short proboscis, narrow frons and narrow parafacial in lateral view (Fig. 22A), arista almost bare, prealar absent in male and weak in female, 2+4 dc and well developed presutural acr. Male with patch of four or five very long, strong apical bristles on fore coxa and with a row of ad on apical 2/3 of T2. This species is similar to Drymeia neoborealis and Drymeia vockerothi sp. nov. but can be distinguished from them in both sexes by the smaller size, in the male by the bristles of the fore coxa and T2, and in the female by the combination of strong presutual acr, completely dusted prementum, and parafacial in lateral view nearly as wide as width of first flagellomere.
Description. Male. Body length: 5.5 mm wing length: 4.5 mm. Head: Ground colour black; eye bare; ocellar triangle, fronto-orbital plate and parafacial silvery pruinose; face grey, gena and lower occiput light grey pruinose; frontoorbital plates touching in the middle; frons at narrowest point as wide as width of anterior ocellus; parafacial very narrow in lateral view, < 1/2 width of first flagellomere along full length (Fig. 22A); lower margin of the face equal to or slightly behind lower level of profrons in lateral view; gena at narrowest point 0.5 × length of first flagellomere, densely setulose and with a group of upcurved setae on anterior part of genal dilation; 12 frontal setae (including interstitials) reaching almost to anterior ocellus; antenna black; first flagellomere 1.3 × as long as wide; arista swollen near base and almost bare, with longest hair much shorter than basal diameter of arista; palpus black; proboscis short, with prementum shorter than palpus and heavily dusted; labella large and fleshy.
Legs: Black; fore coxa with patch of four or five very long, strong apical bristles, at least as long as length of coxa; T1 with 1 or 2 pv; F2 with a row of short av on apical 1/3, 5 or 6 a on basal 1/2, these longer than width of femur, 3 preapical pd-p, and dense fine pv over most of the surface; T2 without av, with a row of ad on apical 2/3, these longer near apex, 4 pd and 2 pv; F3 with av row stronger on apical 1/3, without pv except for a few hairs near base; T3 with 3 av, 3 long ad, a row each of short erect uneven a and ad, 2 long and 2 or 3 short pd, ventral apical process very short but distinct, apical pv absent (Fig. 22B).
Wing: Mostly dark brown; basicosta and tegula black; costa with short weak spinules and costal spine reduced; calypters with membrane and edges deep yellow.
Abdomen: Conical, ground colour black; densely grey dusted with a black median vitta on tergites II-V; sternite I bare; sternite V as in Fig. 23B.
Female. Body length: 4.9-5.5 mm wing length: 4.3-4.7 mm (Fig. 24A). Differs from the male as follows: Head: Ground colour grey with grey dusting; frontal vitta black with brownish dust; frontal triangle indistinct, ocellar triangle covered with heavy grey or brown dust; parafacial completely dusted but with large area near base of antenna appearing velvety brown in anterior and lateral views (Fig. 24B, C); frons at midpoint approximately 0.4 × as wide as head and approximately 0.75 × as long as wide; fronto-orbital plate narrow, at midpoint 0.16-0.2 × as wide as frontal vitta; gena approximately as high as length of first flagellomere; five or six medioclinate frontal setae (including interstitials), three orbital setae, the upper two reclinate and lateroclinate, the lower one proclinate.
Thorax: Scutum with whitish grey dust and, when viewed from the front, with 3 brown vittae with diffuse margins (Fig. 24B); postpronotum, notopleuron, postalar lobe and pleuron grey dusted; acr as in male but with slightly uneven rows; prealar distinct but weak in all females from the type series, always much shorter than 2 nd notopleural.
Legs (chaetotaxy described in full): Fore coxa without patch of long strong bristles; T1 with 1 or 2 pv; F2 with a row of short av, no longer than width of femur and a row of delicate pv-v, no longer than diameter of femur; T2 usually with a row of 5-7 irregular pd, (these sometimes reduced) and 1 pv; F3 with av row stronger on apical 1/3, without pv except for a few hairs near base, T3 with 2-3 av, 5-6 short irregular ad and 5-6 pd, apical pv usually absent but if visible, then no longer than 1/2 the length of apical av.
Wing: Clear with slight pale brown to pale yellow tinge, veins pale brown; calypters whitish.
Abdomen: Densely whitish grey dusted with slight brownish tinge along posterior margins of tergites II-IV.
Remarks. See notes on problematic taxa.

Notes on problematic North American taxa
Huckett (1965a: 297) provided a key (but no descriptions) to females for three unnamed species of northern Drymeia with a short prealar, short proboscis and proclinate anterior orbital setae (as Trichopticoides sp. A, sp. B, sp. C). While we could not trace the specimens listed by the author, females of Drymeia woodorum sp. nov. (Fig. 24A-C) would key out to Trichopticoides species B (recorded from Alaska and the Yukon Territory) and one paratype from Whitehorse, Yukon Territory, bears a yellow label written "Trichopticoides'' in what appears to be Huckett's handwriting. The taxa referred to as Drymeia sp.1 (BOLD:ACA9236) and D. sp. 2 (BOLD:ACZ1539) in the present work (Fig. 25) would also have keyed out to "Trichopticoides" in Huckett (1965a); specimens of D. sp. 1 were too damaged for additional assessment but those of D. sp. 2 (two females from Kluane, northern British Columbia) would key out to "Trichopticoides" sp. C, recorded from nearby Alaska. Unfortunately, as we have not located the specimens listed in Huckett (1965a) we remain unable to assess if one or more of his unnamed species are a true match for some of our material. Drymeia vockerothi sp. nov. is another species with a short prealar, short proboscis and proclinate orbital setae but it is much larger than the other taxa discussed above and does not fit the features listed in the key or the distribution of any Trichopticoides spp. mentioned in Huckett (1965a).

Morphology
Three new synonymies are established here (the recognition of D. longiseta Sorokina & Pont, 2015 as a junior synonym of D. spinitarsis (Aldrich, 1918), of D. rivalis (Huckett, 1966) as a junior synonym of D. amnicola (Huckett, 1966), and of Drymeia alpicola (Rondani, 1871) as a junior synonym of Drymeia glacialis (Rondani, 1866)). These synonymies are based exclusively on morphology and details are provided in the checklist section of this work.
Most species of Drymeia show distinctive and discrete morphological differences between species (at least in the males). However, the four species described in Huckett (1966) from California turned out to be especially challenging. Specimens of D. profrontalis (Huckett, 1966) can usually be distinguished from other similar species but we found that even following the synonymy of D. rivalis with D. amnicola, some specimens (mostly females but even some males) could not be unambiguously assigned to either D. amnicola or D. santamonicae (Huckett, 1966). While the type series for these taxa are generally long, we could not find recently collected material; we therefore plan to prioritise western California in future collecting efforts.

DNA barcoding
A total of 175 sequences were included in the DNA barcode data set (Suppl. material 1: Table S1). The majority (150) were from specimens assigned to named species while the remaining 25 could not be matched to any known taxa. This material came from Canada, Russia, Greenland, the United States, and Finland; COI sequences were assigned to 24 BINs in BOLD (Fig. 25).
Eighteen named species and four unknown taxa matched a single BIN. These numbers were compiled after we re-examined material (all paratypes) from BOLD:ACT1697 (males and females assigned to D. grandis Sorokina & Pont, 2015) and BOLD:ACT1698 (a mixture of males D. cilitarsis Sorokina & Pont, 2015, and females originally assigned to D. grandis). The female of D. cilitarsis is currently undescribed and the re-examination of specimens from these two BINs led us to the conclusion that the paratype series of D. grandis is mixed, containing properly identified females as well as some belonging to D. cilitarsis (both taxa have similar distributions). The description of the female of D. cilitarsis, a Palaearctic species, is outside the scope of the present work and will be completed in a separate contribution.
Following the examination of material from the four BINs containing specimens we could not assign to known species based on morphology we described one new species, Drymeia hucketti sp. nov. (BOLD:ACA9214). Due to the very poor condition of available specimens (BOLD:ACA9236), combined with a lack of males for BOLD:ACZ1539 we chose to leave these two clusters unassigned as Drymeia sp. 1 and Drymeia sp. 2, respectively (Fig. 25), until more material becomes available. Both were relatively small (< 6.5 mm), had short mouthparts, non-projecting faces, prealar absent, and 2+4 dc; in males of sp. 1, the apical ventral process on T3 was very short (see notes on problematic North American taxa for additional details).
The four specimens (two from each sex) available for BOLD:ACY7523 (Drymeia sp. nr. aldrichi), were most similar to D. aldrichi  but differed slightly from material examined (holotype male + two females from same series) in the following features: females with very short av and pv near the base of the mid femur and with tarsomere 5 not flattened, males with projection of frons slightly more pronounced and with slightly longer pv on the apical 1/2 of the mid femur. Since no DNA barcodes were available for material matching the morphology of specimens of D. aldrichi examined here, we chose to leave this cluster unresolved as we are unable at present time to determine if the material from BOLD:ACY7523 belongs to a new distinct species or if D. aldrichi is more variable morphologically than currently understood.
While the congruence between BIN assignment and morphology was good for 71% (22/31) of species studied here (including the three unnamed taxa), a significant proportion of species (9/31) presented no or very low interspecific distance to other morphologically distinct taxa. These taxa were grouped under two BIN "merges" including two (BOLD:AAC1021) and seven (BOLD:AAD7664) species (Fig. 25). In both cases, obvious and discrete morphological differences (at least in the males) exist between all species from these mixed clusters (see identification keys; remarks under D. cantabrigensis, D. glacialis, D. pribilofensis and D. segnis;and Sorokina and Pont (2015)).
Low interspecific distance between sequences from Canadian specimens of D. pribilofensis and D. segnis had previously been flagged by Renaud et al. (2012) but our wider taxon sampling uncovered a more complex issue, now with seven species clustering together in BOLD:AAD7664 with p-distances ranging from 0.0% to 1.83%. Of these taxa, only D. cantabrigensis is restricted to the Nearctic region. Drymeia pribilofensis, D. setibasis, and D. segnis are Holarctic and the remaining species (D. cristata Sorokina & Pont, 2015, D. grisea Sorokina & Pont, 2015and D. puchokana Sorokina & Pont, 2015 are known only from the eastern Palaearctic region. Drymeia pribilofensis was the only species from BOLD:AAD7664 where all sequences in the data set shared unique nucleotides at several positions in the alignment (50, 58, 300, 475) but these differences were not enough to provide an unambiguous species-level match in BOLD and classify these sequences in a separate group under the BIN system (Ratnasingham and Hebert 2013).
Male specimens were included in our dataset for all species involved in the two BIN merges except D. puchokana; we therefore exclude misidentification as a possible explanation for the mismatches between BIN assignment and morphology reported here. Low genetic divergence in COI sequences between closely related species has been reported in the literature for different groups of animals (e.g., Whitworth et al. 2007;Gibbs 2018;Tizard et al. 2019) and can result from a number of non-exclusive causes such as polymorphism, hybridisation and/or incomplete lineage sorting (Funk and Omland 2003). Additional data based on other genetic markers and/ or reproductive compatibility experiments will therefore be required before modifying the current taxonomy for members of the two BIN merges reported here in Drymeia.

Conclusions
We found both advantages and limitations to using DNA barcodes and BIN assignments for the identification of Drymeia species. They allowed us to resolve the issue of male/female associations in the Palaearctic species D. grandis and D. cilitarsis, to discover one new species, D. hucketti sp. nov., and to focus our attention on three potentially new taxa (Drymeia sp. 1, D. sp. 2, D. sp. nr. aldrichi). BINs were also a good match for 22 of the 31 species included in our data set, regardless of geographical distance. However, the nine remaining species could not be distinguished from at least one other species; we therefore caution against using BIN assignments based on COI DNA barcoding as the only determination tool for Drymeia material without prior knowledge of its limitations for certain species groups. This is especially important for the Nearctic fauna where only 65% (11/17) of the species represented in our DNA barcode library (including the three unnamed taxa) have a match between BIN and morphology while the remaining six show no or very low sequence divergence with one or more morphologically distinct species.
Considering the taxonomic changes presented here, the North American fauna of Drymeia now includes 24 species; 22 are found in the Nearctic region, D. ponti sp. nov. is recorded only from Neotropical Mexico, and D. aterrima has been recorded from both the Nearctic and Neotropical regions. Furthermore, the synonymy of D. longiseta with D. spinitarsis now brings the number of Holarctic species to ten. Whereas the present study provides the first comprehensive taxonomic treatment of Drymeia for North America, ten of the 24 named species of Drymeia are still missing DNA barcodes and some additional work remains. The most important gap in knowledge concerns the fauna of the southwestern United States and Mexico and the collection of additional specimens from those areas are likely to result in further changes to the taxonomy of the genus. Financial support was provided by an NSERC Discovery grant (Canada) to J. Savage, Bishop's University, a CanaColl grant (CanaColl foundation) to V.S. Sorokina and the work of V.S. Sorokina was carried out within the framework of the State Research Program FWSG-2021-0004. The authors have declared that no competing interests exist.