Integrative taxonomy of the genus Dyscolus (Coleoptera, Carabidae, Platynini) in Ecuadorian Andes

The genus Dyscolus Dejean, 1831 is a highly speciose taxon of neotropical Carabidae and the major component of high-altitude ground beetle communities in the tropical Andes. The aim of this study is threefold: (i) refi ne the taxonomic position of the equatorial members of Dyscolus using molecular data, (ii) provide a delimitation of the species found in Ecuador in páramo and montane forest environments based on a robust combination of molecular and morphological data, (iii) describe the new species and take the nomenclatural decisions made necessary by the results of this study. The seclusion of Dyscolus from more basal platynine clades including Platynus, Batenus and Glyptolenus, is supported by a phylogenetic analysis of the COI marker. Twenty-fi ve new species of Dyscolus, most of them microendemic, are described and illustrated: D. aquator Moret sp. nov. (Tandayapa, Pichincha), D. arauzae Moret sp. nov. (Mt Cayambe, Pichincha), D. arborarius Moret sp. nov. (Parque Nacional Yacuri, Loja), D. barragani Moret sp. nov. (Mt Ayapungu, Chimborazo), D. crespoae Moret sp. nov. (Parque Nacional Podocarpus, Zamora-Chinchipe), D. danglesi Moret sp. nov. (Parque Nacional Yacuri, Loja), D. donosoi Moret sp. nov. (Parque Nacional Podocarpus, Zamora-Chinchipe), D. eleonorae Moret sp. nov. (Cotopaxi and Pichincha provinces), D. famelicus Moret sp. nov. (Papallacta, Napo), D. giselae Moret sp. nov. (Reserva Otonga, Cotopaxi), D. globoculus Moret sp. nov. (Parque Nacional Podocarpus, Zamora-Chinchipe), D. gobbii Moret sp. nov. (Guamaní and Mt Antisana, Pichincha), D. incommunis Moret sp. nov. (Tandayapa, Pichincha), D. marini Moret sp. nov. (Parque Nacional Podocarpus, Loja), D. piscator Moret sp. nov. (Guamaní, Napo), D. placitus Moret sp. nov. (Guamaní, Napo), D. ravidus Moret sp. nov. (Parque Nacional Podocarpus, Loja), D. rivinus Moret sp. nov. (Reserva Otonga, Cotopaxi), D. rugitarsis Moret sp. nov. (Parque Nacional Yacuri, Loja), D. ruizi Moret sp. nov. (Parque Nacional Podocarpus, Loja and Zamora), D. salazarae Moret sp. nov. (Parque Nacional Podocarpus, Loja), D. silvestris Moret sp. nov. (Papallacta, Napo), D. sulcipedis Moret sp. nov. (Parque Nacional Yacuri, Loja), D. verecundior Moret sp. nov. (Mt Illiniza and Mt Corazón, Pichincha) and D. verecundissimus Moret sp. nov. (Mt Chimborazo, Chimborazo). Dyscolus palatus Moret, 1998 is newly synonymized with D. denigratus (Bates, 1891). We demonstrate the subgenus Hydrodyscolus Moret, 1996 to be polyphyletic and therefore consider it a junior synonym of Dyscolus Dejean, 1831. European Journal of Taxonomy 646: 1–55 (2020) 2


Introduction
is the most diverse and speciose genus of the ground beetle tribe Platynini in tropical America. Its taxonomic position is still a matter of discussion. From Maximilien de Chaudoir′s world-wide revision (1859,1878) to the mid-twentieth century, most neotropical platynines were placed in a huge, pantropical genus Colpodes McLeay, 1825. Moret (1989: 135) and Liebherr (1998) showed that the genus-level name Colpodes is restricted to a small clade of highly distinctive Javanese species and cannot be used in the broad sense of Chaudoir (1878). Whitehead (1973), followed by Liebherr (1986Liebherr ( , 1992, transferred all the Neotropical Colpodes to the genus Platynus Bonelli, 1810, until then restricted to a moderately diverse lineage of holarctic platynines. However, Moret (1989Moret ( , 1999 argued that characters of submental setation, tarsal morphology and female genitalia made it impossible to include the Neotropical representatives of the former Colpodes complex in the genus Platynus. He removed Dyscolus from synonymy with Colpodes to designate most of the South American platynines previously placed either in Colpodes or in Platynus (Moret 1989). Additionally, fi ve lineages closely related to Dyscolus were isolated in separate genera by Moret or Perrault: Scaphiodactylus Chaudoir, 1838(Moret 1996, Stenocnemion Moret, 1989 (Moret 1991), Dyscolidion Moret, 1990, Andinocolpodes Perrault, 1990 and Glyptolenopsis Perrault, 1991(Perrault 1991. At fi rst, these taxonomic changes did not receive wide agreement. All these taxa were treated by Liebherr as species groups within Platynus (Liebherr 1992), whereas Dyscolidion and Glyptolenopsis were considered subgenera of Platynus in Lorenz (2005), and Dyscolus was treated as a subgenus of Platynus in Bousquet (2012b).
A later decisive contribution  showed that the derived condition of male monorchy (i.e., the presence of only one testis) provides evidence for the monophyly of the Colpodes-Dyscolus complex, whereas Platynus shares with other platynine genera a plesiomorphic diorchid condition. Although the taxonomic consequences of this important fi nding were not explicitly drawn by the authors of this study, the genus name Dyscolus is increasingly used for neotropical platynines of the Colpodes-Dyscolus complex in checklists (Martínez & Ball 2003;Lorenz 2019) and ecological surveys (Maveety & Browne 2014). However, the name Platynus has also been used in similar contexts (Camero 2003;Erwin et al. 2015).
This taxonomic uncertainty is all the more problematic since Dyscolus is a hyperdiverse taxon with 445 accepted described species (Lorenz 2019) and several hundreds of undescribed species. Widespread in subtropical and tropical America from Southern United States to northern Argentina, it is the major component of high-altitude ground beetle communities in the equatorial Andes, both in the upper belt of the montane forest above 2000 m a.s.l. (Maveety & Browne 2014) and in the tropical alpine páramo ecosystem above 3500 m a.s.l. (Moret 2005;Gobbi et al. 2018).
The aim of the present study is threefold: (i) refi ne the taxonomic position of the genus Dyscolus within the tribe Platynini using molecular data; (ii) provide species delimitation for the Ecuadorian highaltitude members of the genus, based on a robust combination of molecular and morphological data; (iii) describe the new species and take the nomenclatural decisions based on the results of the present study. Among this material, 218 specimens identifi ed as Dyscolus were selected for molecular analysis. They were euthanized with ethyl acetate and then immediately stored in 96% ethanol. After removing one leg for DNA extraction, voucher specimens of the new species were dried, pinned on mounting boards and deposited in public entomological collections (Table 1; further details in the taxonomic section). The vouchers of already described species are kept in 96% ethanol at Toulouse University. Prior to molecular analyses, specimens were identifi ed to species level following Moret (2005).

Amplifi cation and sequencing
Total DNA was extracted from one leg. After discarding the ethanol used for fi xing the specimens, samples were dried 15 minutes at 70°C to ensure complete removal of the ethanol. We used the Macherey Nagel XS kit following manufacturer′s protocol. Final elution was performed in 30 μL followed by a concentration of 5 minutes at 90°C (fi nal volume 15-20 μL). Purifi ed genomic DNA was used as a template for PCR amplifi cation. The standard 658 base pairs barcode of the mitochondrial proteinencoding gene cytochrome c oxidase subunit I [COI hereafter] was amplifi ed using the primer pair LCO1490 (5′ GGT CAA CAA ATC ATA AAG ATA TTG G 3′) / HCO2198 (5′ TAA ACT TCA GGG TGA CCA AAA AAT CA 3′) (Folmer et al. 1994) using the FastStart PCR Master (Roche, Mannheim, Germany). The total PCR volume was 25 μL and consisted of 2.5 μL of 10X reaction buffer, 2 μL of 10 mM dNTPs, 1.5 μL of 25 mM MgCl 2 , 1 μL of each primer, 8.2 μL of H2O, 0.2 μL of Taq Polymerase and 10 μL of template DNA. The PCR cycles were as follow 94°C for 4 min, 35 cycles at 94°C for 30 s, 55°C for 30 s and 72°C for 1 min; and then a fi nal extension at 72°C for 7 min. The PCR products were verifi ed by capillary migration on a QIAxcel Advanced system (Qiagen) and DNA quantitation was performed using PicoGreen (Turner BioSystems) on an Infi nite F400 plate reader (Tecan). The purifi ed PCR products were sequenced directly with the same primer pairs as used for amplifi cation. The sequencing reactions were performed using BigDye Terminator ver. 3.1 (Applied Biosystems). The BigDye-labelled PCR products were purifi ed using Sephadex G-50 medium (Sigma). The sequence reaction products were then analyzed using an ABI Prism 3730xl Genetic Analyzer (Applied Biosystems) by Genoscreen (Lille, France). Forward and reverse sequences were edited and assembled using Geneious 9 (Kearse et al. 2012). The 212 obtained sequences (sequencing failed for six specimens) were uploaded to the Barcode of Life Data Systems (Ratnasingham & Hebert 2007), and can be found in the BOLD projects "DS-GLYPTO" and "DS-DYSGLYP".
The DS-GLYPTO dataset thus includes 93 specimens, 79 of which are members of the Dyscolus complex. After initial alignment with Muscle (Edgar 2004), a Maximum Likelihood phylogenetic tree was reconstructed using raxml-ng (Kozlov et al. 2019) based on 100 random parsimony starting trees, applying a GTR+G model and performing an automatic bootstrapping procedure (autoMRE option) using the Transfert Bootstrap Expectation support metric (Lemoine et al. 2018).

Species delimitation
In the following step of the analysis, we only included representatives of the Dyscolus genus but used all the available sequences, 212, thus including several sequences per nominal species. In addition to our own sequences, we also included nine sequences provided by Carlos Ruiz Carreira, from carabid beetles collected during his project "Estudio de la riqueza de especies y estructura fi logenética en comunidades de carábidos a lo largo de un gradiente altitudinal en el Parque Nacional Podocarpus" (Universidad Técnica Particular de Loja, 2013-2014:  Among the algorithms available for molecular species delimitation, we chose to use a distance-based method that does not rely on ad hoc similarity thresholds and does not require parameters that are diffi cult to select a priori. We used the REfi n Single Linkage clustering approach (RESL; Ratnasingham & Hebert 2013) as implemented in the Barcoding of Life Data System [BOLD] to defi ne Barcode Index Numbers (BINs) based on our COI dataset. The RESL algorithm has the advantage of using a two-step procedure: an initial clustering at a 2.2% divergence threshold followed by a refi nement step using Markov clustering. In addition, it uses all of the sequences present in the BOLD database for clustering, allowing for a direct comparison of our dataset with sequences produced from other barcoding projects.

Morphological examination and measurements
Specimens were examined and measured using a Zeiss Stemi 508 trinocular stereo microscope. Photographs were taken with a Zeiss AxioCam 105 microscope camera. Z-series of photographs were stacked using Helicon Focus Software (Helicon Soft Ltd), then enhanced with Photoshop © . Body size was measured from the tip of mandibles in closed position to the apex of elytra. The width of the pronotum and elytra was measured at their widest points. The length of the pronotum was measured along the median line. The length of the elytra was measured from the basal bead to the apex of the longer elytron.

Taxonomic treatment
Following publication procedures aimed at speeding up the description of species in hyperdiverse genera and providing a partial solution to the taxonomic impediment (Butcher et al. 2012;Riedel et al. 2013), we combined concise diagnostic descriptions with DNA-barcoding and high-resolution photographs of type specimens. Morphological descriptions are therefore limited to diagnostic characters allowing a sure separation from any other currently known taxon.
The studied material is housed in the following museums and collections:

Systematic position of Dyscolus
Even though we acknowledge that a molecular study restricted to COI data might present limited phylogenetic information, our analysis based on a broad taxon sampling (dataset DS-GLYPTO, Fig. 2) confi rms morphology-based statement that the genus Dyscolus is only distantly related to Platynus. In particular, our results are congruent with those based on testicular morphology . All the diorchid taxa of our dataset -i.e., those with two testes, which is the plesiomorphic state of this character  (Liebherr & Will 1996;Bousquet 2012aBousquet , 2012bLorenz 2019): P. brunneomarginatus, P. ovipennis and P. tenuicollis. Together with these three species whose monorchid condition has been experimentally demonstrated , other North American species currently placed in Platynus (Bousquet 2012b) are found far from the Platynus-Batenus clade and close to Dyscolus: P. cohni Liebherr & Will, 1996, P. falli (Darlington, 1936 [registered as Dyscolus in Lorenz 2019], P. lyratus (Chaudoir, 1879) and P. megalops (Bates, 1882) [registered as Dyscolus in Lorenz 2019].
Platynus in its currently accepted defi nition thus appears to be polyphyletic. As already suggested (Moret 1989(Moret , 1999, Platynus should be restricted to species with only one pair of setae on the submentum and possessing both apical and subapical pairs of setae on the dorsolateral margin of the fourth metatarsomere, as observed in the type species, Platynus assimilis (Paykull, 1790), and in a few related North American species. The species exhibiting a quadrisetose submentum and with only one pair of setae on the dorsolateral margin of the fourth metatarsomere, which is the case of P. brunneomarginatus, P. cohni, P. falli, P. lyratus, P. megalops, P. ovipennis and P. tenuicollis in our dataset, should be removed from Platynus. A more comprehensive analysis is needed to fi nd out if all of these species can fi t into a broadly defi ned genus Dyscolus, or if some of them should be placed in another taxon (an option could be Scaphiodactylus, as defi ned in Moret 1999). We therefore decided not to modify the binomials retrieved from Lorenz (2019) in our phylogenetic tree (Fig. 2).
With regard to the other platynine genera included in our analysis, it should be noted that Glyptolenus and Glyptolenoides are found in different branches, which suggests that the shared derived character of externally canaliculate foretibiae is likely to be a homoplasy. Conversely, Incagonum appears to be closely related to Glyptolenus, rather than to the Rhadine-Tanystoma lineage as previously hypothesized (Liebherr 1994). The melanistic base of the male genitalia, a character shared by Incagonum and Glyptolenus, is likely to give a morphological support to this relationship.

Subgeneric divisions of Dyscolus
Our results show that representatives of the supraspecifi c taxa Dyscolidion (GMADN010-16) and Glyptolenopsis (NEONU670-11) are nested within the Dyscolus clade. These positions are compatible with their current treatment as subgenera of Dyscolus (Lorenz 2019). More research is needed to determine whether they should be maintained as subgenera or downgraded to species groups. The supraspecifi c taxon Stenocnemion could not be included in this study, as no specimens preserved in pure ethanol were available. However, the monorchid condition of at least one Central American Stenocnemion species  suggests that this lineage, defi ned by two highly distinctive synapomorphies (parascutellar area of elytron bisetose and ventral face of female gonocoxite plurisetose), can be provisionally treated as a subgenus of Dyscolus (Moret 2019).
Ten species of the studied dataset fi t the defi nition of the subgenus Hydrodyscolus: head large with long and acute mandibles, fourth metatarsomere lacking dorsolateral setae but with a dense ventral pubescence, riparian way of life. These species turned out to be scattered in fi ve different clades: i/ D. irriguus Moret, 2005 and D. lutarius Moret, 2005; ii/ D. gobbii ; iii/ D. arauzae ; iv/ D. asphaltinus (Chaudoir, 1878), D. aquator Moret sp. nov., D. giselae Moret sp. nov. and D. incommunis ; v/ D. gaujoni (Perrault, 1989). Even though COI barcoding does not provide suffi cient support to resolve the relationships between species groups within the genus Dyscolus, the mentioned combination of characters appears to be homoplasic. As the loss of the dorsolateral setae is associated in most species with an increasing density and length of the ventral pubescence, it can be assumed that this modifi cation of the tarsal morphology is an adaptive character, related to the capacity of riparian species to move on or in the water. Similarly, the lengthening of the mandibles could be related to the habit of feeding on aquatic arthropods or crustaceae (personal observation in Ecuador, Ayapungu, 4180 m a.s.l.; 24 Mar. 2015: an adult of D. irriguus plunging his head into the water to capture an amphipod of the genus Hyalella). Accordingly, the subgenus Hydrodyscolus appears to be polyphyletic and we therefore propose to treat it as a junior synonym of Dyscolus. The revision of the sequenced species led us to re-examine related material and to describe three new species of which we do not have DNA data, but which can be readily separated from the sequenced taxa (Dyscolus placitus Moret sp. nov., D. rivinus Moret sp. nov. and D. verecundissimus Moret sp. nov.).

Taxonomic descriptions
All the species under study share the following characters, which are, therefore, not repeatedly mentioned in the descriptions of the new species: labrum with six setae near apical margin; clypeus with two pairs of setae; two supraorbital setae each side; mentum with one pair of setae, submentum with two pairs of setae (except in the species of the hirsutus and verecundus groups, which may have additional submental setae); pronotum with two pairs of lateral setae, the distal one located near the widest point of pronotum, the basal one near or at the laterobasal angle (except D. donosoi Moret sp. nov. which lacks the basal seta); parascutellar setiferous pore present; metathoracic wings reduced and metepisternum very short (except in D. eleonorae Moret sp. nov.).
The order in which these 25 new species are presented below is based on broad biogeographic and ecological groupings and on resemblances in external morphology, in order to make comparisons and identifi cations easier. The fi ve resulting sections are: 1/ full-winged montane forest species (pp. 11-13); 2/ riparian wingless species of the montane forest (pp. 13-18); 3/ wingless, non-riparian species of the montane forest (pp. 18-36); 4/ riparian species of the páramo (pp. 36-42); 5/ non-riparian species of the páramo (pp. 42-51).
All the new species are authored by P. Moret.

Diagnostic description
Habitus: Fig. 3. Fully winged. Body length: 10.8-13.7 mm. Shiny black with a strong metallic iridescence on the elytra, legs dark brown to piceous black, antennomeres brownish with the basal third reddish brown, mouthparts dark brown. Head constricted basally, vertex convex, eyes big, genae short. Pronotum broader at base than at apex, sides arcuate, hind angles rounded, laterobasal impressions deeply impressed; two pairs of lateral setae. Elytra convex, subparallel, humeri broadly bulging, striae fi nely punctate, intervals fl at to slightly convex; third interval with three setae, the second one after the middle, the third one near apex; a marked protuberance at the beginning of apical third along intervals 7 and 8; subapical sinuation deep; each elytron with a separate, rounded apical extension. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs thin; fourth metatarsomere with a subapical dorsolateral seta only on the inner lobe, apical lobes elongate, parallel, asymmetrical, the outer lobe 2.5 times longer than the inner lobe. Male genitalia: Fig. 4. Median lobe short, stout, with a short apex; endophallus densely spiculate, with longer and more sclerotized spikes at apex. Female genitalia: unstudied.

Comparisons
The asymmetrical lobes of the fourth metatarsomere, with one dorsolateral seta on the inner lobe and none on the outer lobe, are similar to those in the subgenus Dyscolidion Moret, 1990. However, in Dyscolidion the dorsolateral seta has an apical position, while in this species it is subapical, and the DNA barcoding does not suggest a phyletic proximity to Dyscolus (Dyscolidion) viridiauratus (Bates, 1878) [Process ID in BOLD: GMADN010-16, recorded as Platynus sp.].

Habitat
Humid montane forest, from 1300 to 2500 m a.s.l. Probably arboreal. One of the paratypes has been observed sitting on a leave of the lower vegetation stratum, during the day (personal observation at the Otonga Reserve).

Geographic distribution
Pacifi c slope of the Western Cordillera throughout Ecuador. Most specimens were collected at the latitude of Quito, but one paratype is from the southernmost part of the country (El Oro Province). Possibly present across the border in southern Colombia. The record from Chone, in a drier environment in the coastal lowlands of the Manabí Province, is doubtful.

Diagnostic description
Habitus: Fig. 5. Wingless. Body length: 10.2-12.2 mm. Shiny, piceous black with a faint metallic lustre on the elytra, varying from dark green to copper or pink; legs dark brown, antennae and mouthparts reddish brown to brownish. Elytral microsculpture transverse, shallowly impressed. Head broad, convex, eyes bulging, mandibles long and acute. Pronotum convex, rounded, much narrower at base than at apex, not sinuate, hind angles very obtuse, almost rounded; two pairs of lateral setae. Elytra elongate, convex, base narrow, humeri effaced; striae shallowly impressed, fi nely punctate, intervals fl at to subconvex; third interval with three setae, except one male paratype that bears fi ve setae on each side; subapical sinuation deep; apical extension triangular, recurved dorsally, more or less acute at apex. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs moderately slender; fourth metatarsomere without subapical dorsolateral setae, apical lobes parallel, the outer lobe 1.9 times longer than the inner lobe. Male genitalia: Fig. 6. Median lobe swollen in its middle part, abruptly bent ventrally in distal fourth; apical blade long and sinuate in lateral view; endophallus without sclerotized structure. Female genitalia: unstudied.

Comparisons
Metallic coloration and apically subspinose elytra separate D. giselae Moret sp. nov. from all the species (previously placed in the subgenus Hydrodyscolus) which share the derived character of a dorsally asetose fourth metatarsomere.

Habitat
Riparian in the montane forest on the Western slope of the Andes, at around 1800-2000 m a.s.l.

Geographic distribution
Southern end of the Chocó biogeographic region in northwestern Ecuador. Only known from the type locality, probably microendemic.

Diagnostic description
Habitus: Fig. 7. Wingless. Body length: 10.2-11.9 mm. Body shiny black; basal ¾ of the femora and basal ½ of the tibiae dark brown; rest of the legs, antennae and mouthparts reddish brown. Elytral microsculpture transverse. Head average, eyes moderately bulging, mandibles long and acute. Pronotum elongate, cordiform, hind angles right angled, sharp; two pairs of lateral setae. Elytra narrow, humeri wholly effaced, striae deep, entire, with traces of punctation; third interval with three setae in most specimens (four setae on both elytra of one paratype and on the right elytron of another); subapical sinuation very deep; apex with a small acute tooth at sutural angle. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs slender; fourth metatarsomere without subapical dorsolateral setae, apical lobes parallel, the outer lobe 1.5 times longer than the inner lobe. Male genitalia: Fig. 8. Middle section of the median lobe straight in lateral view, feebly arcuate distally, apex triangular in lateral view, very short; endophallus without sclerotized structure, except for a subapical denticle. Female genitalia: unstudied.

Comparisons
See below under D. incommunis Moret sp. nov.

Habitat
Riparian in the montane forest on the Western slope of the Andes, at around 2200-2250 m a.s.l. The 2100 m elevation indicated on the labels of the specimens collected in 2010 is wrong, the correct elevation is 2200 m.

Diagnostic description
Habitus: Fig. 10. Wingless. Body length: 9.6-10.7 mm. Shiny piceous black, femora dark brown, the rest of the legs, antennae and mouthparts reddish brown. Elytral microsculpture transverse. Eyes moderately bulging, genae fl at, mandibles long and acute. Pronotum convex, relatively big, elongate, narrowed backward but not cordiform; sides slightly sinuate posterad, hind angles obtuse and blunt; two pairs of lateral setae. Elytra elongate, humeri obliquely sloped, striae entire, fi nely impressed, with slight traces of punctation; third interval with three setae; subapical sinuation moderate; apical area triangle-shaped, rounded at apex. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs slender; fourth metatarsomere with or without subapical dorsolateral setae (one seta is present on the external lobe of the paratype), apical lobes parallel, the outer lobe 1.5 times longer than the inner lobe. Male genitalia: Fig. 9. Median lobe evenly arcuate, apex fi nely acute and slightly sinuate in lateral view, apex of the endophallus with a sclerotized structure formed of scales and denticles. Female genitalia: unstudied.

Comparisons
Although the BIN analysis suggests that D. aquator Moret sp. nov. and D. incommunis Moret sp. nov. are conspecifi c, the latter is differentiated from the former by important and stable morphological characters: sides of the pronotum less sinuate with obtuse hind angles; elytra broader and more convex, humeri not completely effaced, subapical sinuation weak, no apical tooth. The aedeagus′ median lobe is more arcuate, with a shorter apex and a sclerotized area differently shaped in the endophallus. Both species are sympatric in their type locality (Bellavista Lodge), but they live in different habitats.

Habitat
Montane forest on the Western slope of the Andes, at around 2250 m a.s.l. Unlike its closest relatives, this species has been collected on the forest ground, far from any stream, in the leaf litter and in fallen bromeliads.

Geographic distribution
Southern end of the Chocó biogeographic region in northwestern Ecuador. Only known from the type locality, probably microendemic.

Diagnostic escription
Wingless. Body length: 10.0-11.7 mm. Very similar externally to D. incommunis Moret sp. nov., from which it differs in the following features. Base of the pronotum narrower, lateral margins less expanded near base, hind angles more obtuse and more rounded. Apex of elytra ended with a small but protruding and acute tooth, preceded externally by a short sinuation (Fig. 11). Fourth metatarsomere without subapical setae. Male genitalia (Fig. 12) with apex of the median lobe shorter, less acute, not sinuate in lateral view, endophallus unarmed, without a dense squamose area near apex, at most with a single feebly sclerotized squama. Female genitalia unknown.

Comparisons
Closely related to D. incommunis Moret sp. nov., from which it can be distinguished by a few but constant characters, especially the rounded hind angles of the pronotum, the apical tooth of the elytra and the unarmed endophallus.

Habitat
Riparian in the montane forest on the Western slope of the Andes, at around 1900-2000 m a.s.l. This hygrophile specialisation differs from the more mesic habitat of D. incommunis Moret sp. nov.

Geographic distribution
Southern end of the Chocó biogeographic region in northwestern Ecuador, 75 km south of the type locality of D. incommunis Moret sp. nov. Probably microendemic.

Diagnostic description
Habitus: Fig. 13. Wingless. Body length: 12.2 mm. Body shiny black, with a bluish-violaceous metallic lustre on the elytra; femora and tibiae black, tarsi dark brown; palpi reddish brown, antennomeres 1-3 nigropiceous with a darker apex, 4-11 reddish brown. Elytral microsculpture transverse. Head small, very elongate, almost cylindrical, with an abrupt collar constriction; eyes strongly protruding, almost hemispherical; genae long and fl at, longer than the eyes; mandibles moderately long and acute. Pronotum almost as wide as the elytra, completely rounded basally, with broadly explanate margins and lacking anterior lobes; two pairs of lateral setae; longitudinal basal impressions narrow. Elytra elongate-oval, convex; humeri effaced; subapical sinuation weak; apex separately rounded. Striae deep and punctate, randomly broken by a few interruptions in the distal part of striae 1-4 and all along striae 5-7. Intervals 1-7 strongly convex, the third interval with two discal setae, one after the middle at ca 0.60 of the elytron length, the other in a broader fovea at 0.85; 17 umbilicate setae along the lateral margin. Last visible abdominal ventrite emarginate, with one pair (♂) of setae along its apical margin. Legs elongate and slender; fourth metatarsomere with one pair of subapical dorsolateral setae (the outer seta has fallen out but the pore is clearly visible), apical lobes asymmetrical, the outer lobe more than two times longer than the inner lobe; fi fth tarsomere of all legs setose ventrally. Male genitalia: Fig. 14

Comparisons
Given the unique combination of the following fi ve characters: elytra with violaceous metallic lustre, cylindrical head with hemispherical eyes, rounded pronotum, bisetose third interval of the elytra and setose onychia, Dyscolus globoculus Moret sp. nov. has no close relatives among Neotropical Platynini. Dyscolus globoculus Moret sp. nov. superfi cially resembles Javanese members of the genus Colpodes in its strongly protruded eyes and explanate margins of the pronotum (Liebherr 1998: fi g. 2), but it lacks several derived characters of Colpodes: head with deep sulci bordering the eyes; effaced neck constriction; elytra with a denticle at the sutural apex; fourth tarsomeres without any subapical dorsolateral seta (Moret 1989: fi g. 5;Liebherr 1998: fi gs 4-6). The protuberant eye is probably a homoplastic character which developed independently in these two lineages.

Habitat
Upper montane forest on the Eastern slope of the Andes, at around 2500 m a.s.l.

Geographic distribution
Only known from the type locality in Southern Ecuador, in the Parque Nacional Podocarpus. Probably microendemic.

Etymology
Noun in the genitive case, dedicated to David A. Donoso, assistant professor at the Escuela Politécnica Nacional (Quito), specialist of the ecology of neotropical ants, who took part in the discovery of the holotype of this new species.

Comparisons
Owing to the unique combination of a rounded basal half of pronotum, absence of pronotal hind setae, and asetose third interval of the elytra, Dyscolus donosoi Moret sp. nov. has no close relatives among the described Ecuadorian Dyscolus. In our molecular analysis (Fig. 2) D. donosoi Moret sp. nov. forms a clade with D. saxatilis Moret, 1993, but the bootstrap support for this relationship is very weak. Dyscolus saxatilis is a páramo high-altitude species with black non-metallic integuments, broad head, short mandibles, and short and robust appendages.

Habitat
Montane forest on the Eastern slope of the Andes, at around 2100-2200 m a.s.l. Active at the beginning of the night (7-8.30 pm), on the trunk of a medium-sized tree, one meter above ground.

Geographic distribution
Only known from the type locality in Southern Ecuador, in the Parque Nacional Podocarpus. Probably microendemic.

Comparisons
This species shares with D. caulatus Moret, 1993 the general form of the body and the absence of setae on the third interval of the elytra, but D. caulatus lacks the anterior supraorbital seta and the posterior pronotal seta, and has a conspicuous foveate elytral pattern.

Habitat
Upper montane forest on the Eastern slope of the Andes, at around 2870 m a.s.l. Collected by pyrethrin insecticide fogging on mossy branches and epiphytes, two meters above ground.

Geographic distribution
Only known from the type locality in Southern Ecuador, in Parque Nacional Yacuri. Probably microendemic.

Comparisons
According to the currently available molecular data, Dyscolus marini Moret sp. nov. belongs to the same clade as D. ruizi Moret sp. nov., D. arborarius Moret sp. nov., D. moretianus Perrault, 1993, D. sulcipedis Moret sp. nov. and D. rugitarsis Moret sp. nov. However, by its external characters it has a close resemblance to D. ravidus Moret sp. nov., which belongs to a different clade. Dyscolus marini Moret sp. nov. differs from D. ravidus Moret sp. nov. by the following characters: elytra with the same colour as the pronotum; eyes more convex; hind angles of the pronotum obtuse; elytral striae wider; third interval with at least two discal setae; elytral microsculpture isodiametric; male genitalia: apical blade of the median lobe two times shorter than in D. ravidus Moret sp. nov.

Habitat
Upper montane forest on the Eastern slope of the Andes, 2750-3050 m a.s.l., found at the beginning of the night on the surface of the leaf litter and above ground on mossy trunks or on epiphytes, especially in leaf axils of epiphytic bromeliads.

Geographic distribution
Only known from the type locality in Southern Ecuador. Probably microendemic.

Comparisons
Dyscolus rugitarsis Moret sp. nov. is closely related, within the D. marini Moret sp. nov. clade, to D. moretianus which inhabits the western part of the Loja Province in the Parque Podocarpus, 70 km northwest from the type locality of this new species. Dyscolus rugitarsis Moret sp. nov. differs from the latter by the blunter hind angles of the pronotum, the presence of at least one seta in the basal half of the third elytral interval (third interval asetose in D. moretianus), slightly longer antennae and tarsi, and a markedly longer external lobe at the apex of the fourth metatarsomere. For comparison with D. sulcipedis Moret sp. nov., see below (p. 27).

Habitat
Upper montane forest, at around 3200-3250 m a.s.l. Active at the beginning of the night on the surface of the leaf litter.

Habitat
Upper montane forest, at around 3240 m a.s.l. Active at the beginning of the night on the surface of the leaf litter and on tree trunks 2 m above ground.

Geographic distribution
Only known from the type locality in Southern Ecuador, in Parque Nacional Yacuri. Probably microendemic.

Diagnostic description
Habitus: Fig. 25. Wingless. Body length: 10.8 mm. Body dark brown with the sides of the pronotum and the elytral sutura reddish brown; femora dark brown; rest of the legs, antennae and mouthparts reddish brown. Elytral microsculpture shallowly impressed, transverse. Head large and convex; eyes prominent, very convex; basal constriction strong; mandibles relatively short. Antennae thin, very elongate. Pronotum subcordiform, sides moderately arcuate anterad, slightly sinuate posterad; hind angles obtuse; two pairs of lateral setae. Elytra elongate-ovate with effaced humeri, subapical sinuation obsolete; striae narrow and shallowly impressed; intervals fl at; third elytral interval with two or three discal setae. Last visible abdominal ventrite of the male with two pairs of apical setae. Legs thin and elongate. Fourth metatarsomere oval-shaped, with one pair of dorsolateral subapical setae and long apical lobes, the external lobe nearly two times longer than the inner lobe. Male genitalia: Fig. 26. Median lobe weakly arcuate, apex short, acute, endophallus with one subapical sclerotized structure. Female genitalia: unknown.

Comparisons
According to the COI gene, D. arborarius Moret sp. nov. belongs to the marini clade, close to D. ruizi Moret sp. nov. It can be easily distinguished from the other species of this clade by its slender body, large head with very prominent eyes, and effaced humeri.

Habitat
Upper montane forest on the Eastern slope of the Andes, at around 2870 m a.s.l. Collected by pyrethrin insecticide fogging on mossy branches and epiphytes, two metres above ground.

Comparisons
The slender, elongate body form, the dashed elytral setae and the hooked apex of the aedeagus separate D. famelicus Moret sp. nov. from any other species having an asetose third elytral interval. Our molecular analysis (Fig. 2) places this species close to D. gobbii Moret sp. nov. which lives in the páramo and has a very different external morphology, but the bootstrap support for this relationship is weak.

Habitat
Upper montane forest on the Eastern slope of the Andes, at around 2700 m a.s.l., collected above ground at the beginning of the night, on mossy trunks and on epiphytes.

Comparisons
This species belongs to the same clade as D. ravidus Moret sp. nov. Both have the same general aspect and share the character of an asetose third interval, but in D. crespoae Moret sp. nov. the dorsal surface is shiny with a transverse microsculpture on the elytra, the elytra are more convex, the hind angles of the pronotum more obtuse, the ventral pubescence of the tarsi denser, the apical lobes of the fourth metatarsomere more asymmetrical.

Habitat
Montane forest on the Eastern slope of the Andes, at around 2100 m a.s.l. Active at the beginning of the night (7.30-9 pm) on the surface of the leaf litter above ground.

Geographic distribution
Only known from the type locality in Southern Ecuador, in the Parque Nacional Podocarpus. Probably microendemic. Moret sp. nov. urn:lsid:zoobank

Comparisons
According to the currently available molecular data, Dyscolus ravidus Moret sp. nov. belongs to the same clade as D. crespoae Moret sp. nov. (see above in the description of the latter for the morphological differences between them). However, by its external characters, D. ravidus Moret sp. nov. has a closer resemblance to D. marini Moret sp. nov., which belongs to a different clade. Dyscolus ravidus Moret sp. nov. differs from D. marini Moret sp. nov. by the following characters: elytra brownish in contrast with a black or piceous black pronotum; eyes less convex; hind angles of the pronotum sharper, right angled; elytral striae narrower and more shallowly impressed; third interval without discal setae; elytral microsculpture transverse; male genitalia: apical blade of the median lobe two times as long as in D. marini Moret sp. nov.

Habitat
Upper montane forest on the Eastern slope of the Andes, between 2700 and 2900 m a.s.l. Active at the beginning of the night (7.30-9 pm) on the ground and in leaf axils of epiphytic bromeliads.

Geographic distribution
Only known from the Parque Nacional Podocarpus in Southern Ecuador. Probably microendemic. Moret sp. nov. urn:lsid:zoobank

Diagnostic description
Habitus: Fig. 34. Wingless. Body length: 9.2-11.3 mm. Entirely black or piceous black. Elytral microsculpture fi nely transverse, weakly impressed. Head broad basally, feebly constricted, eyes small, genae long and convex; antennae short. Pronotum elongate, wider at apex than at base; sides moderately arcuate apically, straight or feebly sinuate basally; hind angles obtuse and blunt; two pairs of lateral setae, the basal one anterad the hind angle. Elytra narrow depressed medially; no subapical sinuation; striae entire, well impressed; intervals fl at to subconvex. Third elytral interval with two setae, one near base and one near apex, or with only one subapical seta. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs robust, anterior femora swollen; dorsal surface of tarsi rugose, almost squamose, not sulcate; fourth metatarsomere with one pair of dorsolateral subapical setae, apical lobes short, the external lobe two times longer than the inner lobe; fi fth metatarsomere asetose ventrally. Male genitalia: Fig. 35. Median lobe arcuate, apex robust, acute, endophallus without sclerotized structure. Female genitalia: unstudied.

Comparisons
Dyscolus ruizi Moret sp. nov. belongs to the same clade as D. marini Moret sp. nov., D. arborarius Moret sp. nov., D. moretianus, D. sulcipedis Moret sp. nov. and D. rugitarsis Moret sp. nov. However, by its robust legs, short antennae, small eyes and convex genae, D. ruizi Moret sp. nov. is more similar to species living at higher elevations in the páramo, as D. denigratus (Bates 1891), than to these species which share the same forest environment beneath 3000 m. Dyscolus ruizi Moret sp. nov. differs from D. denigratus by its elongate pronotum and the absence of the middle seta on the third interval.

Habitat
Upper montane forest and subpáramo on the eastern slope of the Andes, at around 2800-3000 m a.s.l. The specimens from Saraguro-Huashapamba were collected in pitfall traps baited with dung.

Geographic distribution
Only known from the Parque Nacional Podocarpus and surroundings, in southern Ecuador. Probably microendemic. Moret sp. nov. urn:lsid:zoobank

Diagnostic description
Habitus: Fig. 32. Wingless. Body length: 9.3-9.6 mm. Head, pronotum and elytra brunneopiceous; legs, antennae and mouthparts reddish brown. Elytral microsculpture isodiametric. Head convex, markedly constricted basally, eyes small, not bulging, genae almost fl at in dorsal view, as long as the eyes. Pronotum elongate, convex; sides weakly arcuate anterad, almost straight in basal fourth, with a very short sinuation at level of the basal seta, hind angles obtuse and blunt, anterior angles pronounced, broadly rounded; two pairs of lateral setae. Elytra elongate, subparallel; base relatively narrow, shoulders rounded; sides arcuate apically without subapical sinuation; striae entire, well impressed, not punctate; intervals slightly convex. Third elytral interval asetose. Legs rather short, meso-and metatarsomeres 1-3 with a lateral sulcus on each side; fourth metatarsomere with one pair of dorsolateral subapical setae, its apical lobes very short, the outer lobe twice as large as the inner lobe; fi fth metatarsomere asetose ventrally. Last visible abdominal ventrite of the male with one pair of setae along its apical margin. Male genitalia: Fig. 33. Median lobe almost straight in its basal half, shortly arcuate before apex; apex slightly refl exed upward, narrow and acute in lateral view; endophallus without sclerotized structure. Female genitalia: unknown.

Comparisons
Closely related to D. moreti Perrault, 1993, but smaller (most specimens of D. moreti range from 10 to 11 mm), the lateral margin of the pronotum narrower, the elytra narrower and less convex, the apex of the aedeagus shorter, more acute, and not bent ventrally as in D. moreti.

Habitat
Upper montane forest on the Eastern slope of the Andes, at around 2850 m a.s.l.

Geographic distribution
Only known from the type locality in Southern Ecuador. Probably microendemic.

Riparian species of the páramo
Dyscolus arauzae Moret sp. nov., D. piscator Moret sp. nov. and D. placitus Moret sp. nov. belong to the group of D. hirsutus Moret, 2005, which differs from all the other equatorial clades of Dyscolus by the combination of four characters: elongate body with slender appendages; ventral abdominal sclerites entirely or partly reddish; submentum with supernumerary setae; fi fth elytral interval with setae. Dyscolus hirsutus was described on specimens from two distant páramos, both in the Eastern Cordillera: Guamaní and El Altar, 150 km apart. A closer examination of the type series shows that the Guamaní population (North) and the El Altar population (South) belong to two different species. As the holotype of D. hirsutus came from El Altar, the Northern population had to be described as new. It has been named D. placitus Moret sp. nov. Furthermore, the material collected in 2015-2017 revealed the existence of a second species, D. piscator Moret sp. nov., sympatric with D. placitus Moret sp. nov. at Guamaní ca 3900 m a.s.l., but living in a different microhabitat. A third species closely related to D. piscator Moret sp. nov.and D. placitus Moret sp. nov., D. arauzae Moret sp. nov., lives at a higher elevation (4300 m a.s.l.) on Mt Cayambe.

Diagnostic description
Habitus: Fig. 38. Wingless. Body length: 8.8-9.6 mm. Dorsal surface of the body entirely dark brown; ventral surface mostly dark brown, with reddish areas on the sides of the abdominal sclerites and on most parts of the coxae and trochanters; legs, antennae and mouthparts reddish brown. Elytral microsculpture made of slightly oblong sculpticells. Head convex, collar constriction strong, eyes moderately bulging, genae fl at in dorsal view, mandibles moderately long and acute; submentum with two or three pairs of setae, mentum with one pair of setae. Pronotum cordiform, hind angles right angled, blunt; two pairs of lateral setae. Elytra elongate-oval, narrow, without subapical sinuation, rounded at apex; humeri almost entirely effaced. Striae well impressed, intervals feebly convex. Third interval with 3-6 setae, fi fth interval with 0-3 setae, seventh interval with 0 or 1 seta; 17 to 20 umbilicate setae along the lateral margin. Last visible abdominal ventrite with one pair (♂) or two pairs (♀) of setae along its apical margin. Legs slender; fourth metatarsomere without subapical dorsolateral setae, apical lobes parallel, the outer lobe 1.5 times longer than the inner lobe. Male genitalia: Fig. 41. Apex of the median lobe bent downwards; endophallus without sclerotized structure. Female genitalia: unstudied.

Comparisons
Dyscolus hirsutus, now geographically restricted to the El Altar volcano, differs from D. placitus Moret sp. nov. by the following characters: upper surface bicolorous with the head and the elytra dark brown, the pronotum reddish; ventral surface (including the epipleura) entirely reddish, except for the head which is dark brown; mandibles with a long and very acute tip; eyes more bulging; hind angles of the pronotum broadly rounded; humeri more effaced; seventh interval with 1 or 2 setae; apex of the median lobe broader, apically rounded (slightly hooked in D. placitus Moret sp. nov.). The differences with D. arauzae Moret sp. nov. and D. piscator Moret sp. nov. are set forth in the diagnoses of these species.

Habitat
Hygrophilous in the upper montane forest, at around 3800-3900 m a.s.l. The specimens collected in 2017 were found in the understory of a Polylepis forest, beneath soaked mosses on the ground.

Geographic distribution
Microendemic species, restricted to the Guamaní area in the Eastern Cordillera.

Etymology
Latin adjective meaning 'fi sherman', in allusion to the feeding habits of this riparian species which preys on small water invertebrates.

Habitat
Riparian in the upper montane forest with Polylepis, at around 3900 m a.s.l. Under stones at the edge of a small stream. This biotope is 50 m away from that of D. placitus Moret sp. nov.

Geographic distribution
Microendemic species, restricted to the Guamaní area in the Eastern Cordillera. Moret sp. nov. urn:lsid:zoobank

Habitat
Riparian along glacier-fed streams in the humid superpáramo at 4300-4400 m a.s.l.

Geographic distribution
Microendemic species, presumably restricted to the Cayambe mountain. Moret sp. nov. urn:lsid:zoobank.org:act:787DC50E-8874-4E7F-BCAD-FE33230508A1 Fig. 42 Etymology Noun in the genitive case, dedicated to Mauro Gobbi, specialist of the ecology of glacier-foreland invertebrates, who took part in the discovery of this species.

Diagnostic description
Habitus: Fig. 42. Wingless. Body length: 10.2 mm (holotype) to 12.3 mm (paratype). Body piceous black, legs brownish with reddish-brown femora. Elytral microsculpture made of oblong sculpticells, shallowly impressed. Head big, convex dorsally; mandibles long and sharp, as long as the length of the head from base to apex of the labium; eyes moderately bulging, genae oblique, slightly convex in dorsal view. Pronotum transverse, subquadrate, narrower at base than at apex; lateroapical lobes broadly rounded; hind angles obtusely rounded; two pairs of lateral setae. Elytra oval-shaped, convex, with effaced humeri. Striae weakly impressed, with traces of punctuation; intervals 1-4 slightly convex, 5-8 fl at; third interval with fi ve or six discal setae. Last visible abdominal ventrite of the female with two pairs of setae along its apical margin. Legs slender but short; fourth metatarsomere with convex sides, without subapical dorsolateral setae (holotype) or with weak and short subapical dorsolateral setae (paratype), apical lobes ovoid, the outer lobe slightly longer than the inner lobe. Male genitalia: unknown. Female genitalia: unstudied.

Comparisons
This isolated species remotely resembles Dyscolus segnipes Moret, 1990, with which it coexists in the Guamaní páramo. Dyscolus segnipes has much shorter mandibles and well developed subapical dorsolateral setae on the fourth metatarsomere. The COI tree places gobbii and segnipes in two clades far apart from each other.

Habitat
Humid superpáramo, from 4200 to 4400 m a.s.l. The Guamaní specimen was collected under a stone on the shore of a small lake; the Antisana specimen was collected by pitfall trapping in a mesic environment, 20 metres from a glacier-fed stream.

Geographic distribution
Microendemic species, restricted to the Antisana and Guamaní areas in the Eastern Cordillera.

Diagnostic description
Habitus: Fig. 43. Wingless. Body length: 8.9-9.2 mm. Body with a varying coloration, from reddish brown to piceous black, the sides of the pronotum and of the elytra lighter than the discal area; legs, antennae and mouthparts testaceous to reddish brown. Elytral microsculpture isodiametric, distinctly impressed. Head broad and convex, neck markedly constricted, eyes moderately bulging, genae oblique, almost fl at in dorsal view; antennae moniliform, very short. Pronotum transverse, quadrate; lateroapical lobes small, not protruding; sides feebly arcuate distally, almost straight in basal third; hind angles bluntly obtuse, almost rounded; two pairs of lateral setae. Elytra subparallel, convex; preapical sinuation obsolete; striae fi ne and shallowly impressed, almost obsolete; intervals fl at or subconvex; third elytral interval without discal setae. Legs short and robust; fourth metatarsomere triangular, with one pair of apical dorsolateral setae; apical lobes very short, the outer lobe slightly larger than the inner lobe; fi fth metatarsomere asetose ventrally. Last visible abdominal ventrite of the male with one pair of setae along its apical margin. Male genitalia: Fig. 44. Median lobe weakly arcuate with a long apical blade, endophallus unarmed. Female genitalia: unknown.

Comparisons
This species closely resembles Dyscolus saxatilis Moret, 1993 in its general sturdy aspect, asetose third elytral interval, shallow striae, and in the form of the fourth metatarsomere. It differs from it by its smaller size (D. saxatilis: 10.0-10.8 mm), elytral microsculpture less granulose, eyes bigger, genae shorter and fl at (convex in D. saxatilis), lateroapical lobes of the pronotum smaller. Our molecular analysis (Fig. 2) puts D. barragani Moret sp. nov. closer to D. hebeculus (Bates, 1891), with a very weak bootstrap support. Dyscolus hebeculus shares with D. barragani Moret sp. nov. an asetose third elytral interval, but otherwise the morphological differences are many. Dyscolus hebeculus has smaller eyes, longer antennae, a more elongate pronotum with more developed hind angles, elytral striae well impressed, and markedly asymmetrical lobes at the apex of the fourth metatarsomere.

Geographic distribution
Microendemic species, restricted to the Ayapungu massif. (Bates,  Dyscolus denigratus is a highly variable species, with a broad range of variation within each population regarding body size, form of the pronotum, length of the legs, and form of the apex of the male aedeagus (48)(49)(50). The tree based on the COI sequences separates two clades (Fig. 2): one in the south-west part of the area of D. denigratus, represented in our dataset by the Cotopaxi and Pichincha populations, vs a north and north-east zone represented by the Cotacachi, Cayambe and Guamaní populations. Additionally, three species-level units are recognized through BIN analysis: the fi rst one embraces the populations of the south-western clade, including specimens identifi ed as D. palatus. The northern clade is split into two BIN units: one for the Guamaní population, and another for the Cotacachi and Cayambe populations. However, we were unable to fi nd any morphological character that reliably refl ects these divisions. According to the form of the apex of the median lobe, the Cotacachi population ( Fig. 49) seems to be closer to the Pichincha population (Fig. 50) than to the Guamaní one (Fig. 48), in contradiction with the molecular results. Obviously, the phylogeny of the denigratus complex cannot be resolved from the COI sequences alone.

Dyscolus denigratus
We thus take a conservative approach, with no taxonomic decision regarding the specifi c or subspecifi c status of the components of the denigratus complex, except for the synonymy of Dyscolus palatus with Dyscolus denigratus. Dyscolus palatus was described from specimens from the Pichincha, Atacazo and Corazón volcanoes in the Western Cordillera (Fig. 47). This taxon was then thought to be more closely related to D. altarensis (Bates, 1891) than to D. denigratus, due to its relatively broader and shorter body. The specimen of D. palatus included in our analysis (COI voucher PM359-01, BOLD sequence SUM168-18) instead suggests that D. palatus and D. denigratus are conspecifi c. A re-examination of the type series of D. palatus and of large series of specimens recently collected on the Pichincha shows that the body dimensions of the two taxa partly overlap (Fig. 51). Owing to the existence of these transition forms, D. palatus cannot be maintained as a good species and we therefore treat it as a synonym of D. denigratus.

Comparisons
Dyscolus silvestris Moret sp. nov. is similar to the shorter and broader individuals of Dyscolus denigratus. The main differences are: head slightly bigger and broader; sides of the pronotum sinuate basally, hind angles sharply obtuse (completely rounded in D. denigratus); femora of the male not swollen as in D. denigratus. The apex of the median lobe of the male genitalia has a shorter apex, and the endophallus is entirely unarmed (with a heavily sclerotized area in D. denigratus).

Habitat
Upper montane forest on the Amazonian slope of the Andes, from 2700 to 3400 m a.s.l. This species lives in the leaf litter.

Geographic distribution
Only known from the type locality around Papallacta in Northern Ecuador, probably microendemic.

The group of Dyscolus verecundus
Dyscolus verecundus Moret, 1998 was originally thought to be a morphologically variable species, widespread in the northern Andes of Ecuador. The type series includes specimens from fi ve volcanoes: Pichincha, Atacazo, Illiniza, Cotopaxi and Chimborazo. The COI sequences revealed a strong separation between two specimens collected on Mt Pichincha and a third one from Mt Illiniza. The re-examination of the material used to describe this uncommon species confi rms the existence of several microendemic taxa, two of which are described herein. Further investigation will probably reveal a highly fragmented pattern, with vicariant species on each mountain. Furthermore, the two specimens of Mt Pichincha (PM360-01 / SUM194-18 from Rucu Pichincha, the western summit of this mountain, and PM 385-02 / SUM250-18 from Guagua Pichincha, its eastern summit) are also separated by relatively long branches, and the BIN analysis assigns them to different species. However, no morphological differences could be detected between these two populations. The distance from Rucu Pichincha to Guagua Pichincha is only 4 km, with no discontinuity of the páramo ecosystem at the elevations where the specimens have been found (4200 to 4500 m). Pending more comprehensive analyses, we kept a conservative position and decided not to split the Pichincha population of D. verecundus into two species.
Another important result of the molecular analysis is that D. verecundus is found near the base of the tree, far from the rest of the páramo Dyscolus species, among clades of Central and North America. It shares with these low altitude platynines, traditionally placed in the genus Platynus, an emarginate, triangularshaped fourth metatarsomere, without protruding apical lobes and with dorsal setae in a lateroapical position, two character states that can be considered plesiomorphic, whereas the rest of the páramo lineages of Dyscolus have a bilobed fourth metatarsomere with dorsolateral setae in a subapical position.

Diagnostic description
Habitus: Fig. 54. Wingless. Body length: 10.8-11.0 mm. Body colour variable, from reddish brown to piceous black, the vertex and the sutural zone of the elytra reddish; legs, antennae and mouthparts reddish. Elytral microsculpture isodiametric, distinctly impressed. Head robust, convex, neck strongly constricted, eyes moderately bulging, genae almost fl at in dorsal view; antennae short. Pronotum transverse, subquadrate; sides moderately arcuate distally, weakly sinuate in basal third; hind angles variable, obtuse or almost quadrate, but always blunt; two pairs of lateral setae. Elytra subparallel, disc almost depressed; humeri rounded but not effaced, subapical sinuation obsolete; striae thin, shallowly impressed; intervals fl at; third elytral interval with 3-5 discal setae. Fourth metatarsomere triangular, with one pair of lateroapical setae and without subapical dorsolateral setae; apical lobes very short, the inner one not extending beyond the insertion point of the lateroapical seta; fi fth metatarsomere asetose ventrally. Last visible abdominal ventrite of the male with one pair of setae along its apical margin (♂). Male genitalia: Fig. 58. Median lobe weakly arcuate; apical blade bluntly triangular in dorsal view; endophallus with two sclerotized squamose areas, ovoid in shape, median and subapical. Female genitalia: unknown.

Comparisons
This species closely resembles D. verecundus, now restricted to Mt Pichincha, its type locality. Dyscolus verecundior Moret sp. nov. differs from D. verecundus in a few external and genital characters: genae almost fl at (convex in D. verecundus); antennae slightly thicker, the second antennomere of the male 1.41 to 1.49 times as long as wide (1.50 to 1.65 times in D. verecundus); fourth metatarsomere thinner, 1.16 to 1.20 times as long as wide (1.07 to 1.13 times in D. verecundus); apical blade of the median lobe thinner in lateral view (compare Figs 56-57 and 58); sclerotized areas of the endophallus bigger.

Geographic distribution
Microendemic species, restricted to the Illiniza and Corazón volcanoes in the Western Cordillera. Moret sp. nov. urn:lsid:zoobank

Geographic distribution
Microendemic species, restricted to the Chimborazo volcano in the Western Cordillera.

Comments
The isolated population we raise here to species rank was initially assigned to D. verecundus by Moret (2005). There are more morphological differences between D. verecundissimus Moret sp. nov. and the pair of sister species D. verecundus and D. verecundior Moret sp. nov. than between the two latter, especially in the male aedeagus.

Discussion
The species delimitations generated by the Barcode Index Number (BIN) System (Ratnasingham & Hebert 2013) are most often congruent with the results of a previous taxonomic study, only based on characters of external and genital morphology (Moret 2005). Discrepancies with the morphology-based identifi cations occur in 14 cases, out of 64 morphospecies under study. All these cases are examined below, fi rst the cases of species merged by the Bin analysis, then the species split by the BIN analysis. As a general rule, our taxonomic decisions were based on a conservative approach: we only followed the indications of molecular analysis if they have the support of stable and unambiguous morphological characters.

Species merged by the BIN analysis
In four cases, pairs of taxa are merged into one by BIN assignments: D. denigratus with D. palatus, D. fucatus Moret, 2005 with D. robiginosus Moret, 2005, D. diopsis (Bates, 1891) with D. oreas (Bates, 1891) and D. aquator Moret sp. nov. with D. incommunis Moret sp. nov. In the fi rst case, we follow the results of the barcoding-based delimitation. In the three others, robust morphological characters support the separation of two species.
1/ The re-examination of the type material has confi rmed that D. palatus and D. denigratus are conspecifi c. We therefore proposed to synonymize D. palatus with D. denigratus, the former being a junior subjective synonym of the latter (see the discussion above, pp. 43-45).
2/ Dyscolus fucatus and D. robiginosus are sympatric in the Ayapungu massif. After examination of the type specimens and of new material collected in 2015-2017, we maintain the existence of two different species. The characters already mentioned in the original description (Moret 2005) are confi rmed, and several more can be added. In D. fucatus the overall body size is bigger (9.6-11.0 mm vs 8.6-9.5 mm in D. robiginosus); the upper surface is shinier, due to a shallower microsculpture with more oblong sculpticells; the antennae are longer; the elytra are less elongate (length from basal bead to apex / maximum width = 1.47-1.48 in D. fucatus, 1.53-1.62 in D. robiginosus); the fourth metatarsomere is thinner with much more asymmetrical apical lobes. Above all, the number of umbilicate setae along the lateral margin of the elytron is always higher (14-19 in D. fucatus, 10-12 in D. robiginosus).
3/ Dyscolus diopsis and D. oreas are closely related taxa, both adapted to extreme environments at the highest elevations of Northern Ecuador′s superpáramos. Dyscolus diopsis has been found on several mountains (Cotacachi, Pichincha, Corazón, Illiniza, Cayambe, Guamaní, Antisana, Cotopaxi, Quilindaña, Llanganatis), generally above 4500 m a.s.l. (Moret et al. 2016). Being probably a southern vicariant of D. diopsis, D. oreas is a microendemic taxon, restricted to the connected superpáramos of Mt Chimborazo and Mt Carihuairazo. The morphological differences between D. diopsis and D. oreas, although slight, are constant and higher than the differences among the currently isolated populations of D. diopsis (Moret 2005). 4/ We found strong morphological arguments to distinguish two new species, D. aquator Moret sp. nov. and D. incommunis Moret sp. nov., among the material of recent samplings in the montane forest of Western Ecuador (see above pp. 14-16), although the BIN algorithms give them the same number, therefore suggesting that they are conspecifi c.
In the fi rst seven cases, the species, as currently defi ned, proves to be monophyletic, and most of the included BINs correspond to high-altitude isolated populations that have probably begun a genetic differentiation. However, the morphological characters that could support the recognition of a species or subspecies status for these BIN-based entities are extremely weak (fusipalpis, orthomus, oopteroides) or could not be found at all (diopsis, rotundiceps, lucifuga, alpinus). We therefore decided to be conservative and not to split these species into different taxa. The more complex case of D. denigratus is discussed above (pp. 43-45).
In only two cases, those of D. hebeculus and D. verecundus, have we seen the need to split the existing taxon into two species, in agreement with the BIN analysis. Regarding D. verecundus Moret, 1998, please refer to the discussion in pp. 47-51 (D. verecundior Moret sp. nov.). The description of a new sibling species is postponed in the case of Dyscolus hebeculus (Bates, 1891). The holotype of this species was collected by Edward Whymper in 1880 "between Antisanilla and Piñantura, 11,000 feet"; in the vast subpáramo grasslands that extend west of Mt Antisana (Pichincha Province). The distribution area of D. hebeculus was later enlarged to the Cotopaxi Province (Moret 2005). Two specimens assigned to this species have been sequenced, one from Papallacta (COI voucher PM041-05, BOLD sequence SUM126-18), the other from the lower slopes of Mt Cotopaxi (COI voucher PM511, BOLD sequence SUM311-18), revealing a long molecular distance between them. As the type locality of D. hebeculus is just halfway between the sample sites of these two vouchers, a re-examination of the unique male holotype, held in the Bates Collection in Paris (MNHN), will be required to assign one of the barcoded specimens to the existing taxon, and describe the other as a new species.