A new cryptic species and review of the east-Andean leaf chafer genus Mesomerodon Ohaus, 1905 (Coleoptera, Scarabaeidae, Rutelinae)

Abstract The Neotropical scarab beetle genus Mesomerodon Ohaus (Scarabaeidae: Rutelinae: Rutelini) is distributed in the western (lowland) Amazonian region from Colombia to Bolivia. Based on our research, the genus includes three species including a new cryptic species from Ecuador. We use niche modeling to predict potential suitable habitat and identify environmental factors associated with the distribution of Mesomerodon species. We characterize the genus, provide a key to species, diagnose each species, describe a new species, provide spatial and temporal distributions, and discuss distributions of the species in relation to Amazonian landscape biodiversity.


Introduction
The South American genus Mesomerodon Ohaus (Figs 1-8) is a member of the pelidnotine leaf chafer scarabs, a polyphyletic assemblage of beetles which are in need of comprehensive revision (Moore et al. 2017). Members of the genus Mesomerodon are ovate, 17-25 mm in length, and cream-colored when alive . After death the color fades to testaceous or yellowish with weak metallic reflections. The genus is sexually dimorphic: males possess elongated, spinose elytral apices as well as an acute process on the posterior margin of the mesofemur (both lacking in females). The unusual form of the spinose elytral apex is a character state that is shared with males of the Neotropical leaf chafer Hoplopelidnota metallica (Laporte, 1840). Sister group relationships have not been addressed, host plant information is lacking, and larvae are undescribed. Members of this distinctive but poorly known group are distributed in lowland Amazonian regions (ca. 150-760 m elevation) from Colombia to Bolivia and are collected at light at night. In overall body form, the genus Mesomerodon resembles some species of Pelidnota MacLeay, 1819 (with which it is closely allied). Ohaus (1905) established the genus Mesomerodon Ohaus and included in it one species with 'peculiar sexual characteristics' from Peru. Nearly 100 years after Ohaus' (1905) description of the genus, Soula (2008) described a second species, Mesomerodon gilletti from Ecuador. In an overview of pelidnotine leaf chafers (Moore et al. 2017), the distribution of the genus was expanded to include Colombia and Bolivia. The distributional data provided for the genus Mesomerodon (Moore et al. 2017) were derived from the data in this study and are therefore given for the first time with specimen associations.
Species in the genus Mesomerodon possess many external similarities in form, but the male genitalia are sufficiently different as to warrant species status for three, distinct operational taxonomic units that we treat as species. Soula (2008) recognized M. gilletti as distinct from M. spinipenne based entirely on the form of the male genitalia ( Fig. 23 versus Fig. 24). Cryptic species such as these are difficult or sometimes impossible to distinguish morphologically, and they are often incorrectly classified as a single taxon (Beheregaray andCaccone 2007, Bickford et al. 2007). Our synthesis of information on this group of beetles, which is based on 302 specimens and morphological data, led to the unveiling of an additional cryptic species in the genus.
We provide a synthesis of the biodiversity of the genus, including descriptions, key to species, diagnoses, and images. As a result of our research, the genus Mesomerodon includes three species, all of which are distributed in the western (lowand) Amazonia, including a new unexpected and cryptic species.

Material and methods
Characters. Morphological characters formed the basis of this work. The broadest range of potentially phylogenetically informative morphological characters was used for morphological analyses and comparisons. Morphological terminology is based primarily on Jameson (1998), however we use the term venter instead of sternum and antennomeres instead of antennal segments. Antennomeres are defined as the pedicel plus flagellum (or flagellum and club). Consistent with use of venter, the term mesometasternum is replaced with mesometaventrum and the term sternite is replaced with ventrite. For measurements, we used an ocular micrometer. Body measurements, puncture density, puncture size, and density of setae are based on the following standards. Body length was measured from the apex of the clypeus to the apex of the pygidium. Body width was measured at the widest width of the elytra. Puncture density was considered 'dense' if punctures were nearly confluent to less than two puncture diameters apart, 'moderately dense' if punctures were from two to six puncture diameters apart, and 'sparse' if punctures were separated by more than six puncture diameters. Puncture size was defined as 'small' if punctures were 0.02 mm in diameter or smaller; 'moderate' if 0.02-0.07 mm, 'moderately large' if 0.07-0.12 mm, and 'large' if 0.12 mm or larger. Setae density was defined as 'dense' if the surface was not visible through the setae, 'moderately dense' if the surface was visible but with many setae, and 'sparse' if there were few setae. It should be noted that setae are subject to wear and may be abraded away. Elytral striae are defined as the striae located between the elytral suture and the elytral humerus. The interocular width measures the number of transverse eye diameters that span the width on the frons between the eyes. This was measured by placing the ocular micrometer in a position such that it intersects the frons and eyes (dorsal view), focusing on the surface of the frons, and then measuring the width of the frons and width of the eyes without adjusting the focus. Sclerotized portions of the male genitalia are used for diagnosis and identification. This includes the parameres, phallobase (or "basal piece" [d'Hotman and Scholtz 1990]), and the ventral sclerite of the phallobase (e.g., Fig. 22). Mouthparts, wings, and genitalia were examined and card-mounted beneath the specimen.
Characters and specimens were observed with 6-48× magnification and fiber-optic illumination. Digital images of specimens and structures were captured using the Leica Application Suite V3.8. Images were edited in Adobe Photoshop CS2 (background removed, contrast manipulated).
Species concept. Species are characterized by combinations of characters including the form the male protarsomeres and form of the male parameres in caudal and lateral views, and form of the ventral sclerite of the phallobase. Identification of female specimens required associated males from the same collecting event (place and date). We use the phylogenetic species concept (Wheeler and Platnick 2000) in this work: "A species is the smallest aggregation of (sexual) populations or (asexual) lineages diagnosable by a unique combination of character states." Locality data. Locality data for all specimens examined as part of this study were translated into decimal latitude and longitude using GoogleEarth (http://www.google. com/earth/index.html) and provided (Suppl. material 1). If latitude and longitude data were not included in label data or were too vague to be informative, then we searched for GPS data in GoogleEarth and GoogleMaps. Maps were generated by entering the coordinates into an Excel sheet. These files were subsequently used to construct distribution maps using an R script which plots the distribution points on an elevational map constructed using Global Land One-kilometer Base Elevation (GLOBE) data (Hastings et al. 1999). For species accounts, locality data are recorded for country, department and province (Bolivia), region and province (Peru), department (Colombia), or province (Ecuador). Additional locality data are provided (Suppl. material 1).
Distribution modelling. To model the potential distribution and to identify environmental factors associated with Mesomerodon species, we used the maximum entropy algorithm (MaxEnt; Phillips et al. 2006, Elith et al. 2011 for the species distribution modeling and followed the workflow of Fikáček et al. (2014). For occurrence data, we summarized all available records for all species of Mesomerodon. We ran three independent analyses: a genus-level distribution analysis that included data for all species and unidentified specimens, an analysis of the distribution of Mesomerodon spinipenne, and a combined analysis of the Ecuadorian and Colombian species (M. gilletti + M. barclayi sp. n.). A separate analysis for either M. gilletti or M. barclayi sp. n. was not conducted due to the few number of data points per species. Because two species occur in sympatry and in similar climatic conditions (M. gilletti and M. barclayi sp. n.), we concluded that the combined species analysis is justified. Furthermore, a combined analysis of the latter two species made it possible to include data points with unassociated species (female specimens which could not be identified). We employed the high-resolution climate data available in the Worldclim database (Hijmans et al. 2005) containing 19 layers of climatic variables. Analyses were performed in R (MaxEnt command in Dismo package). After mapping the ecological niche of the genus and/ or species, prediction values were converted into binary values (presence and absence) using the threshold for maximum training sensitivity and specificity provided by the outputs of the resulting models (Figs 19-21).
Type specimens. Friedrich Ohaus provided a legacy for understanding the biodiversity of Rutelinae with over 170 published papers and research collections (for biography see Smith 2003). Perhaps due to concern with destruction of museums during World War II, Ohaus often labeled specimens as types long after publication (e.g., Kuijten 1988Kuijten , 1992Jameson 1990Jameson , 1998Smith 2003). These erroneous type specimens can be recognized because label data are incongruous with data in the original description. As part of this research, we found 10 specimens that were labeled as type specimens. Six of these specimens do not belong to the syntype series of M. spinipenne (Suppl. material 1): one specimen from ZMHB, four specimens from ZSM, one specimen from NHMB). To reduce future confusion, these were labeled "NOT a type specimen of Mesomerodon spinipenne, Ohaus, 1905, des. Seidel 2016" (see "Remarks" for M. spinipenne").

Collections (Suppl. material 1).
This research is based on 302 specimens in 25 collections. The material examined in the present study is housed in the following collections and was provided by the curators and/or collections managers:
Etymology. From the Greek, "mesos" meaning middle or in the middle, "mero" meaning femur, and "odon" meaning tooth. The name refers to the spinose process on the posterior margin of the mesofemur in males, a synapomorphy for species in the genus. The gender is neuter.
Composition and distribution (Fig. 18). Three species distributed on western (lowland) Amazonia from Colombia, Ecuador, Peru, and Bolivia. An erroneous record from Brazil (Ohaus 1905) was repeatedly cited by subsequent authors (Blackwelder 1944, Ohaus 1934, 1952, Machatschke 1972, Krajcik 2008, Soula 2008, Moore et al. 2017) (see Mesomerodon spinipenne type material). We record the genus from elevations between 150 to 762 m. A record of the genus occurring at 2800 m (Paucar-Cabrera 2005) is beyond the limits of the genus, and we consider it erroneous. A locality record of M. gilletti from Loja (Ecuador) (Fig. 18 [indicated with question mark]) waits for confirmation through future collecting since a short series of specimens supposedly collected from Loja province (west side of the Andes) seems to be out of the altitudinal and longitudinal reach of the genus. The ranges of two species of Mesomerodon overlap in aseasonal Ecuador in a region known for the highest levels of mammal and plant species diversity (Hoorn et al. 2010). Rutelinae biodiversity in Ecuador is the highest recorded in South America, with 53 genera and 298 species (Paucar-Cabrera 2005). Of these, 92 species of Rutelinae (or 36%) are endemic to the country (Paucar-Cabrera 2005). The distribution of the genus is restricted to low elevations alongside the Andes without extending eastward into the Brazilian Amazon. Mesomerodon exhibits a distributional gap between M. spinipenne and the Ecuadorian species in northern Peru.
Niche modeling. Within the Andean corridor, the genus level distribution model is congruent with the specimen-based distribution (Fig. 19 vs. Fig. 18). Therefore, the apparent distributional gap between Mesomerodon spinipenne and the Ecuadorian species in northern Peru is unlikely a result of a lack of sampling. The distribution model of the Ecuadorian species (Fig. 20) suggests that either M. barclayi sp. n. or M. gilletti extend into northern Peru. The distribution model for M. spinipenne (Fig. 21) shows a continuous distribution in central and southern Peru with a disconnected population in Bolivia (also corroborated with the specimen-based distribution). Specimen-level data do not corroborate the occurrence of M. spinipenne in western Brazil, northern Colombia, or Venezuela. Collecting may reveal occurrence of the taxon in western Brazil, but we consider it unlikely that the taxon occurs in Colombia or Venezuela because these countries have been well collected.
Diagnosis. Males of Mesomerodon barclayi sp. n. are differentiated from other Mesomerodon species by the following combination of characters: Protarsomere 2 ventrally lacking a striated region at the ventral apex (striated in M. spinipenne; apical region lacking striae in M. gilletti ) and form of the male genitalia (Fig. 22 versus Fig. 24 in M. spinipenne and Fig. 23 in M. gilletti). Females can only be confidently determined when associated to male specimens from the same collecting event.
Etymology. It is our honor to dedicate this species to Max Barclay (Curator, Coleoptera, Department of Entomology), who invited the first and second authors to the 1 st Scarab Symposium at the BMNH in 2014 where cooperation on this work was initiated. Max Barclay has led the way in making biodiversity science more accessible to scientists and citizens alike.
Distribution (Fig. 18). Known from western (lowland) Amazonia areas in Ecuador and occurring in apparent sympatry with M. gilletti. Ohaus (1934Ohaus ( , 1952 recorded M. spinipenne from Sarayacu, Ecuador. We recovered these specimens in ZMHB and ZSM and identified them as belonging to M. barclayi sp. n. Locality data (Suppl. material 1). 20 specimens from 9 collections. ECUADOR: Morona-Santiago, Orellana, Pastaza Temporal data. Based on label data, this species is known to be active in the months of February, August, and November.
Natural history. Based on label data, adult M. barclayi sp. n. is usually collected at light, thus suggesting activity and feeding at night. Individuals probably occur throughout the year. Immature stages are unknown. Specimens have been recorded at an elevation of 300 m.
Diagnosis. Mesomerodon gilletti males are differentiated from other Mesomerodon species by the following combination of characters: Protarsomere 2 of male ventrally with striated region poorly defined or lacking at apex (striated in M. spinipenne; lacking in M. barclayi sp. n.; ) and parameres, and ventral sclerite of phallobase ( Fig. 23 versus Fig. 24 in M. spinipenne and Fig. 22 in M. barclayi sp. n.). Mesomerodon gilletti females can only be confidently determined when associated with male specimens from the same collecting event.
Diagnosis. Mesomerodon spinipenne males are differentiated from other Mesomerodon species by the following combination of characters: Protarsomere 2 with well-defined striae at ventral apex (  (Figs 6-7). Distribution (Fig. 18). Mesomerodon spinipenne is the most broadly distributed species in the genus, and it occurs from central Peru to central Bolivia in the western (lowland) Amazonia (from 230 to 762 m elevation). The species has been recorded from Brazil, "Rio Purus" (Blackwelder 1944, Ohaus 1905, 1934, 1952, Machatschke 1972, Krajcik 2008, Soula 2008, Moore et al. 2017), but in our view this is an erroneous record. The record is based on a paralectotype female collected at the Rio Purus that rises in the Uyacali Region in Peru and flows into Brazil and which could, therefore, have been collected either in Peru or Brazil. Based on our examination of 213 specimens, we think that this specimen represents a Peruvian locality. Records of the species from Ecuador (Blackwelder 1944, Ohaus 1918, 1934, 1952, Machatschke 1972, Moore et al. 2017, Paucar-Cabrera 2005 are incorrect. These are records for either M. gilletti or M. barclayi sp. n.
Natural history. Based on label data, adult M. spinipenne are active at night and can be collected at lights. Immature stages are unknown.
Remarks. Based on body measurements provided by Ohaus (1905), we conclude that he had a minimum of 3 specimens: 1 male (length 18.5mm, width 10.5mm) and 2 females (length 22-23.5mm, width 12.5-13.5mm) from two localities (Chuchuras and Rio Purus). The lectotype specimen at ZMHB was designated by Soula (2008), and it is a male specimen labeled by Ohaus as "type" ("Mesomerodon spinipenne, Type ♂ Ohs.") and with the locality label "bei Pozuzu, Eckardt S." and "O. Peru, Chuchurras". As part of this research, we found and labeled specimens that were invalidly labeled as type specimens and that do not belong to the type series (see "Type Specimens").
It is possible that the specific epithet, "spinipenne", refers to the apex of the elytra in the male which possess an apical spine or tubercle. The Latin root "spini" refers to spine or thorn, and the Latin root "penna" refers to wing. This character state is not unique to M. spinipenne; instead, it is a synapomorphy for all species in the genus.

Discussion
The genus Mesomerodon is composed of three very similar species, two of which that have evaded discovery since the description of the genus by Ohaus (1905). We discovered specimens of our new species, M. barclayi sp. n. in collections that were studied by Ohaus and Soula, both experts who were not able to detect this cryptic species based on external features. Stasis in external morphology, in combination with the apparent sympatry of two Mesomerodon species in Ecuador, corroborate two hypotheses for generation of cryptic species: nonvisual mating signals and ecological specialization in similar niches (Bickford et al. 2007). Sympatric distribution and external similarity of M. barclayi sp. n. and M. gilletti suggest that sexual selection might be a driver for diversification in the genus. Only aedeagal characters differ between species, thus sexual selection by female choice may drive the evolution of male genitalia (Eberhard 1985), and this could be accompanied with differences in mating pheromones or mating calls (Bickford et al. 2007). It is possible that specialization in food plants or other life history-dependent factors may drive diversification. Studies of presumed dietary generalists in narrow ecological regions have revealed cryptic beetle and butterfly species complexes with dietary specializations (Blair et al. 2005, Hebert et al. 2004. Co-distributed cryptic species complexes may be a function of the western (lowland) Amazonian region with its aseasonal climate, humid forest, and heterogeneous vegetation. The distribution of two species of Mesomerodon in Ecuador coincides with highest global diversity of passerine birds and anurans (InfoNatura 2007) as well as the region for highest wood biomass productivity (Hoorn and Wesselingh 2010). The Ecuadorian Amazonian region descends from the foothills to elevations of 200-400 m and receives approximately 282 cm of precipitation annually (Dangles et al. 2009). The absence of a prolonged dry season, varied topography, and warm temperatures make the region a hotspot for biodiversity (Myers et al. 2000). In this region, small differences in elevation and vegetational cover create refuges (Dangles et al. 2009) that may allow for ecological niche diversification, especially for herbivorous species such as those in the genus Mesomerodon.
Future studies associated with these cryptic species are needed to examine divergence, population structure, and sister group relationships of the genus. Molecular data will allow association of males and females for each species. Focused fieldwork could determine distributional limits and yield ecological data which will assist in understanding the origin and cause of sympatry.