﻿A second species of the pill millipede genus Nearctomeris Wesener, 2012 (Diplopoda, Glomerida) from the Great Smoky Mountains, USA

﻿Abstract We describe a second species of Nearctomeris Wesener, 2012, a genus of pill millipede endemic to the southern Appalachians, based on morphological and molecular evidence. The fauna of Glomerida in America is characterized by its low diversity, and Nearctomerissmokysp. nov. is only the fifth species of the order known from the eastern United States. Our phylogenetic analyses based on COI sequences recover a tentatively monophyletic lineage including both eastern American genera Onomeris Cook, 1896 and Nearctomeris, with a common ancestor in the Late Cretaceous to Mid Eocene and extant diversity within genera dating back to the Miocene. Our results suggest that the observed low diversity of the group in the eastern US is likely caused by extinction events, but it is also possible that new species are yet to be found. We provide new records for Nearctomerisinexpectata Wesener, 2012, Onomerisunderwoodi Cook, 1896 and O.australora Hoffman, 1950; the latter is here reported for the first time from South Carolina. We also present DNA barcoding data for all species of Glomerida present in the US that are not yet publicly available.


Introduction
The order Glomerida, commonly known as pill millipedes, is a small group of Diplopoda with a mostly Holarctic distribution (Enghoff 2015).Its major center of diversity is the Western Palaearctic, especially within the Mediterranean peninsulas (Wesener 2010).In America the group is poorly diversified, and only three genera are present.The genus Glomeroides Chamberlin, 1922, belonging to the family Protoglomeridae, includes 15 species found in Nearctic and Neotropical forests from central Mexico to Guatemala, one present in coastal forests in Central California (Hoffman 1999), and at least one as-yet undescribed species from southern Texas (Wesener 2010).
In the eastern United States, the only known glomeridans belong to the family Glomeridae: three species in the genus Onomeris Cook, 1896, and a single species in the genus Nearctomeris Wesener, 2012 (Wesener 2010(Wesener , 2012)).The available distribution data of all four species is still scarce, but most records are located at low to mid elevations along the southern Appalachians.This region ZooKeys 1166: 333-349 (2023), DOI: 10.3897/zookeys.1166.103516Ernesto Recuero & Michael S. Caterino: New species of Nearctomeris Wesener, 2012 is one of the oldest emergent ranges in the World and a major center of diversity for several groups of organisms (e.g., Crandall and Buhay 2008;Barnes and Clark 2017), including some millipede families (Hoffman 1969;Marek and Bond 2006).The low diversity observed among pill millipedes is thus surprising, although it is possible that further undescribed species are still to be found (Hoffman 1999;Wesener 2010).
The genus Nearctomeris is so far known from a single species, N. inexpectata Wesener, 2012, only recorded from three localities in the southern Appalachian Mountains, in the states of Alabama, Tennessee and North Carolina, often in association with other pill millipede species in the genus Onomeris (Wesener 2012).Both genera are very similar in their general appearance and this, together with its apparently patchy and reduced distribution, may explain how Nearctomeris inexpectata has been overlooked until so recently.However, a closer examination of specimens shows clear differences that allows an easy diagnosis, such as the marked Y-shaped crest and antennal grooves present in Onomeris and missing in Nearctomeris (Wesener 2012).Among species, external differences are more subtle, and identification generally requires the dissection of specimens to examine the male telopods (Wesener 2010).
Here, we describe a second species of the genus Nearctomeris from the Great Smoky Mountains, based on morphological and molecular evidence.Also, we provide new records for N. inexpectata, O. underwoodi Cook, 1896 andO. australora Hoffman, 1950, andcomplete the DNA barcoding data for the known species of American Glomeridae.The molecular data is analyzed to propose the first hypothesis on the relationships and age of diversification within the family in North America.

Material and methods
Specimens of the new species were collected by sifting leaf litter from Whiteoak Sink, near the mouth of Waterfall Cave (Tennessee: Blount Co.), in the western part of the Great Smoky Mountains National Park (Fig. 1).The litter samples were processed using Berlese-Tullgren funnels and specimens collected directly into 100% ethanol.
General habitus pictures were generated by focus stacking 20 images taken with a Nikon EOS 6D camera equipped with a Tamron AF 1.4× teleconverter and a Canon MP-E 65 mm macro lens, using a Visionary Digital Passport system, and combined with Helicon Focus software v.8.1.1 (HeliconSoft, Ukraine).Morphological examination was performed using an Olympus SZX7 stereomicroscope, and a Zeiss Axioskop 50 compound microscope.Drawings were prepared using a camera lucida and digitized using a Wacom Intuos Pro tablet.Tegument details were examined using a Hitachi S3400N variable pressure scanning electron microscope (SEM), using uncoated specimens.
We studied additional material (Table 1, Suppl.material 1) corresponding to other species of Glomeridae from eastern US, including samples collected by the authors and colleagues, and others found among sample residues (i.e., leftovers of former samplings after pulling out specimens of immediate interest) stored in the Clemson University Arthropod Collection and made readily searchable online by posting photos of the different samples (https:// www.cuacinsects.org/databases.html).We also collected records of eastern US Glomeridae from published papers (Loomis 1943;Hoffman 1950;Causey 1959;Shelley 2000;Wesener 2010Wesener , 2012) ) and the citizen science initiative iNaturalist (https://www.inaturalist.org)(Suppl.material 1).Maps showing these records were generated using QGIS v.3.30(http://qgis.org/).The type material of the new species is deposited at the U.S. National Entomological Collection (USNM, Smithsonian National Museum of Natural History), the collections of the Great Smoky Mountains National Park (GRSM) and the Clemson University Arthropod Collection (CUAC) (see examined material).All non-type material is deposited at CUAC.
We extracted DNA from several specimens (Table 1), including the new species, using the GeneJET Genomic DNA Purification Kit (Thermo Fisher Scientific, Waltham, MA, USA), following standard protocol, and eluting in molecular grade water.We amplified a 658 bp fragment of the barcoding region of the Cytochrome Oxidase Subunit I (COI) mitochondrial gene using the primers LCO1490 and HCO2198 (Folmer et al. 1994).Polymerase chain reaction (PCR) reactions were performed in a 25 μL volume using the conditions described in Recuero and Rodríguez-Flores (2020).We visualized PCR products in a 1% agarose gel electrophoresis to check PCR success, and sent them to Psomagen, Inc. (Maryland, USA) for cleaning and sequencing.Sequences are deposited in GenBank (Table 1).
Sequences were compiled, assembled, and edited using Sequencher v.5.4.1 (Gene Codes Corporation), and aligned manually, including several Glomerida taxa available in GenBank from previous works (Spelda et al. 2011;Wesener 2012Wesener , 2015aWesener , 2015bWesener , 2022;;Oeyen and Wesener 2015;Wilbrandt et al 2015;Wesener and Conrad 2016;Reip and Wesener 2018;Liu and Golovatch 2020;Nguyen et al. 2021;Kuroda et al. 2022aKuroda et al. , 2022b) ) (Table 1).Sequences were translated to amino acids to check the presence of stop codons; this way, we were able to identify a few of our fragments as nuclear mitochondrial DNA (NUMT), which were removed from our analyses (Table 1).Mean uncorrected pairwise genetic distances (p-distances) among species were calculated using MEGA6 (Tamura et al. 2013).Phylogenetic hypotheses were generated on an unpartitioned matrix using maximum likelihood (ML) and a Bayesian approach.ML analyses were performed with W-IQ-TREE (available at http://iqtree.cibiv.univie.ac.at; Trifinopoulos et al. 2016), allowing the program to select the best fitting substitution model, and measuring branch support with 1000 ultrafast bootstrap replicates.Bayesian phylogenetic inference, including estimates of time to the most recent common ancestor (TMRCA), was performed with BEAST v.1.10.4 (Drummond et al. 2012); we used a GTR+G+I substitution model as estimated in the previous analysis, a birth-death tree prior and a lognormal relaxed molecular clock; given our limited taxonomic sampling and data, the few available fossil records for Glomerida could not be adequately incorporated as a calibration point into the analysis (Wesener 2019), so we fixed a substitution rate of 2.3%/Myr (0.0115 substitution/site/Myr), widely used for Arthropoda COI sequence data (Brower 1994).The analysis was run for 100 million generations, sampling every 10000, and repeated independently four times to check consistency of the results, yielding in all cases high effective sample sizes (> 200) for all parameters as checked with Tracer v.1.7.2 (Rambaut et al. 2018).A maximum clade credibility tree was built with TreeAnnotator v.1.10.4 considering a 25% burn-in.

Results
The phylogenetic analyses results are limited by using just a 658 bp fragment of the mitochondrial COI gene, and most clades lack support in either ML or BEAST trees, especially in the basal relationships (Fig. 2).As expected, we observe a sister relationship between Nearctomeris inexpectata and N. smoky sp.nov., with a mean genetic p-distance of 6.85%.The clade including all Onomeris species is also well supported, with O. australora being the sister lineage to a clade containing O. sinuata (Loomis, 1943)

Hyleoglomeris lobus Tonkinomeris huzhengkun Tonkinomeris napoensis Rhopalomeris sauda Rhopalomeris nagao
Peplomeris magna Hyperglomeris sp.New records of Nearctomeris inexpectata, Onomeris underwoodi and Onomeris australora are provided in Table 1 and shown in Fig. 1.Onomeris australora is reported here for the first time in the state of South Carolina, within the range of the Blue Ridge Mountains.
Diagnosis.Externally similar to Nearctomeris inexpectata; they can be differentiated in the shape of the femoral process (distal finger) of telopods (Fig. 5E), blunt and rounded in N. inexpectata, elongated and curved in the N. smoky sp.nov., and of the syncoxite (Fig. 5D), distally broader in N. inexpectata than in N. smoky sp.nov., with rounded central lobe in N. inexpectata and bilobed in N. smoky sp.nov., and lateral processes shorter in N. inexpectata than in N. smoky sp.nov.The observed mean COI pairwise uncorrected p-distance between both species is 6.85%.
Name.Smoky, a noun in apposition, refers to the Great Smoky Mountains where the species lives.
Description.Body with 12 segments (including collum).Length of largest male (holotype), 2.9 mm; width at thoracic shield 1.8 mm, at tergite five 2 mm; height of thoracic shield 1.3 mm.Length of largest female 3.2 mm; width at thoracic shield 2 mm, at tergite five 2.1 mm; height of thoracic shield 1.3 mm.
General coloration of adults (Fig. 3) brown to dark brown dorsally; collum with a large, central, off-white area; tergites 2-11 with lateral, transversely oval, off-white areas; lateral and posterior margins of tergites translucent; head brown, more or less mottled with white, labrum and organ of Tömösváry white, ocular field black; antennae brown mottled with white; ventrally off-white, legs white, mottled or not with brown, with brown tarsi.Juveniles with 11 segments with similar pattern but much lighter than adults, and even lighter in juveniles with 10 segments (Fig. 4A).
Collum (Fig. 4B): surface smooth, finely and densely punctured dorsally with minute pits, as the rest of segments, with two well-marked, transverse striae.
Thoracic shield (Fig. 4B): with schisms rounded posteriorly, well-differentiated but not protruding beyond tergite contour.Schism impression broad and well-developed.Only three striae transversely crossing the shield; no trace of a central, incomplete stria.One strong, one very weak lateral stria below the schism impression.3A, 4A): with soft, shiny appearance; surface densely covered with small pits that, observed through the transparent first layer of the tegument, seem to be the opening of pore canals connecting the epidermis with the exterior.There is no trace of setae on the tergites.Tergite 11 is partially hidden under tergite 10.

Tergites (Figs
Mid-body legs (Fig. 5A): relatively slender, femur about 2.5 times as long as wide, tarsus 5-5.5 times as long as wide.Ventral margin of prefemur and femur with numerous strong setae.Tarsus with 5-6 ventral, 2-3 dorsal spiniform setae, mostly set in the distal half.Claw 4-4.5 times as long as wide.Anal shield (Fig. 3B): with evenly rounded posterior border, with no sign of a notch or concavity.
Male leg pair 17 (Fig. 5B): with a broad coxal lobe, mesally with a spiniform seta; telopodite formed by three strongly reduced podomeres, first and second with mesal spiniform setae, third with an apical spiniform seta.

Discussion
Some of the oldest fossils of Oniscomorpha, a probably non-monophyletic group including the different orders of pill millipedes (Benavides et al. 2023), have been found at the Middle Pennsylvanian deposits of Mazon Creek, Illinois, US, with an age of over 300 Myr (Hannibal and Feldmann 1981).Although most of the species belong to the already extinct order Amynilyspedida, some fossils were tentatively assigned to Sphaeroteriida and, according to some authors, they could even correspond to Glomerida (Shelley and Golovatch 2011, but see Racheboeuf et al. 2004).However, the current diversity of oniscomorph millipedes in America is notably impoverished, with only a few known species of Glomerida (Hoffman 1999).
In the eastern United States, with only five species in two genera, the low known diversity of pill millipedes could be explained by different reasons.On one hand it could be caused by a relatively recent colonization, with little time for subsequent diversification.It has been hypothesized that both genera might not represent sister lineages and that they could be more closely related to Asian taxa such as Hyleoglomeris Verhoeff, 1910 andHyperglomeris Silvestri, 1917 rather than to each other (Wesener 2012, but see Liu and Golovatch 2020).This hypothesis is compatible with the recent colonization scenario, but also with old events of dispersal and extinction on intervening geographic areas, as has been proposed for some species in the millipede genus Brachycybe Wood, 1864 (Brewer et al. 2012).
Alternatively, the current species could represent relics of a formerly more diverse fauna affected by high extinction rates, in which case we would expect a sister relationship between Onomeris and Nearctomeris.Our phylogenetic analyses favor this idea, supporting a monophyletic lineage including both genera, with a common ancestor dating back most likely to the Mid Eocene, and no close relationship with any of the Palaearctic taxa included in the analyses.However, the proposed relationships are tentative, being based in a single locus, likely affected by saturation, and suffering from reduced taxonomic sampling considering the global diversity of Glomerida, as has happened in previous attempts to resolve the phylogeny of the order using molecular data (Oeyen and Wesener 2015;Liu and Golovatch 2020).Diversification of extant lineages within Onomeris and Nearctomeris occurred during the Middle and Upper Miocene, resulting in long branches that are suggestive of diversity loss, and supporting the hypothesis of old relict species of a more diverse fauna affected by high extinction rates.The clades including the Asian Hyleoglomeris, or the European Glomeris Latreille, 1802, exhibit long but densely bifurcated branches, suggesting that extinction has been considerably lower in those lineages than in eastern US taxa.
Our age estimates must be considered as tentative, considering the limitation of our dataset and that they are based on a substitution rate that, even if widely used across different groups of Arthropoda including millipedes (e.g., Brewer et al. 2012;Nielsen et al. 2022), could be different in a group like Glomerida.It has been found that inter-and intraspecific distances in other glomeridan genera, such as Glomeris, are unusually high for COI sequences (Wesener 2015a;Wilbrandt et al. 2015;Wesener and Conrad 2016;Reip and Wesener 2018).This could indicate either a faster substitution rate or, alternatively, old speciation events or presence of cryptic species not yet delimited; however, we still do not have the necessary data to test these scenarios.In the case of a faster substitution rate, our estimates would be overly old, and the actual ages of the clades would be more recent.
However, we must not discount the likelihood that there is still some diversity that has not been identified yet.In fact, the strong morphological conservatism typical of both genera could be hiding a higher diversity than reflected by current taxonomy, and it has been generally assumed that further species should be found and described (Hoffman 1999;Shelley 2000).An integrative taxonomic approach could help detect those cases, both in Onomeris and Nearctomeris, as has been shown in other Appalachian millipedes, both helping describe new species (e.g., Means et al. 2021) or synonymizing taxa (e.g., Vasquez-Valverde and Marek 2022).
In the case of Onomeris underwoodi a moderate interspecific morphological variability has been described (Causey 1959;Wesener 2010); considering its wide distribution it could be possible that there are morphologically similar or even cryptic species under that name.Unfortunately, we still lack sufficient sampling to tackle this point.The best represented species in our analyses, Onomeris australora, has shown shallow genetic divergences in the southern portion of its distribution, but we have no data from populations further north.Onomeris sinuata is known from just a few, isolated and widely separated localities; it could be that the species has not been found yet in intervening areas, but if the isolation is real then the existence of independent lineages or even species may be predicted; a similar pattern is observed within Nearctomeris inexpectata (Wesener 2012).Glomerids are widespread in the eastern US, ranging from Florida and Mississippi in the south to Kentucky in the north, but within that huge region there are large areas where no reports are available.Further sampling, together with detailed morphological and molecular characterization are still necessary to determine the real diversity of pill millipedes in the eastern US.

Figure 4 .
Figure 4. Nearctomeris smoky sp.nov.A habitus of a 10-segment juvenile (CUAC) in lateral and ventral views B SEM image of collum and thoracic shield of a female (USNM), showing details of tegument, striae and schism.

Table 1 .
Extracted and sequenced specimens of Nearctomeris Wesener, 2012 and Onomeris Cook, 1896, and GenBank sequences included in the phylogenetic analyses.